diff --git a/data.json b/data.json
index e6ee076d3841f4c2f28a5b55a9cec9a3d06d2ced..1ca75b8e0c2d46cf87c1934b0b87597e80ef23a6 100644
--- a/data.json
+++ b/data.json
@@ -1 +1 @@
-{"rows": [["EvollaModel", "Reachable", "", 732, "westlake-repl/Evolla-10B-hf"], ["afmoe", "Reachable", "", 84004, "optimum-intel-internal-testing/tiny-random-trinity"], ["aimv2", "Reachable", "", 10627, "apple/aimv2-large-patch14-224-lit"], ["aimv2_text_model", "Unreachable", "component config only", 0, ""], ["aimv2_vision_model", "Reachable", "", 3412, "apple/aimv2-large-patch14-224"], ["albert", "Done", "", 2009876, "albert/albert-base-v2"], ["align", "Unreachable", "forward fails after shrink", 21429, "kakaobrain/align-base"], ["align_text_model", "Unreachable", "component config only", 0, ""], ["align_vision_model", "Unreachable", "component config only", 0, ""], ["altclip", "Reachable", "", 7652, "BAAI/AltCLIP"], ["altclip_text_model", "Unreachable", "component config only", 0, ""], ["altclip_vision_model", "Unreachable", "component config only", 0, ""], ["apertus", "Done", "", 457103, "swiss-ai/Apertus-8B-Instruct-2509"], ["arcee", "Reachable", "", 64187, "optimum-intel-internal-testing/tiny-random-ArceeForCausalLM"], ["aria", "Unreachable", "forward fails after shrink", 29760, "rhymes-ai/Aria"], ["aria_text", "Unreachable", "forward fails after shrink", 88, "hf-tiny-v2/tiny-random-AriaTextForCausalLM"], ["audio-spectrogram-transformer", "Unreachable", "forward fails after shrink", 1651174, "MIT/ast-finetuned-audioset-10-10-0.4593"], ["audioflamingo3", "Done", "", 144463, "nvidia/audio-flamingo-3-hf"], ["audioflamingo3_encoder", "Unreachable", "forward fails after shrink", 27, "hf-tiny-v2/tiny-random-AudioFlamingo3Encoder"], ["autoformer", "Unreachable", "synthetic inputs rejected", 79649, "huggingface/autoformer-tourism-monthly"], ["axk1", "Reachable", "", 5094, "skt/A.X-K1"], ["axk2", "Reachable", "", 90892, "skt/A.X-K2-NVFP4"], ["aya_vision", "Reachable", "", 3689, "CohereLabs/aya-vision-8b"], ["bailing_hybrid", "Unreachable", "forward fails after shrink", 727122, "rapid-mlx/Ling-3.0-tiny-MLX-4bit"], ["bamba", "Reachable", "", 11728, "ibm-ai-platform/Bamba-9B-v1"], ["bark", "Unreachable", "forward fails after shrink", 28264, "suno/bark"], ["bart", "Done", "", 6328395, "facebook/bart-large-mnli"], ["beam", "Reachable", "", 20667, "reflection/Beam-1.0-501B-A23B"], ["beit", "Done", "", 169127, "microsoft/dit-base-finetuned-rvlcdip"], ["bert", "Done", "", 527874026, "sentence-transformers/all-MiniLM-L6-v2"], ["bert-generation", "Reachable", "", 3386, "ybelkada/random-tiny-BertGenerationModel"], ["big_bird", "Done", "", 194505, "google/bigbird-roberta-base"], ["bigbird_pegasus", "Reachable", "", 71110, "optimum-intel-internal-testing/tiny-random-bigbird_pegasus"], ["biogpt", "Reachable", "", 56866, "optimum-intel-internal-testing/tiny-random-BioGptForCausalLM"], ["bit", "Unreachable", "model build fails", 8676, "google/bit-50"], ["bitnet", "Reachable", "", 35767, "optimum-intel-internal-testing/tiny-random-bitnet"], ["blenderbot", "Done", "", 235225, "optimum-intel-internal-testing/tiny-random-BlenderbotModel"], ["blenderbot-small", "Reachable", "", 5857, "facebook/blenderbot_small-90M"], ["blip", "Unreachable", "forward fails after shrink", 2359364, "Salesforce/blip-image-captioning-base"], ["blip-2", "Unreachable", "forward fails after shrink", 457477, "Salesforce/blip2-opt-2.7b"], ["blip_2_qformer", "Unreachable", "forward fails after shrink", 145, "hf-tiny-v2/tiny-random-Blip2QFormerModel"], ["blip_2_vision_model", "Unreachable", "component config only", 24, "tmnam20/blip2-opt-2.7b-vision"], ["blip_text_model", "Unreachable", "component config only", 0, ""], ["blip_vision_model", "Unreachable", "component config only", 0, ""], ["bloom", "Done", "", 1527823, "bigscience/bloomz-560m"], ["blt", "Unreachable", "forward fails after shrink", 6999, "itazap/blt-1b-hf"], ["blt_global_transformer", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["blt_local_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["blt_local_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["blt_patcher", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["bridgetower", "Reachable", "", 80, "regisss/bridgetower-newyorker-a100-8x"], ["bridgetower_text_model", "No checkpoint", "no Hub checkpoint has this model_type (documented BridgeTower/bridgetower-base is another type)", 0, ""], ["bridgetower_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (documented BridgeTower/bridgetower-base is another type)", 0, ""], ["bros", "Unreachable", "forward fails after shrink", 42661, "naver-clova-ocr/bros-base-uncased"], ["camembert", "Done", "", 2226781, "almanach/camembert-base"], ["canary", "Unreachable", "forward fails after shrink", 237192, "nvidia/canary-1b-v2"], ["canary_decoder", "Unreachable", "component config only", 0, ""], ["canine", "Done", "", 167492, "google/canine-c"], ["chameleon", "Reachable", "", 8504, "facebook/chameleon-7b"], ["chameleon_vqgan", "Unreachable", "component config only", 33, "leloy/Anole-7b-v0.1-vqvae-hf"], ["chinese_clip", "Done", "", 276376, "OFA-Sys/chinese-clip-vit-base-patch16"], ["chinese_clip_text_model", "Unreachable", "component config only", 0, ""], ["chinese_clip_vision_model", "Reachable", "", 57, "hf-tiny-v2/tiny-random-ChineseCLIPVisionModel"], ["chmv2", "Unreachable", "forward fails after shrink", 1526, "facebook/dinov3-vitl16-chmv2-dpt-head"], ["clap", "Done", "", 9104856, "laion/clap-htsat-fused"], ["clap_audio_model", "Unreachable", "component config only", 257, "NGDtuanh/abook-music-student"], ["clap_text_model", "Unreachable", "component config only", 61, "werent4/larger_clap_general_text_model"], ["clip", "Done", "", 37612455, "openai/clip-vit-base-patch32"], ["clip_text_model", "Done", "", 940714, "peft-internal-testing/tiny-clip-text-2"], ["clip_vision_model", "Done", "", 432156, "xingpng/OneIG-StyleEncoder"], ["clipseg", "Done", "", 545844, "CIDAS/clipseg-rd64-refined"], ["clipseg_text_model", "Unreachable", "component config only", 0, ""], ["clipseg_vision_model", "Unreachable", "component config only", 0, ""], ["clvp", "Unreachable", "forward fails after shrink", 5530, "susnato/clvp_dev"], ["clvp_decoder", "Unreachable", "component config only", 0, ""], ["clvp_encoder", "Unreachable", "component config only", 0, ""], ["codegen", "Unreachable", "forward fails after shrink", 168810, "Salesforce/codegen2-16B_P"], ["cohere", "Done", "", 166589, "optimum-intel-internal-testing/tiny-random-CohereForCausalLM"], ["cohere2", "Done", "", 162977, "optimum-intel-internal-testing/tiny-random-aya-base"], ["cohere2_moe", "Reachable", "", 15203, "CohereLabs/North-Mini-Code-1.0"], ["cohere2_vision", "Reachable", "", 37679, "CohereLabs/command-a-plus-05-2026-bf16"], ["cohere_asr", "Unreachable", "forward fails after shrink", 350010, "CohereLabs/cohere-transcribe-arabic-07-2026"], ["cohere_compass", "Reachable", "", 34840, "CohereLabs/North-Micro-Vision-Instruct"], ["cohere_compass_text", "Reachable", "", 41, "hf-tiny-v2/tiny-random-CohereCompassTextForSequenceClassification"], ["cohere_compass_vision", "Unreachable", "forward fails after shrink", 15, "hf-tiny-v2/tiny-random-CohereCompassVisionModel"], ["colmodernvbert", "Unreachable", "no usable checkpoint", 1649, "Qdrant/colmodernvbert"], ["colpali", "Reachable", "", 8568, "vidore/colpali-v1.3-hf"], ["colqwen2", "Reachable", "", 36700, "vidore/colqwen2-v1.0-hf"], ["conditional_detr", "Reachable", "", 62673, "yainage90/fashion-object-detection"], ["convbert", "Done", "", 509373, "YituTech/conv-bert-base"], ["convnext", "Done", "", 192305, "facebook/convnext-tiny-224"], ["convnextv2", "Done", "", 330963, "facebook/convnextv2-base-22k-384"], ["cosmos3_edge", "Reachable", "", 83288, "nvidia/Cosmos3-Edge"], ["cosmos3_edge_text", "Reachable", "", 14, "hf-tiny-v2/tiny-random-Cosmos3EdgeTextModel"], ["cosmos3_edge_vision", "Unreachable", "forward fails after shrink", 12, "hf-tiny-v2/tiny-random-Cosmos3EdgeVisionModel"], ["cosmos3_omni", "Done", "", 244937, "nvidia/Cosmos3-Nano"], ["cpmant", "Reachable", "", 3742, "openbmb/cpm-ant-10b"], ["csm", "Done", "", 145078, "sesame/csm-1b"], ["csm_depth_decoder_model", "Unreachable", "checkpoint gated", 0, "sesame/csm-1b"], ["ctrl", "Reachable", "", 17032, "Salesforce/ctrl"], ["cvt", "Reachable", "", 12319, "microsoft/cvt-13"], ["cwm", "Unreachable", "no usable checkpoint", 1007, "facebook/cwm"], ["d_fine", "Unreachable", "forward fails after shrink", 126298, "ustc-community/dfine-xlarge-obj2coco"], ["dab-detr", "Unreachable", "config load fails", 7050, "IDEA-Research/dab-detr-resnet-50"], ["dac", "Unreachable", "forward fails after shrink", 72000, "descript/dac_44khz"], ["data2vec-audio", "Reachable", "", 39013, "optimum-intel-internal-testing/tiny-random-Data2VecAudioModel"], ["data2vec-text", "Reachable", "", 7219, "optimum-intel-internal-testing/tiny-random-Data2VecTextModel"], ["data2vec-vision", "Unreachable", "shrink fails", 17284, "optimum-intel-internal-testing/tiny-random-Data2VecVisionModel"], ["dbrx", "Reachable", "", 16983, "trl-internal-testing/tiny-DbrxForCausalLM"], ["deberta", "Done", "", 692267, "microsoft/deberta-large-mnli"], ["deberta-v2", "Done", "", 21746039, "microsoft/mdeberta-v3-base"], ["decision_transformer", "Unreachable", "synthetic inputs rejected", 6158, "edbeeching/decision-transformer-gym-hopper-medium"], ["deepseek_ocr2", "Unreachable", "synthetic inputs rejected", 22922, "optimum-intel-internal-testing/tiny-random-deepseek-ocr-2"], ["deepseek_ocr2_encoder", "Unreachable", "component config only", 0, ""], ["deepseek_ocr2_sam_vision_model", "Unreachable", "component config only", 0, ""], ["deepseek_ocr2_text", "Unreachable", "component config only", 0, ""], ["deepseek_ocr2_vision", "Unreachable", "component config only", 0, ""], ["deepseek_v2", "Done", "", 668975, "deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct"], ["deepseek_v3", "Done", "", 4950570, "deepseek-ai/DeepSeek-R1"], ["deepseek_v32", "Done", "", 6161743, "deepseek-ai/DeepSeek-V3.2"], ["deepseek_v4", "Done", "", 6213673, "deepseek-ai/DeepSeek-V4-Flash-0731"], ["deepseek_vl", "Reachable", "", 8828, "deepseek-community/deepseek-vl-1.3b-chat"], ["deepseek_vl_hybrid", "Reachable", "", 8249, "deepseek-community/deepseek-vl-7b-chat"], ["deformable_detr", "Reachable", "", 23013, "SenseTime/deformable-detr"], ["deimv2", "Unreachable", "forward fails after shrink", 4393, "alessioarcara/deimv2-deimv2_dinov3_x_coco"], ["deit", "Reachable", "", 50030, "facebook/deit-base-distilled-patch16-384"], ["depth_anything", "Unreachable", "forward fails after shrink", 2447209, "depth-anything/Depth-Anything-V2-Small-hf"], ["depth_pro", "Unreachable", "forward fails after shrink", 52957, "apple/DepthPro-hf"], ["detr", "Done", "", 385370, "TahaDouaji/detr-doc-table-detection"], ["dia", "Unreachable", "no usable checkpoint", 16396, "nari-labs/Dia-1.6B-0626"], ["dia_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented nari-labs/Dia-1.6B is another type)", 0, ""], ["dia_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented nari-labs/Dia-1.6B is another type)", 0, ""], ["diffllama", "Reachable", "", 4593, "kajuma/DiffLlama-0.3B-handcut"], ["diffusion_gemma", "Unreachable", "forward fails after shrink", 1034537, "google/diffusiongemma-26B-A4B-it"], ["diffusion_gemma_text", "Unreachable", "component config only", 0, ""], ["dinat", "Unreachable", "model build fails", 209, "shi-labs/dinat-mini-in1k-224"], ["dinov2", "Done", "", 7235906, "facebook/dinov2-base"], ["dinov2_with_registers", "Done", "", 354429, "facebook/dinov2-with-registers-base"], ["dinov3_convnext", "Reachable", "", 95750, "facebook/dinov3-convnext-tiny-pretrain-lvd1689m"], ["dinov3_vit", "Done", "", 2089303, "facebook/dinov3-vitl16-pretrain-lvd1689m"], ["distilbert", "Done", "", 20324420, "distilbert/distilbert-base-uncased"], ["doge", "Reachable", "", 8362, "SmallDoge/Doge-320M"], ["donut-swin", "Reachable", "", 2963, "optimum-intel-internal-testing/tiny-random-DonutSwinModel"], ["dots1", "Reachable", "", 6731, "rednote-hilab/dots.llm1.base"], ["dpr", "Reachable", "", 53829, "facebook/dpr-question_encoder-single-nq-base"], ["dpt", "Done", "", 620365, "Intel/dpt-hybrid-midas"], ["edgetam", "Unreachable", "no usable checkpoint", 9579, "onnx-community/EdgeTAM-ONNX"], ["edgetam_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_video", "Unreachable", "forward fails after shrink", 11709, "yonigozlan/EdgeTAM-hf"], ["edgetam_video_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_video_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_vision_model", "Reachable", "", 53, "hf-tiny-v2/tiny-random-EdgeTamVisionModel"], ["efficientloftr", "Unreachable", "shrink fails", 16245, "zju-community/efficientloftr"], ["efficientnet", "Unreachable", "forward fails after shrink", 168152, "docling-project/DocumentFigureClassifier-v2.5"], ["electra", "Done", "", 28465435, "google/electra-base-discriminator"], ["embedding_gemma2", "Done", "", 719174, "onnx-community/embeddinggemma-2-ONNX"], ["embedding_gemma2_text", "Reachable", "", 171, "HukukcuApp/tr-legal-embedding-gemma-2"], ["emu3", "Reachable", "", 20741, "BAAI/Emu3-Chat-hf"], ["emu3_text_model", "Unreachable", "component config only", 118, "hf-tiny-v2/tiny-random-Emu3TextModel"], ["emu3_vqgan", "Unreachable", "component config only", 0, ""], ["encodec", "Unreachable", "forward fails after shrink", 105065, "facebook/encodec_24khz"], ["encoder-decoder", "Done", "", 29649, "optimum-internal-testing/tiny-random-encoder-decoder-gpt2-bert"], ["eomt", "Unreachable", "forward fails after shrink", 67972, "tue-mps/coco_panoptic_eomt_large_640"], ["eomt_dinov3", "Unreachable", "forward fails after shrink", 21640, "tue-mps/eomt-dinov3-ade-semantic-large-512"], ["ernie", "Reachable", "", 73981, "shibing624/text2vec-base-chinese-paraphrase"], ["ernie4_5", "Reachable", "", 14731, "baidu/ERNIE-4.5-0.3B-PT"], ["ernie4_5_moe", "Unreachable", "forward fails after shrink", 123822, "baidu/ERNIE-4.5-21B-A3B-Thinking"], ["ernie4_5_vl_moe", "Unreachable", "forward fails after shrink", 45, "hf-tiny-v2/tiny-random-Ernie4_5_VLMoeForConditionalGeneration"], ["ernie4_5_vl_moe_text", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/ERNIE-4.5-VL-28B-A3B-PT is another type)", 0, ""], ["ernie4_5_vl_moe_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/ERNIE-4.5-VL-28B-A3B-PT is another type)", 0, ""], ["esm", "Done", "", 7030784, "facebook/esm2_t33_650M_UR50D"], ["esmc", "Done", "", 393143, "biohub/ESMC-300M"], ["esmfold2", "Unreachable", "shrink fails", 388538, "biohub/ESMFold2"], ["eurobert", "Done", "", 847801, "OpenMed/OpenMed-NER-ChemicalDetect-EuroMed-212M"], ["evolla", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["exaone4", "Reachable", "", 64834, "LGAI-EXAONE/EXAONE-4.0-32B"], ["exaone4_5", "Reachable", "", 39320, "LGAI-EXAONE/EXAONE-4.5-33B"], ["exaone4_5_vision", "Unreachable", "forward fails after shrink", 31, "hf-tiny-v2/tiny-random-Exaone4_5_VisionModel"], ["exaone_moe", "Reachable", "", 18548, "LGAI-EXAONE/K-EXAONE-236B-A23B"], ["falcon", "Done", "", 350854, "tiiuae/falcon-7b"], ["falcon_h1", "Done", "", 104762, "tiiuae/Falcon-H1-3B-Instruct"], ["falcon_mamba", "Reachable", "", 61205, "trl-internal-testing/tiny-FalconMambaForCausalLM"], ["fast_vlm", "Reachable", "", 5841, "KamilaMila/FastVLM-7B"], ["fastspeech2_conformer", "Unreachable", "forward fails after shrink", 2880, "espnet/fastspeech2_conformer"], ["fastspeech2_conformer_hifigan", "No checkpoint", "no Hub checkpoint has this model_type (documented espnet/fastspeech2_conformer_with_hifigan is another type)", 0, ""], ["fastspeech2_conformer_with_hifigan", "Unreachable", "forward fails after shrink", 2800, "espnet/fastspeech2_conformer_with_hifigan"], ["flaubert", "Reachable", "", 74651, "flaubert/flaubert_base_uncased"], ["flava", "Reachable", "", 32370, "facebook/flava-full"], ["flava_image_codebook", "Unreachable", "component config only", 23, "facebook/flava-image-codebook"], ["flava_image_model", "Unreachable", "component config only", 0, ""], ["flava_multimodal_model", "Unreachable", "component config only", 0, ""], ["flava_text_model", "Unreachable", "component config only", 0, ""], ["flex_olmo", "Reachable", "", 12810, "allenai/FlexOlmo-7x7B-1T"], ["flint", "Reachable", "", 60545, "RecursiveSI/flint"], ["florence2", "Done", "", 4293112, "microsoft/Florence-2-base"], ["florence_vision", "Unreachable", "component config only", 0, ""], ["fnet", "Done", "", 478381, "google/fnet-base"], ["focalnet", "Reachable", "", 11569, "microsoft/focalnet-tiny"], ["fsmt", "Reachable", "", 57368, "facebook/wmt19-en-de"], ["fun_asr_nano", "Unreachable", "shrink fails", 162557, "FunAudioLLM/Fun-ASR-Nano-2512-hf"], ["fun_asr_nano_encoder", "Unreachable", "forward fails after shrink", 21, "hf-tiny-v2/tiny-random-FunAsrNanoEncoder"], ["funnel", "Reachable", "", 11894, "funnel-transformer/small"], ["fuyu", "Reachable", "", 12098, "adept/fuyu-8b"], ["gemma", "Done", "", 611743, "google/gemma-2b"], ["gemma2", "Done", "", 2138107, "hugging-quants/gemma-2-9b-it-AWQ-INT4"], ["gemma3", "Done", "", 6276380, "google/gemma-3-4b-it"], ["gemma3_text", "Done", "", 3985142, "google/embeddinggemma-300m"], ["gemma3n", "Done", "", 360811, "unsloth/gemma-3n-E4B-it"], ["gemma3n_audio", "Unreachable", "synthetic inputs rejected", 63, "hf-tiny-v2/tiny-random-Gemma3nAudioEncoder"], ["gemma3n_text", "Reachable", "", 2175, "optimum-intel-internal-testing/tiny-random-gemma3n-text"], ["gemma3n_vision", "Reachable", "", 0, ""], ["gemma4", "Done", "", 47591452, "google/gemma-4-31B-it"], ["gemma4_assistant", "Unreachable", "forward fails after shrink", 683069, "google/gemma-4-26B-A4B-it-assistant"], ["gemma4_audio", "Unreachable", "forward fails after shrink", 138, "rnagabh/gemma4-audio-encoder"], ["gemma4_text", "Reachable", "", 41837, "AlexKoff88/gemma4_tinystories"], ["gemma4_unified", "Done", "", 4749964, "google/gemma-4-12B-it"], ["gemma4_unified_assistant", "Unreachable", "forward fails after shrink", 37821, "google/gemma-4-12B-it-assistant"], ["gemma4_unified_audio", "Unreachable", "component config only", 0, ""], ["gemma4_unified_text", "Reachable", "", 2960, "jgeuter/Jeff-1.0-Medium"], ["gemma4_unified_vision", "Unreachable", "component config only", 0, ""], ["gemma4_vision", "Unreachable", "forward fails after shrink", 214, "rnagabh/gemma4-vision-encoder"], ["git", "Reachable", "", 46165, "microsoft/git-base"], ["git_vision_model", "Unreachable", "component config only", 0, ""], ["glm", "Unreachable", "no usable checkpoint", 43219, "optimum-intel-internal-testing/tiny-random-glm-edge"], ["glm4", "Done", "", 139087, "optimum-intel-internal-testing/tiny-random-glm4"], ["glm46v", "Unreachable", "no usable checkpoint", 170, "hf-tiny-v2/tiny-random-Glm46VModel"], ["glm4_moe", "Done", "", 1607130, "zai-org/GLM-4.5-Air"], ["glm4_moe_lite", "Done", "", 1802094, "zai-org/GLM-4.7-Flash"], ["glm4v", "Unreachable", "forward fails after shrink", 277739, "zai-org/GLM-4.1V-9B-Thinking"], ["glm4v_moe", "Unreachable", "forward fails after shrink", 35179, "zai-org/GLM-4.5V"], ["glm4v_moe_text", "Unreachable", "forward fails after shrink", 58, "hf-tiny-v2/tiny-random-Glm4vMoeTextModel"], ["glm4v_moe_vision", "Unreachable", "no usable checkpoint", 38, "hf-tiny-v2/tiny-random-Glm4vMoeVisionModel"], ["glm4v_text", "Unreachable", "forward fails after shrink", 58, "hf-tiny-v2/tiny-random-Glm4vTextModel"], ["glm4v_vision", "Unreachable", "forward fails after shrink", 38, "hf-tiny-v2/tiny-random-Glm4vVisionModel"], ["glm5_next", "Done", "", 7619607, "zai-org/GLM-5.3-Flash"], ["glm5_next_text", "Reachable", "", 23, "hf-tiny-v2/tiny-random-Glm5NextTextModel"], ["glm5_next_vision", "Unreachable", "forward fails after shrink", 65, "immanuelpeter/GLM-5.3-Flash-Vision"], ["glm_image", "Unreachable", "no usable checkpoint", 111, "hf-tiny-v2/tiny-random-GlmImageModel"], ["glm_image_text", "Unreachable", "no usable checkpoint", 55, "hf-tiny-v2/tiny-random-GlmImageTextModel"], ["glm_image_vision", "Unreachable", "no usable checkpoint", 35, "hf-tiny-v2/tiny-random-GlmImageVisionModel"], ["glm_image_vqmodel", "Unreachable", "no usable checkpoint", 6, "hf-tiny-v2/tiny-random-GlmImageVQVAE"], ["glm_moe_dsa", "Done", "", 7133417, "zai-org/GLM-5.3"], ["glm_ocr", "Done", "", 1798225, "zai-org/GLM-OCR"], ["glm_ocr_text", "Reachable", "", 51, "hf-tiny-v2/tiny-random-GlmOcrTextModel"], ["glm_ocr_vision", "Unreachable", "forward fails after shrink", 33, "hf-tiny-v2/tiny-random-GlmOcrVisionModel"], ["glmasr", "Reachable", "", 31331, "zai-org/GLM-ASR-Nano-2512"], ["glmasr_encoder", "Reachable", "", 50, "hf-tiny-v2/tiny-random-GlmAsrEncoder"], ["glmga", "No checkpoint", "no Hub checkpoint has this model_type (documented zai-org/GLM-4.1V-9B-Thinking is another type)", 0, ""], ["glpn", "Reachable", "", 8196, "vinvino02/glpn-kitti"], ["got_ocr2", "Done", "", 56531, "stepfun-ai/GOT-OCR-2.0-hf"], ["got_ocr2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["gpt-sw3", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["gpt2", "Done", "", 30545353, "openai-community/gpt2"], ["gpt_bigcode", "Done", "", 188077, "bigcode/tiny_starcoder_py"], ["gpt_neo", "Unreachable", "shrink fails", 865506, "EleutherAI/gpt-neo-125m"], ["gpt_neox", "Done", "", 7013890, "EleutherAI/pythia-160m"], ["gpt_neox_japanese", "Reachable", "", 46916, "optimum-intel-internal-testing/tiny-random-GPTNeoXJapaneseForCausalLM"], ["gpt_oss", "Done", "", 10738031, "openai/gpt-oss-20b"], ["gptj", "Unreachable", "forward fails after shrink", 188360, "EleutherAI/gpt-j-6b"], ["gradium_phonon_tts", "Unreachable", "shrink fails", 114118, "gradium/phonon-31m-en"], ["granite", "Done", "", 1699090, "ibm-granite/granite-4.1-3b"], ["granite4_vision", "Done", "", 194823, "ibm-granite/granite-vision-4.1-4b"], ["granite4_vision_text", "No checkpoint", "no Hub checkpoint has this model_type (documented ibm-granite4_vision_text/granite4_vision_text-3.0-8b-base does not exist)", 0, ""], ["granite_speech", "Done", "", 494648, "ibm-granite/granite-4.0-1b-speech"], ["granite_speech5_ctc", "Unreachable", "forward fails after shrink", 99951, "ibm-granite/granite-speech-5.0-470m-turboctc"], ["granite_speech5_encoder", "Unreachable", "forward fails after shrink", 17, "hf-tiny-v2/tiny-random-GraniteSpeech5Encoder"], ["granite_speech_encoder", "Unreachable", "component config only", 0, ""], ["granite_speech_plus", "Unreachable", "shrink fails", 29187, "ibm-granite/granite-speech-4.1-2b-plus"], ["granite_speech_plus_encoder", "Unreachable", "component config only", 0, ""], ["granite_swa", "Reachable", "", 7057, "ibm-granite/granite-swash-2b"], ["granitemoe", "Done", "", 244089, "ibm-granite/granite-3.1-3b-a800m-instruct"], ["granitemoe_swa", "Reachable", "", 6507, "ibm-granite/granite-swash-3b-a600m"], ["granitemoehybrid", "Done", "", 332014, "optimum-intel-internal-testing/tiny-random-granitemoehybrid"], ["granitemoeshared", "Unreachable", "no usable checkpoint", 1648, "ibm-research/moe-7b-1b-active-shared-experts"], ["grounding-dino", "Done", "", 2682931, "IDEA-Research/grounding-dino-base"], ["groupvit", "Reachable", "", 6660, "nvidia/groupvit-gcc-yfcc"], ["groupvit_text_model", "Unreachable", "component config only", 0, ""], ["groupvit_vision_model", "Unreachable", "component config only", 0, ""], ["gte", "Reachable", "", 56943, "Snowflake/snowflake-arctic-embed-m-v2.0"], ["helium", "Reachable", "", 6264, "kyutai/helium-1-preview-2b"], ["hgnet_v2", "Unreachable", "forward fails after shrink", 149, "ustc-community/hgnet-v2"], ["hiera", "Reachable", "", 11635, "facebook/hiera-tiny-224-hf"], ["higgs_audio_v2", "Reachable", "", 21331, "bosonai/higgs-tts-2-3b-base"], ["higgs_audio_v2_tokenizer", "Unreachable", "model build fails", 235573, "eustlb/higgs-audio-v2-tokenizer"], ["hrm_text", "Unreachable", "forward fails after shrink", 14676, "sapientinc/HRM-Text-1B"], ["hubert", "Done", "", 1327557, "facebook/hubert-large-ls960-ft"], ["hunyuan_v1_dense", "Done", "", 197468, "optimum-intel-internal-testing/tiny-random-hunyuan-v1-dense"], ["hunyuan_v1_moe", "Unreachable", "forward fails after shrink", 64830, "tencent/Hunyuan-A13B-Instruct"], ["hunyuan_vl", "Done", "", 807278, "tencent/HunyuanOCR"], ["hunyuan_vl_text", "Unreachable", "no usable checkpoint", 9, "hf-tiny-v2/tiny-random-HunYuanVLTextModel"], ["hunyuan_vl_vision", "Unreachable", "no usable checkpoint", 10, "hf-tiny-v2/tiny-random-HunYuanVLVisionTransformer"], ["hy_v3", "Done", "", 1408804, "tencent/Hy3"], ["hy_v4", "Reachable", "", 32339, "tencent/Hy4-preview-FP8"], ["hyperclovax", "Reachable", "", 20512, "naver-hyperclovax/HyperCLOVAX-SEED-Think-14B"], ["hyperclovax_vision_v2", "Reachable", "", 10279, "naver-hyperclovax/HyperCLOVAX-SEED-Think-32B"], ["ibert", "Reachable", "", 40201, "optimum-intel-internal-testing/tiny-random-ibert"], ["idefics", "Unreachable", "forward fails after shrink", 21036, "HuggingFaceM4/idefics-80b-instruct"], ["idefics2", "Reachable", "", 18780, "HuggingFaceM4/idefics2-8b"], ["idefics2_perceiver", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics2-8b is another type)", 0, ""], ["idefics2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics2-8b is another type)", 0, ""], ["idefics3", "Done", "", 776490, "ibm-granite/granite-docling-258M"], ["idefics3_vision", "Unreachable", "no usable checkpoint", 53, "hf-tiny-v2/tiny-random-Idefics3VisionTransformer"], ["idefics_perciever", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics-9b is another type)", 0, ""], ["idefics_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics-9b is another type)", 0, ""], ["ijepa", "Reachable", "", 23643, "facebook/ijepa_vith14_1k"], ["imagegpt", "Reachable", "", 7911, "openai/imagegpt-small"], ["informer", "Unreachable", "synthetic inputs rejected", 30807, "huggingface/informer-tourism-monthly"], ["inkling_audio", "No checkpoint", "no Hub checkpoint has this model_type (documented thinkingmachines/Inkling is another type)", 0, ""], ["inkling_mm_model", "Done", "", 3938529, "thinkingmachines/Inkling"], ["inkling_text", "Unreachable", "no usable checkpoint", 33, "hf-tiny-v2/tiny-random-InklingTextModel"], ["inkling_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented thinkingmachines/Inkling is another type)", 0, ""], ["instructblip", "Unreachable", "forward fails after shrink", 28294, "Salesforce/instructblip-flan-t5-xl"], ["instructblip_qformer", "Unreachable", "component config only", 0, ""], ["instructblip_vision_model", "Unreachable", "component config only", 0, ""], ["instructblipvideo", "Unreachable", "no usable checkpoint", 64, "hf-tiny-v2/tiny-random-InstructBlipVideoForConditionalGeneration"], ["instructblipvideo_qformer", "No checkpoint", "no Hub checkpoint has this model_type (documented Salesforce/instructblip-flan-t5-xl is another type)", 0, ""], ["instructblipvideo_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (documented Salesforce/instructblip-flan-t5-xl is another type)", 0, ""], ["internvl", "Done", "", 873854, "OpenGVLab/InternVL3_5-GPT-OSS-20B-A4B-Preview-HF"], ["internvl_vision", "Reachable", "", 105, "hf-tiny-v2/tiny-random-InternVLVisionModel"], ["jais2", "Reachable", "", 5408, "inception42/Jais-2-8B-Chat"], ["jamba", "Done", "", 210865, "ai21labs/Jamba-tiny-dev"], ["janus", "Done", "", 9275, "deepseek-community/Janus-Pro-1B"], ["janus_vision_model", "Unreachable", "component config only", 0, ""], ["janus_vqgan", "Unreachable", "component config only", 0, ""], ["jetmoe", "Reachable", "", 5517, "jetmoe/jetmoe-8b"], ["jina_embeddings_v3", "Done", "", 245085, "jinaai/jina-embeddings-v3-hf"], ["kimi_k25", "Reachable", "", 1381590, "moonshotai/Kimi-K2.5"], ["kimi_k25_vision", "Unreachable", "no usable checkpoint", 26, "Aquiles-ai/MoonViT-3D"], ["kimi_linear", "Unreachable", "shrink fails", 214668, "moonshotai/Kimi-Linear-48B-A3B-Instruct"], ["kolibri1", "Reachable", "", 52602, "Aleph-Alpha/Kolibri-1"], ["kosmos-2", "Unreachable", "forward fails after shrink", 24555, "microsoft/kosmos-2-patch14-224"], ["kosmos-2.5", "Reachable", "", 7720, "microsoft/kosmos-2.5"], ["kosmos_2_5_text_model", "Unreachable", "component config only", 14, "Fireblossom/kosmos-2.5-4bit-text"], ["kosmos_2_5_vision_model", "Unreachable", "component config only", 15, "Fireblossom/kosmos-2.5-4bit-vision"], ["kosmos_2_text_model", "Unreachable", "component config only", 0, ""], ["kosmos_2_vision_model", "Unreachable", "component config only", 0, ""], ["kyutai_speech_to_text", "Unreachable", "forward fails after shrink", 8225, "kyutai/stt-2.6b-en-trfs"], ["laguna", "Done", "", 3313579, "poolside/Laguna-XS.2"], ["lasr_ctc", "Unreachable", "forward fails after shrink", 13038, "google/medasr"], ["lasr_encoder", "Unreachable", "no usable checkpoint", 40, "hf-tiny-v2/tiny-random-LasrEncoder"], ["layoutlm", "Done", "", 518133, "impira/layoutlm-document-qa"], ["layoutlmv2", "Unreachable", "model build fails", 440458, "microsoft/layoutxlm-base"], ["layoutlmv3", "Unreachable", "forward fails after shrink", 970269, "microsoft/layoutlmv3-base"], ["layoutxlm", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/layoutxlm-base is another type)", 0, ""], ["led", "Unreachable", "model build fails", 33694, "allenai/led-base-16384"], ["levit", "Reachable", "", 19387, "optimum-intel-internal-testing/tiny-random-LevitModel"], ["lfm2", "Done", "", 1255733, "LiquidAI/LFM2.5-230M-Base"], ["lfm2_moe", "Done", "", 688657, "lmstudio-community/LFM2-24B-A2B-MLX-4bit"], ["lfm2_vl", "Done", "", 199290, "LiquidAI/LFM2.5-VL-450M"], ["lightglue", "Unreachable", "forward fails after shrink", 22972, "ETH-CVG/lightglue_superpoint"], ["lighton_ocr", "Reachable", "", 606, "onnx-community/LightOnOCR-2-1B-ONNX"], ["lilt", "Unreachable", "forward fails after shrink", 81714, "SCUT-DLVCLab/lilt-roberta-en-base"], ["llama", "Done", "", 50186457, "meta-llama/Llama-3.1-8B-Instruct"], ["llama4", "Done", "", 373138, "meta-llama/Llama-4-Scout-17B-16E-Instruct"], ["llama4_text", "Reachable", "", 28667, "trl-internal-testing/tiny-Llama4ForCausalLM"], ["llama4_vision_model", "Unreachable", "checkpoint gated", 0, "meta-llama/Llama-4-Scout-17B-16E"], ["llava", "Done", "", 1766490, "llava-hf/llava-1.5-7b-hf"], ["llava_next", "Done", "", 667783, "llava-hf/llava-v1.6-mistral-7b-hf"], ["llava_next_video", "Done", "", 53334, "llava-hf/LLaVA-NeXT-Video-7B-hf"], ["llava_onevision", "Done", "", 262658, "llava-hf/llava-onevision-qwen2-0.5b-ov-hf"], ["longcat_flash", "Unreachable", "forward fails after shrink", 1566, "wangbo981016/longcat2.0-nvfp4"], ["longformer", "Done", "", 1478599, "allenai/longformer-base-4096"], ["longt5", "Done", "", 243258, "pszemraj/long-t5-tglobal-base-sci-simplify"], ["luke", "Done", "", 321239, "studio-ousia/luke-large-finetuned-tacred"], ["lw_detr", "Reachable", "", 13080, "AnnaZhang/lwdetr_small_60e_coco"], ["lw_detr_vit", "Unreachable", "backward fails", 57, "hf-tiny-v2/tiny-random-LwDetrViTBackbone"], ["lxmert", "Unreachable", "forward fails after shrink", 4977, "unc-nlp/lxmert-base-uncased"], ["m2m_100", "Done", "", 2425200, "facebook/nllb-200-distilled-600M"], ["mamba", "Done", "", 138353, "state-spaces/mamba-130m-hf"], ["mamba2", "Unreachable", "shrink fails", 1401, "limloop/whiff-mamba2-50M-v0.1"], ["marian", "Done", "", 5061374, "Helsinki-NLP/opus-mt-nl-en"], ["markuplm", "Reachable", "", 10766, "microsoft/markuplm-base"], ["mask2former", "Done", "", 1060987, "facebook/mask2former-swin-large-ade-semantic"], ["maskformer", "Reachable", "", 16159, "facebook/maskformer-swin-base-ade"], ["maskformer-swin", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/swin-tiny-patch4-window7-224 is another type)", 0, ""], ["mbart", "Done", "", 579384, "ai4bharat/IndicBART"], ["megatron-bert", "Reachable", "", 21960, "KBLab/robust-swedish-sentiment-multiclass"], ["mellum", "Reachable", "", 51107, "JetBrains/Mellum2-12B-A2.5B-Instruct"], ["metaclip_2", "Reachable", "", 92623, "facebook/metaclip-2-worldwide-b16"], ["metaclip_2_text_model", "Unreachable", "component config only", 0, ""], ["metaclip_2_vision_model", "Unreachable", "component config only", 0, ""], ["mgp-str", "Unreachable", "forward fails after shrink", 13719, "alibaba-damo/mgp-str-base"], ["mimi", "Unreachable", "forward fails after shrink", 184266, "kyutai/mimi"], ["mimi_decoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["mimi_encoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["mimo_v2_flash", "Reachable", "", 49711, "XiaomiMiMo/MiMo-V2-Flash"], ["minicpm3", "Done", "", 17559, "openbmb/MiniCPM3-4B"], ["minicpm_ocr", "Reachable", "", 2508, "openbmb/MiniCPM-OCR-1.0"], ["minicpm_ocr_vision", "Unreachable", "forward fails after shrink", 0, ""], ["minicpmv4_6", "Done", "", 372489, "openbmb/MiniCPM-V-4.6"], ["minicpmv4_6_vision", "Unreachable", "component config only", 0, ""], ["minicpmv4_7", "Unreachable", "no usable checkpoint", 775231, "openbmb/MiniCPM-V-4.7-35B-A3B"], ["minicpmv4_7_vision", "Unreachable", "no usable checkpoint", 28, "hf-tiny-v2/tiny-random-MiniCPMV4_7VisionModel"], ["minimax", "Reachable", "", 19547, "MiniMaxAI/MiniMax-Text-01-hf"], ["minimax_m2", "Done", "", 1886356, "MiniMaxAI/MiniMax-M2.5"], ["minimax_m3_vl", "Done", "", 384244, "MiniMaxAI/MiniMax-M3"], ["minimax_m3_vl_text", "Reachable", "", 314, "ProprietaryLegal/minimax-m3-legal-reap-0.25"], ["minimax_m3_vl_vision", "Unreachable", "component config only", 30, "immanuelpeter/MiniMax-M3-Vision"], ["ministral", "Unreachable", "no usable checkpoint", 1857, "saidutta69/Ministral-8B-Instruct-heretic"], ["ministral3", "Done", "", 500650, "nvidia/Nemotron-3-Embed-1B-BF16"], ["mistral", "Done", "", 10522940, "mistralai/Mistral-7B-Instruct-v0.2"], ["mistral3", "Done", "", 2910198, "lightonai/LightOnOCR-2-1B"], ["mistral4", "Reachable", "", 971, "onnx-internal-testing/tiny-random-Mistral4ForCausalLM"], ["mixtral", "Done", "", 1221154, "mistralai/Mixtral-8x7B-Instruct-v0.1"], ["mlcd", "Reachable", "", 6623, "DeepGlint-AI/mlcd-vit-bigG-patch14-336"], ["mlcd_vision_model", "Unreachable", "no usable checkpoint", 107, "hf-tiny-v2/tiny-random-MLCDVisionModel"], ["mllama", "Done", "", 128886, "meta-llama/Llama-3.2-11B-Vision-Instruct"], ["mllama_text_model", "Unreachable", "component config only", 637, "hf-tiny-v2/tiny-random-MllamaTextModel"], ["mllama_vision_model", "Unreachable", "component config only", 65, "hf-tiny-v2/tiny-random-MllamaVisionModel"], ["mm-grounding-dino", "Done", "", 77467, "openmmlab-community/mm_grounding_dino_large_all"], ["mobilebert", "Unreachable", "model build fails", 378729, "google/mobilebert-uncased"], ["mobilenet_v1", "Reachable", "", 22769, "optimum-intel-internal-testing/mobilenet_v1_0.75_192"], ["mobilenet_v2", "Done", "", 116493, "google/mobilenet_v2_1.0_224"], ["mobilevit", "Done", "", 62083, "apple/mobilevit-xx-small"], ["mobilevitv2", "Reachable", "", 4054, "apple/mobilevitv2-1.0-imagenet1k-256"], ["modernbert", "Done", "", 22009085, "ibm-granite/granite-embedding-small-english-r2"], ["modernbert-decoder", "Reachable", "", 13311, "blab-jhu/test-32m-dec"], ["modernvbert", "Reachable", "", 477, "naver/v-splade-quality"], ["molmo2", "Unreachable", "shrink fails", 124222, "allenai/Molmo2-8B"], ["molmo2_adapter", "No checkpoint", "no Hub checkpoint has this model_type (documented allenai/Molmo2-8B is another type)", 0, ""], ["molmo2_text", "Reachable", "", 0, ""], ["molmo2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented allenai/Molmo2-8B is another type)", 0, ""], ["moonshine", "Unreachable", "forward fails after shrink", 120877, "moonshine-ai/moonshine-tiny"], ["moonshine_streaming", "Unreachable", "forward fails after shrink", 167472, "moonshine-ai/moonshine-streaming-tiny"], ["moonshine_streaming_encoder", "Unreachable", "component config only", 0, ""], ["moshi", "Reachable", "", 8294, "kyutai/personaplex-rl-seamless"], ["moshi_depth", "Unreachable", "component config only", 0, ""], ["moss_transcribe_diarize", "Done", "", 139891, "OpenMOSS-Team/MOSS-Transcribe-Diarize"], ["moss_transcribe_diarize_audio", "Reachable", "", 0, ""], ["mpnet", "Done", "", 23509384, "sentence-transformers/all-mpnet-base-v2"], ["mpt", "Reachable", "", 84471, "echarlaix/tiny-mpt-random-remote-code"], ["mra", "Reachable", "", 3305, "uw-madison/mra-base-512-4"], ["mt5", "Done", "", 725133, "google/mt5-small"], ["muse_glimmer", "Done", "", 688700, "RadixArk/Muse-Glimmer-NVFP4"], ["muse_glimmer_assistant", "Unreachable", "synthetic inputs rejected", 220523, "meta-models/Muse-Glimmer-30B-assistant"], ["muse_glimmer_text", "Reachable", "", 210, "OpensourceWTF/Muse-Glimmer-30B-MLX-q4"], ["muse_glimmer_vision", "Unreachable", "forward fails after shrink", 54, "immanuelpeter/Muse-Glimmer-Vision"], ["muse_spark", "Reachable", "", 48269, "meta-models/Muse-Spark-Open-816B-A42B-NVFP4"], ["muse_spark_audio", "Unreachable", "checkpoint gated", 0, "meta-models/Muse-Spark-Open-816B-A42B"], ["muse_spark_text", "Unreachable", "checkpoint gated", 0, "meta-models/Muse-Spark-Open-816B-A42B"], ["muse_spark_vision", "Unreachable", "checkpoint gated", 0, "meta-models/Muse-Spark-Open-816B-A42B"], ["musicflamingo", "Reachable", "", 18191, "nvidia/music-flamingo-2601-hf"], ["musicgen", "Unreachable", "model build fails", 2150786, "facebook/musicgen-medium"], ["musicgen_decoder", "Unreachable", "component config only", 13, "hf-tiny-v2/tiny-random-MusicgenForCausalLM"], ["musicgen_melody", "Unreachable", "model build fails", 12068, "facebook/musicgen-melody"], ["musicgen_melody_decoder", "Unreachable", "component config only", 39, "ylacombe/musicgen-melody-bella-ciao"], ["mvp", "Reachable", "", 4701, "RUCAIBox/mvp"], ["nanochat", "Reachable", "", 5316, "l0ulan/chinese-babylm-nanochat-d22-step10000"], ["nemo_speechlm", "Unreachable", "shrink fails", 5799, "nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B"], ["nemotron", "Reachable", "", 48968, "nvidia/Nemotron-Mini-4B-Instruct"], ["nemotron3_5_asr", "Unreachable", "forward fails after shrink", 842520, "nvidia/nemotron-3.5-asr-streaming-0.6b"], ["nemotron3_5_transcribe_asr_audio", "Unreachable", "forward fails after shrink", 0, ""], ["nemotron3_5_transcribe_audio", "Unreachable", "shrink fails", 0, ""], ["nemotron3_5_transcribe_word_timestamp", "Unreachable", "synthetic inputs rejected", 0, ""], ["nemotron3_diarization", "Unreachable", "shrink fails", 146475, "nvidia/Nemotron-3-Diarization"], ["nemotron3_diarization_audio", "Unreachable", "forward fails after shrink", 34, "hf-tiny-v2/tiny-random-Nemotron3DiarizationAudioModel"], ["nemotron_asr_streaming", "Unreachable", "forward fails after shrink", 103140, "nvidia/nemotron-speech-streaming-en-0.6b"], ["nemotron_asr_streaming_encoder", "Unreachable", "forward fails after shrink", 6, "hf-tiny-v2/tiny-random-NemotronAsrStreamingEncoder"], ["nemotron_h", "Done", "", 12146358, "nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16"], ["nemotron_h_omni", "Unreachable", "no usable checkpoint", 113462, "nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16"], ["neomme", "Reachable", "", 6699, "Hcompany/NeoMME-260M"], ["neucodec", "Unreachable", "forward fails after shrink", 48553, "neuphonic/neucodec"], ["nllb-moe", "Reachable", "", 3303, "facebook/nllb-moe-54b"], ["nomic_bert", "Done", "", 14371774, "nomic-ai/nomic-embed-text-v1.5"], ["nougat", "Unreachable", "config load fails", 8, "zphilip48/nougat-middle-cn"], ["nystromformer", "Reachable", "", 5529, "uw-madison/nystromformer-512"], ["olmo", "Done", "", 125508, "optimum-intel-internal-testing/tiny-random-olmo-hf"], ["olmo2", "Done", "", 442211, "allenai/OLMo-2-0425-1B"], ["olmo3", "Done", "", 1055647, "allenai/Olmo-3-7B-Think"], ["olmo_hybrid", "Reachable", "", 19286, "allenai/Olmo-Hybrid-7B"], ["olmoe", "Done", "", 132339, "allenai/OLMoE-1B-7B-0924"], ["omdet-turbo", "Unreachable", "config load fails", 35542, "omlab/omdet-turbo-swin-tiny-hf"], ["oneformer", "Unreachable", "synthetic inputs rejected", 390959, "shi-labs/oneformer_cityscapes_swin_large"], ["openai-gpt", "Reachable", "", 85250, "openai-community/openai-gpt"], ["openai_privacy_filter", "Done", "", 634682, "OpenMed/privacy-filter-nemotron-v2"], ["opt", "Done", "", 5616003, "facebook/opt-125m"], ["ovis2", "Reachable", "", 6898, "thisisiron/Ovis2-1B-hf"], ["ovis2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented thisisiron/Ovis2-1B-hf is another type)", 0, ""], ["owlv2", "Done", "", 1501028, "google/owlv2-base-patch16-ensemble"], ["owlv2_text_model", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["owlv2_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["owlvit", "Done", "", 220839, "google/owlvit-base-patch32"], ["owlvit_text_model", "Unreachable", "component config only", 0, ""], ["owlvit_vision_model", "Unreachable", "component config only", 0, ""], ["paddleocr_vl", "Done", "", 227189, "BDRC/tibetan-ocr"], ["paddleocr_vl_text", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PaddleOCR-VL is another type)", 0, ""], ["paddleocr_vl_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PaddleOCR-VL is another type)", 0, ""], ["paligemma", "Done", "", 490686, "google/paligemma-3b-pt-224"], ["parakeet_ctc", "Unreachable", "forward fails after shrink", 105992, "nvidia/parakeet-ctc-1.1b"], ["parakeet_encoder", "Unreachable", "forward fails after shrink", 125, "nithinraok/parakeet-tdt-v3-encoder"], ["parakeet_rnnt", "Unreachable", "forward fails after shrink", 40081, "nvidia/parakeet-rnnt-0.6b"], ["parakeet_tdt", "Unreachable", "forward fails after shrink", 694911, "nvidia/parakeet-tdt-0.6b-v3"], ["patchtsmixer", "Unreachable", "synthetic inputs rejected", 15970, "ibm-granite/granite-timeseries-patchtsmixer"], ["patchtst", "Unreachable", "synthetic inputs rejected", 28739, "GilpinLab/panda"], ["pe_audio", "Unreachable", "no usable checkpoint", 89572, "facebook/pe-a-frame-large"], ["pe_audio_encoder", "Unreachable", "forward fails after shrink", 31, "hf-tiny-v2/tiny-random-PeAudioEncoder"], ["pe_audio_video", "Unreachable", "forward fails after shrink", 21786, "facebook/pe-av-large"], ["pe_audio_video_encoder", "Unreachable", "forward fails after shrink", 21, "hf-tiny-v2/tiny-random-PeAudioVideoEncoder"], ["pe_video", "Unreachable", "no usable checkpoint", 25, "hf-tiny-v2/tiny-random-PeVideoModel"], ["pe_video_encoder", "Unreachable", "synthetic inputs rejected", 25, "hf-tiny-v2/tiny-random-PeVideoEncoder"], ["pegasus", "Done", "", 120276, "optimum-intel-internal-testing/tiny-random-pegasus"], ["pegasus_x", "Reachable", "", 6499, "google/pegasus-x-large"], ["perceiver", "Done", "", 41482, "optimum-intel-internal-testing/tiny-random-language_perceiver"], ["perception_lm", "Reachable", "", 1028, "facebook/Perception-LM-8B"], ["persimmon", "Reachable", "", 50277, "optimum-intel-internal-testing/tiny-random-PersimmonForCausalLM"], ["phi", "Done", "", 575882, "microsoft/phi-2"], ["phi3", "Done", "", 2440907, "microsoft/Phi-3.5-mini-instruct"], ["phi4_multimodal", "Done", "", 549, "huggingface-mic/Phi-4-multimodal-instruct-mic"], ["phi4_multimodal_audio", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/Phi-4-multimodal-instruct is another type)", 0, ""], ["phi4_multimodal_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/Phi-4-multimodal-instruct is another type)", 0, ""], ["phimoe", "Done", "", 169903, "optimum-intel-internal-testing/phi-3.5-moe-tiny-random"], ["pi0", "Unreachable", "synthetic inputs rejected", 181, "yinchenghust/openpi_libero_all"], ["pix2struct", "Unreachable", "synthetic inputs rejected", 116238, "optimum-intel-internal-testing/pix2struct-tiny-random"], ["pix2struct_text_model", "Unreachable", "component config only", 0, ""], ["pix2struct_vision_model", "Unreachable", "component config only", 22, "UiPath/pix2struct-vision-base"], ["pixio", "Reachable", "", 163, "LiheYoung/pixio-vith16"], ["pixtral", "Reachable", "", 940, "mlx-community/pixtral-12b-4bit"], ["plbart", "Reachable", "", 11136, "uclanlp/plbart-base"], ["pocket_tts", "Unreachable", "shrink fails", 2949, "mlx-community/pocket-tts"], ["pocket_tts_audio_tokenizer", "Unreachable", "forward fails after shrink", 471, "gradium/phonon-mimi"], ["pocket_tts_audio_tokenizer_decoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["pocket_tts_audio_tokenizer_encoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["poolformer", "Reachable", "", 22043, "optimum-intel-internal-testing/tiny-random-PoolFormerModel"], ["pop2piano", "Reachable", "", 105, "hf-tiny-v2/tiny-random-Pop2PianoForConditionalGeneration"], ["pp_chart2table", "Reachable", "", 272, "PaddlePaddle/PP-Chart2Table_safetensors"], ["pp_chart2table_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["pp_doclayout_v2", "Unreachable", "forward fails after shrink", 585, "PaddlePaddle/PP-DocLayoutV2_safetensors"], ["pp_doclayout_v3", "Unreachable", "forward fails after shrink", 932166, "PaddlePaddle/PP-DocLayoutV3_safetensors"], ["pp_doclayout_v4", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PP-DocLayoutV4_safetensors does not exist)", 0, ""], ["pp_formulanet", "Unreachable", "forward fails after shrink", 3829, "PaddlePaddle/PP-FormulaNet_plus-L_safetensors"], ["pp_formulanet_text", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PP-FormulaNet_plus-L_safetensors is another type)", 0, ""], ["pp_formulanet_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PP-FormulaNet_plus-L_safetensors is another type)", 0, ""], ["pp_lcnet", "Reachable", "", 865, "PaddlePaddle/PP-LCNet_x1_0_doc_ori_safetensors"], ["pp_lcnet_v3", "Unreachable", "no usable checkpoint", 112, "hf-tiny-v2/tiny-random-PPLCNetV3Backbone"], ["pp_lcnet_v4", "Unreachable", "no usable checkpoint", 21, "hf-tiny-v2/tiny-random-PPLCNetV4Backbone"], ["pp_ocrv5_mobile_det", "Reachable", "", 530, "PaddlePaddle/PP-OCRv5_mobile_det_safetensors"], ["pp_ocrv5_mobile_rec", "Reachable", "", 3902, "PaddlePaddle/latin_PP-OCRv5_mobile_rec_safetensors"], ["pp_ocrv5_server_det", "Reachable", "", 509, "PaddlePaddle/PP-OCRv5_server_det_safetensors"], ["pp_ocrv5_server_rec", "Reachable", "", 585, "PaddlePaddle/PP-OCRv5_server_rec_safetensors"], ["pp_ocrv6_medium_det", "Reachable", "", 2694, "PaddlePaddle/PP-OCRv6_medium_det_safetensors"], ["pp_ocrv6_small_det", "Reachable", "", 1617, "PaddlePaddle/PP-OCRv6_small_det_safetensors"], ["pp_ocrv6_small_rec", "Reachable", "", 3559, "PaddlePaddle/PP-OCRv6_medium_rec_safetensors"], ["pp_ocrv6_tiny_rec", "Reachable", "", 693, "PaddlePaddle/PP-OCRv6_tiny_rec_safetensors"], ["prompt_depth_anything", "Unreachable", "forward fails after shrink", 1530, "depth-anything/prompt-depth-anything-vits-hf"], ["prophetnet", "Reachable", "", 7258, "microsoft/prophetnet-large-uncased"], ["pvt", "Done", "", 8712, "Xrenya/pvt-tiny-224"], ["pvt_v2", "Reachable", "", 9824, "OpenGVLab/pvt_v2_b0"], ["qianfan_ocr", "Reachable", "", 38183, "baidu/Qianfan-OCR"], ["qianfan_ocr_vision", "Reachable", "", 77, "hf-tiny-v2/tiny-random-QianfanOCRVisionModel"], ["qwen2", "Done", "", 52739906, "Qwen/Qwen2.5-7B-Instruct"], ["qwen2_5_omni", "Done", "", 456813, "Qwen/Qwen2.5-Omni-7B"], ["qwen2_5_omni_audio_encoder", "Unreachable", "component config only", 23, "FlexiSLM/Qwen2_5-Omni-Audio_Encoder"], ["qwen2_5_omni_bigvgan", "Unreachable", "component config only", 24, "Ryan241005/scail-2"], ["qwen2_5_omni_dit", "Unreachable", "component config only", 0, ""], ["qwen2_5_omni_talker", "Unreachable", "component config only", 0, ""], ["qwen2_5_omni_text", "Unreachable", "component config only", 59, "AlexHung29629/Qwen-2.5-Omni-ThinkerTextModel"], ["qwen2_5_omni_thinker", "Done", "", 128795, "laion/voiceclap-large-v2"], ["qwen2_5_omni_token2wav", "Unreachable", "component config only", 0, ""], ["qwen2_5_omni_vision_encoder", "Unreachable", "component config only", 0, ""], ["qwen2_5_vl", "Done", "", 15195814, "Qwen/Qwen2.5-VL-7B-Instruct"], ["qwen2_5_vl_text", "Reachable", "", 83, "hf-tiny-v2/tiny-random-Qwen2_5_VLTextModel"], ["qwen2_5_vl_vision", "Unreachable", "no usable checkpoint", 22, "hf-tiny-v2/tiny-random-Qwen2_5_VisionTransformerPretrainedModel"], ["qwen2_audio", "Done", "", 199606, "Qwen/Qwen2-Audio-7B-Instruct"], ["qwen2_audio_encoder", "Reachable", "", 200, "Atotti/qwen2-audio-encoder"], ["qwen2_moe", "Done", "", 194932, "Qwen/Qwen1.5-MoE-A2.7B"], ["qwen2_vl", "Done", "", 3757282, "Qwen/Qwen2-VL-7B-Instruct-AWQ"], ["qwen2_vl_text", "Reachable", "", 72, "hf-tiny-v2/tiny-random-Qwen2VLTextModel"], ["qwen2_vl_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented Qwen/Qwen2-VL-7B-Instruct is another type)", 0, ""], ["qwen3", "Done", "", 108746675, "Qwen/Qwen3-0.6B"], ["qwen3_5", "Done", "", 92274452, "Qwen/Qwen3.8-27B"], ["qwen3_5_moe", "Done", "", 34351908, "Qwen/Qwen3.6-35B-A3B"], ["qwen3_5_moe_text", "Done", "", 717113, "Bahushruth/Qwen3.6-35B-A3B-abliterated-v4"], ["qwen3_5_moe_vision", "Unreachable", "no usable checkpoint", 16, "hf-tiny-v2/tiny-random-Qwen3_5MoeVisionModel"], ["qwen3_5_text", "Done", "", 2017293, "autotrust/JEV-9B"], ["qwen3_5_vision", "Unreachable", "forward fails after shrink", 135, "immanuelpeter/Qwen3.8-27B-Vision"], ["qwen3_asr", "Done", "", 4105092, "Qwen/Qwen3-ASR-1.7B"], ["qwen3_asr_encoder", "Unreachable", "forward fails after shrink", 17, "hf-tiny-v2/tiny-random-Qwen3ASREncoder"], ["qwen3_moe", "Done", "", 8500460, "Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8"], ["qwen3_next", "Done", "", 2235316, "Qwen/Qwen3-Coder-Next-FP8"], ["qwen3_omni_moe", "Unreachable", "forward fails after shrink", 1249447, "Qwen/Qwen3-Omni-30B-A3B-Instruct"], ["qwen3_omni_moe_audio_encoder", "Unreachable", "component config only", 1968, "Atotti/Qwen3-Omni-AudioTransformer"], ["qwen3_omni_moe_talker_code_predictor", "Unreachable", "component config only", 0, ""], ["qwen3_omni_moe_talker_text", "No checkpoint", "no Hub checkpoint has this model_type (documented Qwen/Qwen3-Omni-30B-A3B-Instruct is another type)", 0, ""], ["qwen3_omni_moe_text", "Unreachable", "component config only", 0, ""], ["qwen3_omni_moe_thinker", "Reachable", "", 298, "hf-tiny-v2/tiny-random-Qwen3OmniMoeThinkerForConditionalGeneration"], ["qwen3_omni_moe_vision_encoder", "Unreachable", "component config only", 0, ""], ["qwen3_vl", "Done", "", 19337885, "Qwen/Qwen3-VL-8B-Instruct"], ["qwen3_vl_moe", "Done", "", 1686056, "Qwen/Qwen3-VL-30B-A3B-Instruct"], ["qwen3_vl_moe_text", "Reachable", "", 63, "hf-tiny-v2/tiny-random-Qwen3VLMoeTextModel"], ["qwen3_vl_moe_vision", "Unreachable", "no usable checkpoint", 54, "hf-tiny-v2/tiny-random-Qwen3VLMoeVisionModel"], ["qwen3_vl_text", "Reachable", "", 587, "DuoNeural/Cosmos3-Nano-GPTQ-4bit"], ["qwen3_vl_vision", "Unreachable", "forward fails after shrink", 221, "madrisight/madrimed1.2-vision-embed"], ["qwen4_exp", "Done", "", 5763811, "Qwen/Qwen3.8-Flash-Next"], ["qwen4_exp_text", "Reachable", "", 17917, "Jab1718/qwen3.8-flash-coder-44gb-selective-int8"], ["qwen4_exp_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented Qwen/Qwen4-Exp does not exist)", 0, ""], ["radio", "Unreachable", "no usable checkpoint", 3, "hf-tiny-v2/tiny-random-RadioModel"], ["rag", "Unreachable", "no usable checkpoint", 3924, "facebook/rag-token-nq"], ["recurrent_gemma", "Done", "", 4729, "google/recurrentgemma-2b-it"], ["reformer", "Unreachable", "model build fails", 17043, "google/reformer-crime-and-punishment"], ["regnet", "Reachable", "", 7237, "facebook/regnet-y-040"], ["rembert", "Reachable", "", 11890, "google/rembert"], ["resnet", "Done", "", 1174379, "microsoft/resnet-18"], ["rf_detr", "Unreachable", "forward fails after shrink", 28618, "Roboflow/rf-detr-base"], ["rf_detr_dinov2", "Unreachable", "forward fails after shrink", 70, "hf-tiny-v2/tiny-random-RfDetrDinov2Backbone"], ["roberta", "Done", "", 46597117, "FacebookAI/roberta-large"], ["roberta-prelayernorm", "Reachable", "", 4655, "andreasmadsen/efficient_mlm_m0.40"], ["roc_bert", "Reachable", "", 12437, "weiweishi/roc-bert-base-zh"], ["roformer", "Reachable", "", 45159, "optimum-intel-internal-testing/tiny-random-roformer"], ["rt_detr", "Done", "", 1687337, "PekingU/rtdetr_r101vd_coco_o365"], ["rt_detr_resnet", "Unreachable", "no usable checkpoint", 54, "hf-tiny-v2/tiny-random-RTDetrResNetBackbone"], ["rt_detr_v2", "Done", "forward fails after shrink", 2526763, "docling-project/docling-layout-heron"], ["rwkv", "Reachable", "", 14762, "RWKV/rwkv-4-169m-pile"], ["sam", "Unreachable", "forward fails after shrink", 976537, "facebook/sam-vit-base"], ["sam2", "Done", "", 19375, "onnx-community/sam2.1-hiera-tiny-ONNX"], ["sam2_hiera_det_model", "Unreachable", "forward fails after shrink", 59, "hf-tiny-v2/tiny-random-Sam2HieraDetModel"], ["sam2_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2.1-hiera-tiny is another type)", 0, ""], ["sam2_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2.1-hiera-tiny is another type)", 0, ""], ["sam2_video", "Unreachable", "forward fails after shrink", 701203, "facebook/sam2.1-hiera-large"], ["sam2_video_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2_video.1-hiera-tiny does not exist)", 0, ""], ["sam2_video_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2_video.1-hiera-tiny does not exist)", 0, ""], ["sam2_vision_model", "Unreachable", "forward fails after shrink", 57, "hf-tiny-v2/tiny-random-Sam2VisionModel"], ["sam3", "Unreachable", "forward fails after shrink", 3530, "unsloth/sam3.1"], ["sam3_detr_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_detr_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_geometry_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_lite_text", "Unreachable", "forward fails after shrink", 14775, "vil-uob/sam3-litetext-s0"], ["sam3_lite_text_detr_decoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_detr_encoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_geometry_encoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_mask_decoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_text_model", "Unreachable", "forward fails after shrink", 39, "hf-tiny-v2/tiny-random-Sam3LiteTextTextModel"], ["sam3_mask_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker", "Unreachable", "no usable checkpoint", 1300, "onnx-community/sam3-tracker-ONNX"], ["sam3_tracker_mask_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_prompt_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_video", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_video_mask_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_video_prompt_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_video", "Unreachable", "forward fails after shrink", 2087329, "facebook/sam3"], ["sam3_vision_model", "Unreachable", "no usable checkpoint", 38, "hf-tiny-v2/tiny-random-Sam3VisionModel"], ["sam3_vit_model", "Unreachable", "no usable checkpoint", 37, "hf-tiny-v2/tiny-random-Sam3ViTModel"], ["sam_hq", "Unreachable", "forward fails after shrink", 16860, "syscv-community/sam-hq-vit-base"], ["sam_hq_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented syscv-community/sam-hq-vit-base is another type)", 0, ""], ["sam_hq_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented syscv-community/sam-hq-vit-base is another type)", 0, ""], ["sam_hq_vision_model", "Reachable", "", 138, "hf-tiny-v2/tiny-random-SamHQVisionModel"], ["sam_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["sam_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["sam_vision_model", "Reachable", "", 138, "hf-tiny-v2/tiny-random-SamVisionModel"], ["sapiens2", "Unreachable", "forward fails after shrink", 103690, "facebook/sapiens2-pose-1b"], ["sapiens2_head", "Unreachable", "component config only", 0, ""], ["seamless_m4t", "Reachable", "", 13797, "facebook/hf-seamless-m4t-medium"], ["seamless_m4t_v2", "Done", "", 316927, "facebook/seamless-m4t-v2-large"], ["seed_oss", "Done", "", 128943, "ByteDance-Seed/Seed-OSS-36B-Instruct"], ["segformer", "Done", "", 2291917, "mattmdjaga/segformer_b2_clothes"], ["seggpt", "Unreachable", "forward fails after shrink", 21384, "BAAI/seggpt-vit-large"], ["sew", "Reachable", "", 15680, "optimum-intel-internal-testing/tiny-random-SEWModel"], ["sew-d", "Reachable", "", 18152, "optimum-intel-internal-testing/sew-d-tiny-100k-ft-ls100h"], ["shieldgemma2", "Reachable", "", 6627, "google/shieldgemma-2-4b-it"], ["siglip", "Done", "", 18018379, "prithivMLmods/Gender-Classifier-Mini"], ["siglip2", "Unreachable", "forward fails after shrink", 1622065, "google/siglip2-base-patch16-naflex"], ["siglip2_text_model", "Unreachable", "component config only", 56, "sortify-ai/siglip2-so400m-patch14-384-text"], ["siglip2_vision_model", "Unreachable", "forward fails after shrink", 58, "hf-tiny-v2/tiny-random-Siglip2VisionModel"], ["siglip_text_model", "Unreachable", "component config only", 19, "carsonpoole/binary-siglip-text"], ["siglip_vision_model", "Reachable", "", 1842, "zchuning/pi05-siglip-vision-tower"], ["slanet", "Reachable", "", 6957, "PaddlePaddle/SLANet_plus_safetensors"], ["slanext", "Reachable", "", 212, "PaddlePaddle/SLANeXt_wired_safetensors"], ["slanext_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/SLANeXt_wired_safetensors is another type)", 0, ""], ["smollm3", "Done", "", 1135357, "HuggingFaceTB/SmolLM3-3B"], ["smolvlm", "Done", "", 1742362, "HuggingFaceTB/SmolVLM2-500M-Video-Instruct"], ["smolvlm_vision", "Unreachable", "forward fails after shrink", 5, "hf-tiny-v2/tiny-random-SmolVLMVisionTransformer"], ["solar_open", "Reachable", "", 32622, "upstage/Solar-Open-100B"], ["speech-encoder-decoder", "Unreachable", "synthetic inputs rejected", 5244, "hf-tiny-v2/tiny-random-SpeechEncoderDecoderModel-wav2vec2-bert"], ["speech_to_text", "Unreachable", "forward fails after shrink", 23544, "facebook/s2t-small-librispeech-asr"], ["speecht5", "Done", "", 194296, "optimum-intel-internal-testing/tiny-random-SpeechT5ForTextToSpeech"], ["speecht5_hifigan", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/speecht5_asr is another type)", 0, ""], ["splinter", "Reachable", "", 3741, "tau/splinter-base"], ["squeezebert", "Unreachable", "model build fails", 33330, "optimum-intel-internal-testing/tiny-random-squeezebert"], ["stablelm", "Done", "", 169690, "optimum-intel-internal-testing/tiny-random-StableLmForCausalLM"], ["starcoder2", "Done", "", 152451, "optimum-intel-internal-testing/tiny-random-Starcoder2ForCausalLM"], ["step3p5", "Unreachable", "shrink fails", 266885, "stepfun-ai/Step-3.5-Flash"], ["step3p5_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented stepfun-ai/Step-3.7-Flash is another type)", 0, ""], ["step3p7", "Unreachable", "shrink fails", 26135, "stepfun-ai/Step-3.7-Flash"], ["superglue", "Unreachable", "forward fails after shrink", 8964, "magic-leap-community/superglue_indoor"], ["superpoint", "Unreachable", "backward fails", 9952, "magic-leap-community/superpoint"], ["swiftformer", "Reachable", "", 7504, "MBZUAI/swiftformer-xs"], ["swin", "Done", "", 238869, "Organika/sdxl-detector"], ["swin2sr", "Reachable", "", 56057, "caidas/swin2SR-classical-sr-x2-64"], ["swinv2", "Done", "", 434756, "microsoft/swinv2-tiny-patch4-window16-256"], ["switch_transformers", "Reachable", "", 10063, "google/switch-base-8"], ["t5", "Done", "", 94678669, "google-t5/t5-small"], ["t5_gemma_module", "Unreachable", "component config only", 70, "RE-N-Y/t5gemma-2b-2b-ul2-it"], ["t5gemma", "Unreachable", "no usable checkpoint", 176067, "google/t5gemma-2b-2b-ul2-it"], ["t5gemma2", "Unreachable", "forward fails after shrink", 114374, "google/t5gemma-2-270m-270m"], ["t5gemma2_decoder", "Unreachable", "component config only", 0, ""], ["t5gemma2_encoder", "Reachable", "", 149, "hf-tiny-v2/tiny-random-T5Gemma2Encoder"], ["t5gemma2_text", "Unreachable", "component config only", 30, "amanchoudhri/t5-gemma-2-270m-text-encoder"], ["table-transformer", "Done", "", 1872765, "microsoft/table-transformer-structure-recognition"], ["tapas", "Reachable", "", 37821, "google/tapas-base-finetuned-wtq"], ["textnet", "Unreachable", "forward fails after shrink", 7475, "czczup/textnet-base"], ["time_series_transformer", "Unreachable", "synthetic inputs rejected", 4534, "huggingface/time-series-transformer-tourism-monthly"], ["timesfm", "Done", "", 1854129, "google/timesfm-2.5-200m-pytorch"], ["timesfm2_5", "Unreachable", "synthetic inputs rejected", 37239, "google/timesfm-2.5-200m-transformers"], ["timesformer", "Unreachable", "forward fails after shrink", 25737, "facebook/timesformer-base-finetuned-k400"], ["timm_backbone", "Unreachable", "no usable checkpoint", 133, "hf-tiny-v2/tiny-random-TimmBackbone"], ["timm_wrapper", "Unreachable", "forward fails after shrink", 383060, "Freepik/nsfw_image_detector"], ["tipsv2", "Done", "", 289583, "google/tipsv2-so400m14"], ["tipsv2_dpt", "Unreachable", "forward fails after shrink", 8014, "google/tipsv2-b14-dpt"], ["tipsv2_text_model", "Reachable", "", 4, "hf-tiny-v2/tiny-random-Tipsv2TextModel"], ["tipsv2_vision_model", "Reachable", "", 8, "hf-tiny-v2/tiny-random-Tipsv2VisionBackbone"], ["trocr", "Reachable", "", 53, "sasmithasarasi/trocr-sinhala-handwriting"], ["tvp", "Unreachable", "forward fails after shrink", 13994, "Intel/tvp-base"], ["udop", "Unreachable", "forward fails after shrink", 9051, "microsoft/udop-large"], ["ultrabert", "Reachable", "", 29547, "dmis-lab/UltraBERT-380M"], ["ultravbert", "Reachable", "", 1133, "athrael-soju/ColUltraVBERT"], ["umt5", "Done", "", 5305, "huggingface-mic/umt5-small-mic"], ["unispeech", "Reachable", "", 24130, "optimum-intel-internal-testing/tiny-random-unispeech"], ["unispeech-sat", "Reachable", "", 38482, "optimum-intel-internal-testing/tiny-random-UnispeechSatModel"], ["univnet", "Unreachable", "forward fails after shrink", 5412, "dg845/univnet-dev"], ["unlimited_ocr", "Done", "", 3535621, "baidu/Unlimited-OCR"], ["unlimited_ocr_sam_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["unlimited_ocr_text", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["unlimited_ocr_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["unlimited_ocr_vision_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["upernet", "Unreachable", "forward fails after shrink", 46968, "openmmlab/upernet-convnext-small"], ["uvdoc", "Reachable", "", 431, "PaddlePaddle/UVDoc_safetensors"], ["uvdoc_backbone", "Unreachable", "backward fails", 39, "hf-tiny-v2/tiny-random-UVDocBackbone"], ["vaultgemma", "Reachable", "", 2790, "google/vaultgemma-1b"], ["vibevoice", "Done", "", 1070414, "microsoft/VibeVoice-ASR"], ["vibevoice_acoustic_tokenizer", "Unreachable", "forward fails after shrink", 2151, "microsoft/VibeVoice-AcousticTokenizer"], ["vibevoice_acoustic_tokenizer_decoder", "Unreachable", "no usable checkpoint", 33, "hf-tiny-v2/tiny-random-VibeVoiceAcousticTokenizerDecoderModel"], ["vibevoice_acoustic_tokenizer_encoder", "Unreachable", "no usable checkpoint", 49, "hf-tiny-v2/tiny-random-VibeVoiceAcousticTokenizerEncoderModel"], ["vibevoice_asr", "Reachable", "", 59115, "microsoft/VibeVoice-ASR-HF"], ["vibevoice_diffusion", "No checkpoint", "no Hub checkpoint has this model_type (documented vibevoice/VibeVoice-1.5B-hf is another type)", 0, ""], ["video_llama_3", "Reachable", "", 11087, "lkhl/VideoLLaMA3-2B-Image-HF"], ["video_llama_3_vision", "Unreachable", "forward fails after shrink", 32, "hf-tiny-v2/tiny-random-VideoLlama3VisionModel"], ["video_llava", "Reachable", "", 10045, "LanguageBind/Video-LLaVA-7B-hf"], ["videomae", "Unreachable", "forward fails after shrink", 475037, "MCG-NJU/videomae-base"], ["videomt", "Unreachable", "synthetic inputs rejected", 5328, "tue-mps/videomt-dinov2-small-ytvis2019"], ["videoprism", "Unreachable", "synthetic inputs rejected", 545, "MHRDYN7/videoprism-lvt-base-f16r288"], ["videoprism_text_model", "No checkpoint", "no Hub checkpoint has this model_type (documented google/videoprism-lvt-base-f16r288 has no config.json)", 0, ""], ["videoprism_vision_model", "Unreachable", "synthetic inputs rejected", 191, "MHRDYN7/videoprism-base-f16r288"], ["vilt", "Unreachable", "forward fails after shrink", 89943, "dandelin/vilt-b32-finetuned-vqa"], ["vipllava", "Reachable", "", 7898, "llava-hf/vip-llava-7b-hf"], ["vision-encoder-decoder", "Unreachable", "forward fails after shrink", 2482977, "kha-white/manga-ocr-base"], ["vision-text-dual-encoder", "Unreachable", "model build fails", 3250, "hf-tiny-v2/tiny-random-VisionTextDualEncoderModel-vit-bert"], ["visual_bert", "Reachable", "", 5237, "uclanlp/visualbert-vqa-coco-pre"], ["vit", "Done", "", 23571533, "google/vit-base-patch16-224"], ["vit_mae", "Reachable", "", 43367, "facebook/vit-mae-base"], ["vit_msn", "Reachable", "", 1782, "facebook/vit-msn-small"], ["vitdet", "Unreachable", "no usable checkpoint", 214, "hf-tiny-v2/tiny-random-VitDetBackbone"], ["vitmatte", "Done", "", 383080, "hustvl/vitmatte-small-composition-1k"], ["vitpose", "Unreachable", "forward fails after shrink", 392468, "usyd-community/vitpose-base-simple"], ["vitpose_backbone", "Unreachable", "forward fails after shrink", 124, "hf-tiny-v2/tiny-random-VitPoseBackbone"], ["vits", "Done", "", 455247, "facebook/mms-tts-eng"], ["vivit", "Unreachable", "forward fails after shrink", 89874, "google/vivit-b-16x2"], ["vjepa2", "Unreachable", "synthetic inputs rejected", 426419, "facebook/vjepa2-vitl-fpc64-256"], ["voxtral", "Done", "", 319265, "mzbac/voxtral-mini-3b-4bit-mixed"], ["voxtral_encoder", "Reachable", "", 108, "hf-tiny-v2/tiny-random-VoxtralEncoder"], ["voxtral_realtime", "Unreachable", "forward fails after shrink", 1948459, "mistralai/Voxtral-Mini-4B-Realtime-2602"], ["voxtral_realtime_encoder", "Reachable", "", 80, "hf-tiny-v2/tiny-random-VoxtralRealtimeEncoder"], ["voxtral_realtime_text", "Unreachable", "model build fails", 10, "hf-tiny-v2/tiny-random-VoxtralRealtimeTextModel"], ["wav2vec2", "Done", "", 100317588, "jonatasgrosman/wav2vec2-large-xlsr-53-japanese"], ["wav2vec2-bert", "Unreachable", "model build fails", 1320062, "facebook/w2v-bert-2.0"], ["wav2vec2-conformer", "Unreachable", "model build fails", 160365, "facebook/wav2vec2-conformer-rope-large-960h-ft"], ["wavlm", "Done", "", 1990346, "microsoft/wavlm-large"], ["whisper", "Done", "", 22332636, "openai/whisper-large-v3-turbo"], ["xclip", "Unreachable", "forward fails after shrink", 453798, "microsoft/xclip-base-patch16-zero-shot"], ["xclip_text_model", "Unreachable", "component config only", 0, ""], ["xclip_vision_model", "Unreachable", "component config only", 0, ""], ["xcodec", "Unreachable", "config load fails", 14050, "hf-audio/xcodec-hubert-general"], ["xcodec2", "Unreachable", "forward fails after shrink", 9249, "HKUSTAudio/xcodec2-hf"], ["xglm", "Reachable", "", 64850, "optimum-intel-internal-testing/tiny-random-XGLMForCausalLM"], ["xlm", "Reachable", "", 19454, "FacebookAI/xlm-mlm-en-2048"], ["xlm-roberta", "Done", "", 104687870, "BAAI/bge-m3"], ["xlm-roberta-xl", "Done", "", 101085, "facebook/xlm-roberta-xl"], ["xlnet", "Unreachable", "shrink fails", 166190, "xlnet/xlnet-base-cased"], ["xlstm", "Unreachable", "config load fails", 4922, "NX-AI/xLSTM-7b"], ["xmod", "Unreachable", "forward fails after shrink", 29093, "facebook/xmod-base"], ["yolos", "Done", "", 859794, "hustvl/yolos-small"], ["yoso", "Reachable", "", 5281, "uw-madison/yoso-4096"], ["youtu", "Reachable", "", 12971, "tencent/Youtu-LLM-2B"], ["zamba", "Unreachable", "model build fails", 5935, "Zyphra/Zamba-7B-v1"], ["zamba2", "Unreachable", "forward fails after shrink", 70083, "Zyphra/Zamba2-1.2B-instruct"], ["zaya", "Reachable", "", 8978, "Zyphra/ZAYA1-8B"], ["zoedepth", "Unreachable", "forward fails after shrink", 48545, "Intel/zoedepth-nyu-kitti"]], "meta": {"generated": "2026-10-31", "fleet": "666d54fcc782e473ff276ec9626df0b0a3f0cafd", "transformers": "420dca92a4cf7d96dfbc1cc46f00375b0681ff8f", "sweep": "98a8d1d90b64baf9cbca7197d95f139901ae28f6"}, "shrink": {"albert": [["albert-base-v2-mic", 0]], "apertus": [["Apertus-v1.1-0.5B-Instruct-mic", 0]], "audioflamingo3": [["audio-flamingo-3-hf-mic", 0]], "bart": [["bart-base-mic", 0]], "beit": [["beit-base-patch16-224-mic", 0]], "bert": [["bert-base-uncased-mic", 0], ["bert-mic", 0]], "big_bird": [["bigbird-roberta-base-mic", 0]], "blenderbot": [["blenderbot-400M-distill-mic", 0]], "bloom": [["bloomz-560m-mic", 0]], "camembert": [["camembert-base-mic", 0]], "canine": [["canine-c-mic", 0]], "chinese_clip": [["chinese-clip-vit-base-patch16-mic", 0]], "clap": [["clap-htsat-fused-mic", 0]], "clip": [["clip-mic", 0], ["clip-vit-base-patch32-mic", 0]], "clip_text_model": [["tiny-clip-text-2-mic", 0]], "clip_vision_model": [["clip-vit-base-patch32-vision-mic", 0]], "clipseg": [["clipseg-rd64-refined-mic", 0]], "cohere": [["aya-expanse-8b-ungated-mic", 0]], "cohere2": [["tiny-random-aya-base-mic", 0]], "convbert": [["conv-bert-base-mic", 0]], "convnext": [["convnext-tiny-224-mic", 0]], "convnextv2": [["convnextv2-nano-1k-224-mic", 0]], "cosmos3_omni": [["Cosmos3-Nano-mic", 0]], "csm": [["csm-1b-mic", 0]], "deberta": [["deberta-large-mnli-mic", 0]], "deberta-v2": [["mDeBERTa-v3-base-xnli-multilingual-nli-2mil7-mic", 0]], "deepseek_v2": [["DeepSeek-Coder-V2-Lite-Instruct-mic", 0]], "deepseek_v3": [["DeepSeek-R1-mic", 0]], "deepseek_v32": [["DeepSeek-V3.2-mic", 0]], "deepseek_v4": [["DeepSeek-V4-Flash-0731-mic", 0]], "detr": [["detr-mic", 0]], "dinov2": [["dinov2-small-mic", 0]], "dinov2_with_registers": [["dinov2-with-registers-base-mic", 0]], "dinov3_vit": [["dinov3-vitl16-pretrain-lvd1689m-mic", 0]], "distilbert": [["distilbert-base-uncased-mic", 0]], "dpt": [["dpt-large-mic", 0]], "electra": [["electra-small-discriminator-mic", 0]], "embedding_gemma2": [["embeddinggemma-2-mic", 0]], "encoder-decoder": [["bert2bert-cnn_dailymail-fp16-mic", 0]], "esm": [["esm2_t6_8M_UR50D-mic", 0]], "esmc": [["ESMC-300M-hf-mic", 0]], "eurobert": [["EuroBERT-210m-mic", 0]], "falcon": [["falcon-7b-mic", 0]], "falcon_h1": [["Falcon-H1-Tiny-90M-Instruct-mic", 0]], "florence2": [["Florence-2-base-mic", 0]], "fnet": [["fnet-base-mic", 0]], "gemma": [["gemma-2b-mic", 0]], "gemma2": [["gemma-2-2b-it-mic", 0]], "gemma3": [["gemma3-mic", 0]], "gemma3n": [["gemma-3n-E2B-it-mic", 0]], "gemma4": [["gemma-4-E2B-it-mic", 0]], "gemma4_unified": [["gemma-4-12B-it-mic", 0]], "glm4": [["GLM-4-9B-0414-mic", 0]], "glm4_moe": [["GLM-4.5-Air-mic", 0]], "glm4_moe_lite": [["GLM-4.7-Flash-mic", 0]], "glm5_next": [["GLM-5.3-Flash-mic", 1]], "glm_moe_dsa": [["GLM-5.3-mic", 1]], "glm_ocr": [["GLM-OCR-mic", 0]], "got_ocr2": [["got_ocr2-mic", 0]], "gpt2": [["gpt2-mic", 0]], "gpt_bigcode": [["tiny_starcoder_py-mic", 0]], "gpt_neox": [["pythia-160m-mic", 0]], "gpt_oss": [["gpt_oss-mic", 0]], "granite": [["granite-3.1-2b-instruct-mic", 0]], "granite4_vision": [["granite-vision-4.1-4b-mic", 0]], "granite_speech": [["granite-4.0-1b-speech-mic", 0]], "granitemoe": [["granite-3.0-1b-a400m-instruct-mic", 0]], "granitemoehybrid": [["granite-4.0-h-350m-mic", 0]], "grounding-dino": [["grounding-dino-tiny-mic", 0]], "hubert": [["hubert-large-ls960-ft-mic", 0]], "hunyuan_v1_dense": [["Hunyuan-0.5B-Instruct-mic", 0]], "hunyuan_vl": [["HunyuanOCR-mic", 0]], "hy_v3": [["Hy-MT2-30B-A3B-mic", 0]], "idefics3": [["granite-docling-258M-mic", 0]], "inkling_mm_model": [["Inkling-mic", 0]], "internvl": [["internvl-mic", 0]], "jamba": [["Jamba-tiny-dev-mic", 0]], "janus": [["Janus-Pro-1B-mic", 0]], "jina_embeddings_v3": [["jina-embeddings-v3-hf-mic", 0]], "laguna": [["Laguna-XS.2-mic", 0]], "layoutlm": [["layoutlm-base-uncased-mic", 0]], "lfm2": [["LFM2.5-230M-Base-mic", 0]], "lfm2_moe": [["LFM2-8B-A1B-mic", 0]], "lfm2_vl": [["LFM2.5-VL-450M-mic", 0]], "llama": [["llama-mic", 0], ["smollm-1.7b-mic", 0], ["smollm-mic", 0]], "llama4": [["Llama-4-Scout-17B-16E-Instruct-mic", 0]], "llava": [["llava-mic", 0]], "llava_next": [["llava-v1.6-mistral-7b-hf-mic", 0]], "llava_next_video": [["LLaVA-NeXT-Video-7B-hf-mic", 0]], "llava_onevision": [["llava-onevision-qwen2-0.5b-ov-hf-mic", 0]], "longformer": [["longformer-base-4096-mic", 0]], "longt5": [["long-t5-tglobal-base-mic", 0]], "luke": [["luke-base-mic", 0]], "m2m_100": [["nllb-200-distilled-600M-mic", 0]], "mamba": [["mamba-130m-hf-mic", 0]], "marian": [["opus-mt-fr-en-mic", 0]], "mask2former": [["mask2former-swin-tiny-coco-instance-mic", 0]], "mbart": [["mbart-large-50-mic", 0]], "minicpm3": [["MiniCPM3-4B-mic", 0]], "minicpmv4_6": [["MiniCPM-V-4.6-mic", 0]], "minimax_m2": [["MiniMax-M2.5-mic", 0]], "minimax_m3_vl": [["MiniMax-M3-MXFP8-mic", 0]], "ministral3": [["Nemotron-3-Embed-1B-BF16-mic", 0]], "mistral": [["Mistral-7B-Instruct-v0.2-mic", 1]], "mistral3": [["Ministral-3-3B-Instruct-2512-mic", 0], ["mistral3-mic", 0]], "mixtral": [["Mixtral-8x7B-Instruct-v0.1-mic", 0]], "mllama": [["Llama-3.2-11B-Vision-Instruct-mic", 0]], "mm-grounding-dino": [["mm_grounding_dino_tiny_o365v1_goldg_v3det-mic", 0]], "mobilenet_v2": [["mobilenet_v2_1.0_224-mic", 0]], "mobilevit": [["mobilevit-xx-small-mic", 0]], "modernbert": [["modernbert-mic", 0]], "moss_transcribe_diarize": [["MOSS-Transcribe-Diarize-HF-mic", 0]], "mpnet": [["mpnet-base-mic", 0]], "mt5": [["mt5-small-mic", 0]], "muse_glimmer": [["tiny-random-muse-glimmer-mic", 0]], "nemotron_h": [["NVIDIA-Nemotron-3-Nano-4B-BF16-mic", 0]], "nomic_bert": [["nomic-embed-text-v1.5-mic", 0]], "olmo": [["OLMo-1B-hf-mic", 0]], "olmo2": [["OLMo-2-0425-1B-mic", 0]], "olmo3": [["Olmo-3-7B-Instruct-mic", 0]], "olmoe": [["OLMoE-1B-7B-0924-mic", 0]], "openai_privacy_filter": [["privacy-filter-mic", 0]], "opt": [["opt-125m-mic", 0]], "owlv2": [["owlv2-base-patch16-ensemble-mic", 0]], "owlvit": [["owlvit-base-patch32-mic", 0]], "paddleocr_vl": [["tibetan-ocr-mic", 0]], "paligemma": [["paligemma-3b-pt-224-mic", 0]], "pegasus": [["pegasus-xsum-mic", 0]], "perceiver": [["language-perceiver-mic", 0]], "phi": [["phi-2-mic", 0]], "phi3": [["Phi-3.5-mini-instruct-mic", 0]], "phi4_multimodal": [["Phi-4-multimodal-instruct-mic", 0]], "phimoe": [["Phi-3.5-MoE-0.8B-A0.2B-mic", 0]], "pvt": [["pvt-tiny-224-mic", 0]], "qwen2": [["qwen2-mic", 0]], "qwen2_5_omni": [["qwen2_5_omni-mic", 0]], "qwen2_5_omni_thinker": [["LCO-Embedding-Omni-3B-mic", 0]], "qwen2_5_vl": [["qwen2_5_vl-mic", 0]], "qwen2_audio": [["qwen2_audio-mic", 0]], "qwen2_moe": [["Qwen1.5-MoE-A2.7B-mic", 0]], "qwen2_vl": [["Qwen2-VL-2B-Instruct-mic", 0]], "qwen3": [["Qwen3-0.6B-mic", 0]], "qwen3_5_moe": [["Qwen3.5-35B-A3B-mic", 0]], "qwen3_5_moe_text": [["Qwen3.8-1.0B-A0.6B-mic", 0]], "qwen3_5_text": [["Qwen3.5-0.8B-text-only-mic", 0]], "qwen3_asr": [["Qwen3-ASR-0.6B-hf-mic", 0]], "qwen3_moe": [["Qwen3-30B-A3B-mic", 0]], "qwen3_next": [["Qwen3-Next-80B-A3B-Instruct-mic", 0]], "qwen3_vl": [["Qwen3-VL-2B-Instruct-mic", 0]], "qwen3_vl_moe": [["Qwen3-VL-30B-A3B-Instruct-mic", 0]], "qwen4_exp": [["Qwen3.8-Flash-Next-mic", 0]], "recurrent_gemma": [["recurrentgemma-2b-mic", 0]], "resnet": [["resnet-18-mic", 0]], "roberta": [["roberta-base-mic", 0]], "rt_detr": [["rtdetr_r50vd-mic", 0]], "rt_detr_v2": [["docling-layout-heron-mic", 0]], "sam2": [["sam2.1-hiera-tiny-mic", 4]], "seamless_m4t_v2": [["seamless-m4t-v2-large-mic", 0]], "seed_oss": [["Seed-OSS-36B-Instruct-mic", 0]], "segformer": [["segformer-b0-finetuned-ade-512-512-mic", 0]], "siglip": [["siglip-base-patch16-224-mic", 0]], "smollm3": [["SmolLM3-3B-mic", 1]], "smolvlm": [["smolvlm-mic", 0]], "speecht5": [["speech_t5-mic", 0]], "stablelm": [["stablelm-2-1_6b-mic", 0]], "starcoder2": [["starcoder2-3b-mic", 0]], "swin": [["swin-tiny-patch4-window7-224-mic", 0]], "swinv2": [["swinv2-tiny-patch4-window16-256-mic", 0]], "t5": [["t5-mic", 0], ["t5-small-mic", 0]], "table-transformer": [["table_transformer-mic", 0]], "timesfm": [["timesfm-2.0-500m-pytorch-mic", 0]], "tipsv2": [["tipsv2-b14-mic", 0]], "umt5": [["umt5-small-mic", 0]], "unlimited_ocr": [["Unlimited-OCR-mic", 0]], "vibevoice": [["VibeVoice-1.5B-hf-mic", 0]], "vit": [["vit-mic", 0]], "vitmatte": [["vitmatte-small-composition-1k-mic", 0]], "vits": [["mms-tts-eng-mic", 0]], "voxtral": [["Voxtral-Mini-3B-2507-mic", 0]], "wav2vec2": [["wav2vec2-base-960h-mic", 0], ["wav2vec2-mic", 0]], "wavlm": [["wavlm-base-plus-mic", 0]], "whisper": [["whisper-mic", 0], ["whisper-tiny-mic", 0]], "xlm-roberta": [["xlm-roberta-base-mic", 0]], "xlm-roberta-xl": [["xlm-roberta-xl-mic", 0]], "yolos": [["yolos-small-mic", 0]]}, "gated": {"aya_vision": "auto", "chameleon": "manual", "chmv2": "manual", "cohere_asr": "auto", "csm_depth_decoder_model": "auto", "cwm": "manual", "dinov3_convnext": "manual", "gemma": "manual", "gemma2": "manual", "gemma3": "manual", "gemma3_text": "manual", "gemma3n": "manual", "jais2": "auto", "lasr_ctc": "auto", "llama4_vision_model": "manual", "moshi": "auto", "muse_spark": "manual", "muse_spark_audio": "manual", "muse_spark_text": "manual", "muse_spark_vision": "manual", "neucodec": "auto", "perception_lm": "manual", "recurrent_gemma": "manual", "sam3_detr_decoder": "manual", "sam3_detr_encoder": "manual", "sam3_geometry_encoder": "manual", "sam3_mask_decoder": "manual", "sam3_tracker_mask_decoder": "manual", "sam3_tracker_prompt_encoder": "manual", "sam3_tracker_video": "manual", "sam3_tracker_video_mask_decoder": "manual", "sam3_tracker_video_prompt_encoder": "manual", "sam3_video": "manual", "shieldgemma2": "manual", "t5gemma": "manual", "t5gemma2": "manual", "vaultgemma": "manual"}}
\ No newline at end of file
+{"rows": [["EvollaModel", "Reachable", "", 732, "westlake-repl/Evolla-10B-hf"], ["afmoe", "Reachable", "", 84004, "optimum-intel-internal-testing/tiny-random-trinity"], ["aimv2", "Reachable", "", 10627, "apple/aimv2-large-patch14-224-lit"], ["aimv2_text_model", "Unreachable", "component config only", 0, ""], ["aimv2_vision_model", "Reachable", "", 3412, "apple/aimv2-large-patch14-224"], ["albert", "Done", "", 2009876, "albert/albert-base-v2"], ["align", "Unreachable", "forward fails after shrink", 21429, "kakaobrain/align-base"], ["align_text_model", "Unreachable", "component config only", 0, ""], ["align_vision_model", "Unreachable", "component config only", 0, ""], ["altclip", "Reachable", "", 7652, "BAAI/AltCLIP"], ["altclip_text_model", "Unreachable", "component config only", 0, ""], ["altclip_vision_model", "Unreachable", "component config only", 0, ""], ["apertus", "Done", "", 457103, "swiss-ai/Apertus-8B-Instruct-2509"], ["arcee", "Reachable", "", 64187, "optimum-intel-internal-testing/tiny-random-ArceeForCausalLM"], ["aria", "Unreachable", "forward fails after shrink", 29760, "rhymes-ai/Aria"], ["aria_text", "Unreachable", "forward fails after shrink", 88, "hf-tiny-v2/tiny-random-AriaTextForCausalLM"], ["audio-spectrogram-transformer", "Unreachable", "forward fails after shrink", 1651174, "MIT/ast-finetuned-audioset-10-10-0.4593"], ["audioflamingo3", "Done", "", 144463, "nvidia/audio-flamingo-3-hf"], ["audioflamingo3_encoder", "Unreachable", "forward fails after shrink", 27, "hf-tiny-v2/tiny-random-AudioFlamingo3Encoder"], ["autoformer", "Unreachable", "synthetic inputs rejected", 79649, "huggingface/autoformer-tourism-monthly"], ["axk1", "Reachable", "", 5094, "skt/A.X-K1"], ["axk2", "Reachable", "", 90892, "skt/A.X-K2-NVFP4"], ["aya_vision", "Reachable", "", 3689, "CohereLabs/aya-vision-8b"], ["bailing_hybrid", "Unreachable", "forward fails after shrink", 727122, "rapid-mlx/Ling-3.0-tiny-MLX-4bit"], ["bamba", "Reachable", "", 11728, "ibm-ai-platform/Bamba-9B-v1"], ["bark", "Unreachable", "forward fails after shrink", 28264, "suno/bark"], ["bart", "Done", "", 6328395, "facebook/bart-large-mnli"], ["beam", "Reachable", "", 20667, "reflection/Beam-1.0-501B-A23B"], ["beit", "Done", "", 169127, "microsoft/dit-base-finetuned-rvlcdip"], ["bert", "Done", "", 527874026, "sentence-transformers/all-MiniLM-L6-v2"], ["bert-generation", "Reachable", "", 3386, "ybelkada/random-tiny-BertGenerationModel"], ["big_bird", "Done", "", 194505, "google/bigbird-roberta-base"], ["bigbird_pegasus", "Reachable", "", 71110, "optimum-intel-internal-testing/tiny-random-bigbird_pegasus"], ["biogpt", "Reachable", "", 56866, "optimum-intel-internal-testing/tiny-random-BioGptForCausalLM"], ["bit", "Unreachable", "model build fails", 8676, "google/bit-50"], ["bitnet", "Reachable", "", 35767, "optimum-intel-internal-testing/tiny-random-bitnet"], ["blenderbot", "Done", "", 235225, "optimum-intel-internal-testing/tiny-random-BlenderbotModel"], ["blenderbot-small", "Reachable", "", 5857, "facebook/blenderbot_small-90M"], ["blip", "Unreachable", "forward fails after shrink", 2359364, "Salesforce/blip-image-captioning-base"], ["blip-2", "Unreachable", "forward fails after shrink", 457477, "Salesforce/blip2-opt-2.7b"], ["blip_2_qformer", "Unreachable", "forward fails after shrink", 145, "hf-tiny-v2/tiny-random-Blip2QFormerModel"], ["blip_2_vision_model", "Unreachable", "component config only", 24, "tmnam20/blip2-opt-2.7b-vision"], ["blip_text_model", "Unreachable", "component config only", 0, ""], ["blip_vision_model", "Unreachable", "component config only", 0, ""], ["bloom", "Done", "", 1527823, "bigscience/bloomz-560m"], ["blt", "Unreachable", "forward fails after shrink", 6999, "itazap/blt-1b-hf"], ["blt_global_transformer", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["blt_local_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["blt_local_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["blt_patcher", "No checkpoint", "no Hub checkpoint has this model_type (documented itazap/blt-1b-hf is another type)", 0, ""], ["bridgetower", "Reachable", "", 80, "regisss/bridgetower-newyorker-a100-8x"], ["bridgetower_text_model", "No checkpoint", "no Hub checkpoint has this model_type (documented BridgeTower/bridgetower-base is another type)", 0, ""], ["bridgetower_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (documented BridgeTower/bridgetower-base is another type)", 0, ""], ["bros", "Unreachable", "forward fails after shrink", 42661, "naver-clova-ocr/bros-base-uncased"], ["camembert", "Done", "", 2226781, "almanach/camembert-base"], ["canary", "Unreachable", "forward fails after shrink", 237192, "nvidia/canary-1b-v2"], ["canary_decoder", "Unreachable", "component config only", 0, ""], ["canine", "Done", "", 167492, "google/canine-c"], ["chameleon", "Reachable", "", 8504, "facebook/chameleon-7b"], ["chameleon_vqgan", "Unreachable", "component config only", 33, "leloy/Anole-7b-v0.1-vqvae-hf"], ["chinese_clip", "Done", "", 276376, "OFA-Sys/chinese-clip-vit-base-patch16"], ["chinese_clip_text_model", "Unreachable", "component config only", 0, ""], ["chinese_clip_vision_model", "Reachable", "", 57, "hf-tiny-v2/tiny-random-ChineseCLIPVisionModel"], ["chmv2", "Unreachable", "forward fails after shrink", 1526, "facebook/dinov3-vitl16-chmv2-dpt-head"], ["clap", "Done", "", 9104856, "laion/clap-htsat-fused"], ["clap_audio_model", "Unreachable", "component config only", 257, "NGDtuanh/abook-music-student"], ["clap_text_model", "Unreachable", "component config only", 61, "werent4/larger_clap_general_text_model"], ["clip", "Done", "", 37612455, "openai/clip-vit-base-patch32"], ["clip_text_model", "Done", "", 940714, "peft-internal-testing/tiny-clip-text-2"], ["clip_vision_model", "Done", "", 432156, "xingpng/OneIG-StyleEncoder"], ["clipseg", "Done", "", 545844, "CIDAS/clipseg-rd64-refined"], ["clipseg_text_model", "Unreachable", "component config only", 0, ""], ["clipseg_vision_model", "Unreachable", "component config only", 0, ""], ["clvp", "Unreachable", "forward fails after shrink", 5530, "susnato/clvp_dev"], ["clvp_decoder", "Unreachable", "component config only", 0, ""], ["clvp_encoder", "Unreachable", "component config only", 0, ""], ["codegen", "Unreachable", "forward fails after shrink", 168810, "Salesforce/codegen2-16B_P"], ["cohere", "Done", "", 166589, "optimum-intel-internal-testing/tiny-random-CohereForCausalLM"], ["cohere2", "Done", "", 162977, "optimum-intel-internal-testing/tiny-random-aya-base"], ["cohere2_moe", "Reachable", "", 15203, "CohereLabs/North-Mini-Code-1.0"], ["cohere2_vision", "Reachable", "", 37679, "CohereLabs/command-a-plus-05-2026-bf16"], ["cohere_asr", "Unreachable", "forward fails after shrink", 350010, "CohereLabs/cohere-transcribe-arabic-07-2026"], ["cohere_compass", "Reachable", "", 34840, "CohereLabs/North-Micro-Vision-Instruct"], ["cohere_compass_text", "Reachable", "", 41, "hf-tiny-v2/tiny-random-CohereCompassTextForSequenceClassification"], ["cohere_compass_vision", "Unreachable", "forward fails after shrink", 15, "hf-tiny-v2/tiny-random-CohereCompassVisionModel"], ["colmodernvbert", "Unreachable", "no usable checkpoint", 1649, "Qdrant/colmodernvbert"], ["colpali", "Reachable", "", 8568, "vidore/colpali-v1.3-hf"], ["colqwen2", "Reachable", "", 36700, "vidore/colqwen2-v1.0-hf"], ["conditional_detr", "Reachable", "", 62673, "yainage90/fashion-object-detection"], ["convbert", "Done", "", 509373, "YituTech/conv-bert-base"], ["convnext", "Done", "", 192305, "facebook/convnext-tiny-224"], ["convnextv2", "Done", "", 330963, "facebook/convnextv2-base-22k-384"], ["cosmos3_edge", "Reachable", "", 83288, "nvidia/Cosmos3-Edge"], ["cosmos3_edge_text", "Reachable", "", 14, "hf-tiny-v2/tiny-random-Cosmos3EdgeTextModel"], ["cosmos3_edge_vision", "Unreachable", "forward fails after shrink", 12, "hf-tiny-v2/tiny-random-Cosmos3EdgeVisionModel"], ["cosmos3_omni", "Done", "", 244937, "nvidia/Cosmos3-Nano"], ["cpmant", "Reachable", "", 3742, "openbmb/cpm-ant-10b"], ["csm", "Done", "", 145078, "sesame/csm-1b"], ["csm_depth_decoder_model", "Unreachable", "checkpoint gated", 0, "sesame/csm-1b"], ["ctrl", "Reachable", "", 17032, "Salesforce/ctrl"], ["cvt", "Reachable", "", 12319, "microsoft/cvt-13"], ["cwm", "Unreachable", "no usable checkpoint", 1007, "facebook/cwm"], ["d_fine", "Unreachable", "forward fails after shrink", 126298, "ustc-community/dfine-xlarge-obj2coco"], ["dab-detr", "Unreachable", "config load fails", 7050, "IDEA-Research/dab-detr-resnet-50"], ["dac", "Unreachable", "forward fails after shrink", 72000, "descript/dac_44khz"], ["data2vec-audio", "Reachable", "", 39013, "optimum-intel-internal-testing/tiny-random-Data2VecAudioModel"], ["data2vec-text", "Reachable", "", 7219, "optimum-intel-internal-testing/tiny-random-Data2VecTextModel"], ["data2vec-vision", "Unreachable", "shrink fails", 17284, "optimum-intel-internal-testing/tiny-random-Data2VecVisionModel"], ["dbrx", "Reachable", "", 16983, "trl-internal-testing/tiny-DbrxForCausalLM"], ["deberta", "Done", "", 692267, "microsoft/deberta-large-mnli"], ["deberta-v2", "Done", "", 21746039, "microsoft/mdeberta-v3-base"], ["decision_transformer", "Unreachable", "synthetic inputs rejected", 6158, "edbeeching/decision-transformer-gym-hopper-medium"], ["deepseek_ocr2", "Unreachable", "synthetic inputs rejected", 22922, "optimum-intel-internal-testing/tiny-random-deepseek-ocr-2"], ["deepseek_ocr2_encoder", "Unreachable", "component config only", 0, ""], ["deepseek_ocr2_sam_vision_model", "Unreachable", "component config only", 0, ""], ["deepseek_ocr2_text", "Unreachable", "component config only", 0, ""], ["deepseek_ocr2_vision", "Unreachable", "component config only", 0, ""], ["deepseek_v2", "Done", "", 668975, "deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct"], ["deepseek_v3", "Done", "", 4950570, "deepseek-ai/DeepSeek-R1"], ["deepseek_v32", "Done", "", 6161743, "deepseek-ai/DeepSeek-V3.2"], ["deepseek_v4", "Done", "", 6213673, "deepseek-ai/DeepSeek-V4-Flash-0731"], ["deepseek_vl", "Reachable", "", 8828, "deepseek-community/deepseek-vl-1.3b-chat"], ["deepseek_vl_hybrid", "Reachable", "", 8249, "deepseek-community/deepseek-vl-7b-chat"], ["deformable_detr", "Reachable", "", 23013, "SenseTime/deformable-detr"], ["deimv2", "Unreachable", "forward fails after shrink", 4393, "alessioarcara/deimv2-deimv2_dinov3_x_coco"], ["deit", "Reachable", "", 50030, "facebook/deit-base-distilled-patch16-384"], ["depth_anything", "Unreachable", "forward fails after shrink", 2447209, "depth-anything/Depth-Anything-V2-Small-hf"], ["depth_pro", "Unreachable", "forward fails after shrink", 52957, "apple/DepthPro-hf"], ["detr", "Done", "", 385370, "TahaDouaji/detr-doc-table-detection"], ["dia", "Unreachable", "no usable checkpoint", 16396, "nari-labs/Dia-1.6B-0626"], ["dia_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented nari-labs/Dia-1.6B is another type)", 0, ""], ["dia_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented nari-labs/Dia-1.6B is another type)", 0, ""], ["diffllama", "Reachable", "", 4593, "kajuma/DiffLlama-0.3B-handcut"], ["diffusion_gemma", "Unreachable", "forward fails after shrink", 1034537, "google/diffusiongemma-26B-A4B-it"], ["diffusion_gemma_text", "Unreachable", "component config only", 0, ""], ["dinat", "Unreachable", "model build fails", 209, "shi-labs/dinat-mini-in1k-224"], ["dinov2", "Done", "", 7235906, "facebook/dinov2-base"], ["dinov2_with_registers", "Done", "", 354429, "facebook/dinov2-with-registers-base"], ["dinov3_convnext", "Reachable", "", 95750, "facebook/dinov3-convnext-tiny-pretrain-lvd1689m"], ["dinov3_vit", "Done", "", 2089303, "facebook/dinov3-vitl16-pretrain-lvd1689m"], ["distilbert", "Done", "", 20324420, "distilbert/distilbert-base-uncased"], ["doge", "Reachable", "", 8362, "SmallDoge/Doge-320M"], ["donut-swin", "Reachable", "", 2963, "optimum-intel-internal-testing/tiny-random-DonutSwinModel"], ["dots1", "Reachable", "", 6731, "rednote-hilab/dots.llm1.base"], ["dpr", "Reachable", "", 53829, "facebook/dpr-question_encoder-single-nq-base"], ["dpt", "Done", "", 620365, "Intel/dpt-hybrid-midas"], ["edgetam", "Unreachable", "no usable checkpoint", 9579, "onnx-community/EdgeTAM-ONNX"], ["edgetam_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_video", "Unreachable", "forward fails after shrink", 11709, "yonigozlan/EdgeTAM-hf"], ["edgetam_video_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_video_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented yonigozlan/EdgeTAM-hf is another type)", 0, ""], ["edgetam_vision_model", "Reachable", "", 53, "hf-tiny-v2/tiny-random-EdgeTamVisionModel"], ["efficientloftr", "Unreachable", "shrink fails", 16245, "zju-community/efficientloftr"], ["efficientnet", "Unreachable", "forward fails after shrink", 168152, "docling-project/DocumentFigureClassifier-v2.5"], ["electra", "Done", "", 28465435, "google/electra-base-discriminator"], ["embedding_gemma2", "Done", "", 719174, "onnx-community/embeddinggemma-2-ONNX"], ["embedding_gemma2_text", "Reachable", "", 171, "HukukcuApp/tr-legal-embedding-gemma-2"], ["emu3", "Reachable", "", 20741, "BAAI/Emu3-Chat-hf"], ["emu3_text_model", "Unreachable", "component config only", 118, "hf-tiny-v2/tiny-random-Emu3TextModel"], ["emu3_vqgan", "Unreachable", "component config only", 0, ""], ["encodec", "Unreachable", "forward fails after shrink", 105065, "facebook/encodec_24khz"], ["encoder-decoder", "Done", "", 29649, "optimum-internal-testing/tiny-random-encoder-decoder-gpt2-bert"], ["eomt", "Unreachable", "forward fails after shrink", 67972, "tue-mps/coco_panoptic_eomt_large_640"], ["eomt_dinov3", "Unreachable", "forward fails after shrink", 21640, "tue-mps/eomt-dinov3-ade-semantic-large-512"], ["ernie", "Reachable", "", 73981, "shibing624/text2vec-base-chinese-paraphrase"], ["ernie4_5", "Reachable", "", 14731, "baidu/ERNIE-4.5-0.3B-PT"], ["ernie4_5_moe", "Unreachable", "forward fails after shrink", 123822, "baidu/ERNIE-4.5-21B-A3B-Thinking"], ["ernie4_5_vl_moe", "Unreachable", "forward fails after shrink", 45, "hf-tiny-v2/tiny-random-Ernie4_5_VLMoeForConditionalGeneration"], ["ernie4_5_vl_moe_text", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/ERNIE-4.5-VL-28B-A3B-PT is another type)", 0, ""], ["ernie4_5_vl_moe_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/ERNIE-4.5-VL-28B-A3B-PT is another type)", 0, ""], ["esm", "Done", "", 7030784, "facebook/esm2_t33_650M_UR50D"], ["esmc", "Done", "", 393143, "biohub/ESMC-300M"], ["esmfold2", "Unreachable", "shrink fails", 388538, "biohub/ESMFold2"], ["eurobert", "Done", "", 847801, "OpenMed/OpenMed-NER-ChemicalDetect-EuroMed-212M"], ["evolla", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["exaone4", "Reachable", "", 64834, "LGAI-EXAONE/EXAONE-4.0-32B"], ["exaone4_5", "Reachable", "", 39320, "LGAI-EXAONE/EXAONE-4.5-33B"], ["exaone4_5_vision", "Unreachable", "forward fails after shrink", 31, "hf-tiny-v2/tiny-random-Exaone4_5_VisionModel"], ["exaone_moe", "Reachable", "", 18548, "LGAI-EXAONE/K-EXAONE-236B-A23B"], ["falcon", "Done", "", 350854, "tiiuae/falcon-7b"], ["falcon_h1", "Done", "", 104762, "tiiuae/Falcon-H1-3B-Instruct"], ["falcon_mamba", "Reachable", "", 61205, "trl-internal-testing/tiny-FalconMambaForCausalLM"], ["fast_vlm", "Reachable", "", 5841, "KamilaMila/FastVLM-7B"], ["fastspeech2_conformer", "Unreachable", "forward fails after shrink", 2880, "espnet/fastspeech2_conformer"], ["fastspeech2_conformer_hifigan", "No checkpoint", "no Hub checkpoint has this model_type (documented espnet/fastspeech2_conformer_with_hifigan is another type)", 0, ""], ["fastspeech2_conformer_with_hifigan", "Unreachable", "forward fails after shrink", 2800, "espnet/fastspeech2_conformer_with_hifigan"], ["flaubert", "Reachable", "", 74651, "flaubert/flaubert_base_uncased"], ["flava", "Reachable", "", 32370, "facebook/flava-full"], ["flava_image_codebook", "Unreachable", "component config only", 23, "facebook/flava-image-codebook"], ["flava_image_model", "Unreachable", "component config only", 0, ""], ["flava_multimodal_model", "Unreachable", "component config only", 0, ""], ["flava_text_model", "Unreachable", "component config only", 0, ""], ["flex_olmo", "Reachable", "", 12810, "allenai/FlexOlmo-7x7B-1T"], ["flint", "Reachable", "", 60545, "RecursiveSI/flint"], ["florence2", "Done", "", 4293112, "microsoft/Florence-2-base"], ["florence_vision", "Unreachable", "component config only", 0, ""], ["fnet", "Done", "", 478381, "google/fnet-base"], ["focalnet", "Reachable", "", 11569, "microsoft/focalnet-tiny"], ["fsmt", "Reachable", "", 57368, "facebook/wmt19-en-de"], ["fun_asr_nano", "Unreachable", "shrink fails", 162557, "FunAudioLLM/Fun-ASR-Nano-2512-hf"], ["fun_asr_nano_encoder", "Unreachable", "forward fails after shrink", 21, "hf-tiny-v2/tiny-random-FunAsrNanoEncoder"], ["funnel", "Reachable", "", 11894, "funnel-transformer/small"], ["fuyu", "Reachable", "", 12098, "adept/fuyu-8b"], ["gemma", "Done", "", 611743, "google/gemma-2b"], ["gemma2", "Done", "", 2138107, "hugging-quants/gemma-2-9b-it-AWQ-INT4"], ["gemma3", "Done", "", 6276380, "google/gemma-3-4b-it"], ["gemma3_text", "Done", "", 3985142, "google/embeddinggemma-300m"], ["gemma3n", "Done", "", 360811, "unsloth/gemma-3n-E4B-it"], ["gemma3n_audio", "Unreachable", "synthetic inputs rejected", 63, "hf-tiny-v2/tiny-random-Gemma3nAudioEncoder"], ["gemma3n_text", "Reachable", "", 2175, "optimum-intel-internal-testing/tiny-random-gemma3n-text"], ["gemma3n_vision", "Reachable", "", 0, ""], ["gemma4", "Done", "", 47591452, "google/gemma-4-31B-it"], ["gemma4_assistant", "Unreachable", "forward fails after shrink", 683069, "google/gemma-4-26B-A4B-it-assistant"], ["gemma4_audio", "Unreachable", "forward fails after shrink", 138, "rnagabh/gemma4-audio-encoder"], ["gemma4_text", "Reachable", "", 41837, "AlexKoff88/gemma4_tinystories"], ["gemma4_unified", "Done", "", 4749964, "google/gemma-4-12B-it"], ["gemma4_unified_assistant", "Unreachable", "forward fails after shrink", 37821, "google/gemma-4-12B-it-assistant"], ["gemma4_unified_audio", "Unreachable", "component config only", 0, ""], ["gemma4_unified_text", "Reachable", "", 2960, "jgeuter/Jeff-1.0-Medium"], ["gemma4_unified_vision", "Unreachable", "component config only", 0, ""], ["gemma4_vision", "Unreachable", "forward fails after shrink", 214, "rnagabh/gemma4-vision-encoder"], ["git", "Reachable", "", 46165, "microsoft/git-base"], ["git_vision_model", "Unreachable", "component config only", 0, ""], ["glm", "Unreachable", "no usable checkpoint", 43219, "optimum-intel-internal-testing/tiny-random-glm-edge"], ["glm4", "Done", "", 139087, "optimum-intel-internal-testing/tiny-random-glm4"], ["glm46v", "Unreachable", "no usable checkpoint", 170, "hf-tiny-v2/tiny-random-Glm46VModel"], ["glm4_moe", "Done", "", 1607130, "zai-org/GLM-4.5-Air"], ["glm4_moe_lite", "Done", "", 1802094, "zai-org/GLM-4.7-Flash"], ["glm4v", "Unreachable", "forward fails after shrink", 277739, "zai-org/GLM-4.1V-9B-Thinking"], ["glm4v_moe", "Unreachable", "forward fails after shrink", 35179, "zai-org/GLM-4.5V"], ["glm4v_moe_text", "Unreachable", "forward fails after shrink", 58, "hf-tiny-v2/tiny-random-Glm4vMoeTextModel"], ["glm4v_moe_vision", "Unreachable", "no usable checkpoint", 38, "hf-tiny-v2/tiny-random-Glm4vMoeVisionModel"], ["glm4v_text", "Unreachable", "forward fails after shrink", 58, "hf-tiny-v2/tiny-random-Glm4vTextModel"], ["glm4v_vision", "Unreachable", "forward fails after shrink", 38, "hf-tiny-v2/tiny-random-Glm4vVisionModel"], ["glm5_next", "Done", "", 7619607, "zai-org/GLM-5.3-Flash"], ["glm5_next_text", "Reachable", "", 23, "hf-tiny-v2/tiny-random-Glm5NextTextModel"], ["glm5_next_vision", "Unreachable", "forward fails after shrink", 65, "immanuelpeter/GLM-5.3-Flash-Vision"], ["glm_image", "Unreachable", "no usable checkpoint", 111, "hf-tiny-v2/tiny-random-GlmImageModel"], ["glm_image_text", "Unreachable", "no usable checkpoint", 55, "hf-tiny-v2/tiny-random-GlmImageTextModel"], ["glm_image_vision", "Unreachable", "no usable checkpoint", 35, "hf-tiny-v2/tiny-random-GlmImageVisionModel"], ["glm_image_vqmodel", "Unreachable", "no usable checkpoint", 6, "hf-tiny-v2/tiny-random-GlmImageVQVAE"], ["glm_moe_dsa", "Done", "", 7133417, "zai-org/GLM-5.3"], ["glm_ocr", "Done", "", 1798225, "zai-org/GLM-OCR"], ["glm_ocr_text", "Reachable", "", 51, "hf-tiny-v2/tiny-random-GlmOcrTextModel"], ["glm_ocr_vision", "Unreachable", "forward fails after shrink", 33, "hf-tiny-v2/tiny-random-GlmOcrVisionModel"], ["glmasr", "Reachable", "", 31331, "zai-org/GLM-ASR-Nano-2512"], ["glmasr_encoder", "Reachable", "", 50, "hf-tiny-v2/tiny-random-GlmAsrEncoder"], ["glmga", "No checkpoint", "no Hub checkpoint has this model_type (documented zai-org/GLM-4.1V-9B-Thinking is another type)", 0, ""], ["glpn", "Reachable", "", 8196, "vinvino02/glpn-kitti"], ["got_ocr2", "Done", "", 56531, "stepfun-ai/GOT-OCR-2.0-hf"], ["got_ocr2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["gpt-sw3", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["gpt2", "Done", "", 30545353, "openai-community/gpt2"], ["gpt_bigcode", "Done", "", 188077, "bigcode/tiny_starcoder_py"], ["gpt_neo", "Unreachable", "shrink fails", 865506, "EleutherAI/gpt-neo-125m"], ["gpt_neox", "Done", "", 7013890, "EleutherAI/pythia-160m"], ["gpt_neox_japanese", "Reachable", "", 46916, "optimum-intel-internal-testing/tiny-random-GPTNeoXJapaneseForCausalLM"], ["gpt_oss", "Done", "", 10738031, "openai/gpt-oss-20b"], ["gptj", "Unreachable", "forward fails after shrink", 188360, "EleutherAI/gpt-j-6b"], ["gradium_phonon_tts", "Unreachable", "shrink fails", 114118, "gradium/phonon-31m-en"], ["granite", "Done", "", 1699090, "ibm-granite/granite-4.1-3b"], ["granite4_vision", "Done", "", 194823, "ibm-granite/granite-vision-4.1-4b"], ["granite4_vision_text", "No checkpoint", "no Hub checkpoint has this model_type (documented ibm-granite4_vision_text/granite4_vision_text-3.0-8b-base does not exist)", 0, ""], ["granite_speech", "Done", "", 494648, "ibm-granite/granite-4.0-1b-speech"], ["granite_speech5_ctc", "Unreachable", "forward fails after shrink", 99951, "ibm-granite/granite-speech-5.0-470m-turboctc"], ["granite_speech5_encoder", "Unreachable", "forward fails after shrink", 17, "hf-tiny-v2/tiny-random-GraniteSpeech5Encoder"], ["granite_speech_encoder", "Unreachable", "component config only", 0, ""], ["granite_speech_plus", "Unreachable", "shrink fails", 29187, "ibm-granite/granite-speech-4.1-2b-plus"], ["granite_speech_plus_encoder", "Unreachable", "component config only", 0, ""], ["granite_swa", "Reachable", "", 7057, "ibm-granite/granite-swash-2b"], ["granitemoe", "Done", "", 244089, "ibm-granite/granite-3.1-3b-a800m-instruct"], ["granitemoe_swa", "Reachable", "", 6507, "ibm-granite/granite-swash-3b-a600m"], ["granitemoehybrid", "Done", "", 332014, "optimum-intel-internal-testing/tiny-random-granitemoehybrid"], ["granitemoeshared", "Unreachable", "no usable checkpoint", 1648, "ibm-research/moe-7b-1b-active-shared-experts"], ["grounding-dino", "Done", "", 2682931, "IDEA-Research/grounding-dino-base"], ["groupvit", "Reachable", "", 6660, "nvidia/groupvit-gcc-yfcc"], ["groupvit_text_model", "Unreachable", "component config only", 0, ""], ["groupvit_vision_model", "Unreachable", "component config only", 0, ""], ["gte", "Reachable", "", 56943, "Snowflake/snowflake-arctic-embed-m-v2.0"], ["helium", "Reachable", "", 6264, "kyutai/helium-1-preview-2b"], ["hgnet_v2", "Unreachable", "forward fails after shrink", 149, "ustc-community/hgnet-v2"], ["hiera", "Reachable", "", 11635, "facebook/hiera-tiny-224-hf"], ["higgs_audio_v2", "Reachable", "", 21331, "bosonai/higgs-tts-2-3b-base"], ["higgs_audio_v2_tokenizer", "Unreachable", "model build fails", 235573, "eustlb/higgs-audio-v2-tokenizer"], ["hrm_text", "Unreachable", "forward fails after shrink", 14676, "sapientinc/HRM-Text-1B"], ["hubert", "Done", "", 1327557, "facebook/hubert-large-ls960-ft"], ["hunyuan_v1_dense", "Done", "", 197468, "optimum-intel-internal-testing/tiny-random-hunyuan-v1-dense"], ["hunyuan_v1_moe", "Unreachable", "forward fails after shrink", 64830, "tencent/Hunyuan-A13B-Instruct"], ["hunyuan_vl", "Done", "", 807278, "tencent/HunyuanOCR"], ["hunyuan_vl_text", "Unreachable", "no usable checkpoint", 9, "hf-tiny-v2/tiny-random-HunYuanVLTextModel"], ["hunyuan_vl_vision", "Unreachable", "no usable checkpoint", 10, "hf-tiny-v2/tiny-random-HunYuanVLVisionTransformer"], ["hy_v3", "Done", "", 1408804, "tencent/Hy3"], ["hy_v4", "Reachable", "", 32339, "tencent/Hy4-preview-FP8"], ["hyperclovax", "Reachable", "", 20512, "naver-hyperclovax/HyperCLOVAX-SEED-Think-14B"], ["hyperclovax_vision_v2", "Reachable", "", 10279, "naver-hyperclovax/HyperCLOVAX-SEED-Think-32B"], ["ibert", "Reachable", "", 40201, "optimum-intel-internal-testing/tiny-random-ibert"], ["idefics", "Unreachable", "forward fails after shrink", 21036, "HuggingFaceM4/idefics-80b-instruct"], ["idefics2", "Reachable", "", 18780, "HuggingFaceM4/idefics2-8b"], ["idefics2_perceiver", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics2-8b is another type)", 0, ""], ["idefics2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics2-8b is another type)", 0, ""], ["idefics3", "Done", "", 776490, "ibm-granite/granite-docling-258M"], ["idefics3_vision", "Unreachable", "no usable checkpoint", 53, "hf-tiny-v2/tiny-random-Idefics3VisionTransformer"], ["idefics_perciever", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics-9b is another type)", 0, ""], ["idefics_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented HuggingFaceM4/idefics-9b is another type)", 0, ""], ["ijepa", "Reachable", "", 23643, "facebook/ijepa_vith14_1k"], ["imagegpt", "Reachable", "", 7911, "openai/imagegpt-small"], ["informer", "Unreachable", "synthetic inputs rejected", 30807, "huggingface/informer-tourism-monthly"], ["inkling_audio", "No checkpoint", "no Hub checkpoint has this model_type (documented thinkingmachines/Inkling is another type)", 0, ""], ["inkling_mm_model", "Done", "", 3938529, "thinkingmachines/Inkling"], ["inkling_text", "Unreachable", "no usable checkpoint", 33, "hf-tiny-v2/tiny-random-InklingTextModel"], ["inkling_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented thinkingmachines/Inkling is another type)", 0, ""], ["instructblip", "Unreachable", "forward fails after shrink", 28294, "Salesforce/instructblip-flan-t5-xl"], ["instructblip_qformer", "Unreachable", "component config only", 0, ""], ["instructblip_vision_model", "Unreachable", "component config only", 0, ""], ["instructblipvideo", "Unreachable", "no usable checkpoint", 64, "hf-tiny-v2/tiny-random-InstructBlipVideoForConditionalGeneration"], ["instructblipvideo_qformer", "No checkpoint", "no Hub checkpoint has this model_type (documented Salesforce/instructblip-flan-t5-xl is another type)", 0, ""], ["instructblipvideo_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (documented Salesforce/instructblip-flan-t5-xl is another type)", 0, ""], ["internvl", "Done", "", 873854, "OpenGVLab/InternVL3_5-GPT-OSS-20B-A4B-Preview-HF"], ["internvl_vision", "Reachable", "", 105, "hf-tiny-v2/tiny-random-InternVLVisionModel"], ["jais2", "Reachable", "", 5408, "inception42/Jais-2-8B-Chat"], ["jamba", "Done", "", 210865, "ai21labs/Jamba-tiny-dev"], ["janus", "Done", "", 9275, "deepseek-community/Janus-Pro-1B"], ["janus_vision_model", "Unreachable", "component config only", 0, ""], ["janus_vqgan", "Unreachable", "component config only", 0, ""], ["jetmoe", "Reachable", "", 5517, "jetmoe/jetmoe-8b"], ["jina_embeddings_v3", "Done", "", 245085, "jinaai/jina-embeddings-v3-hf"], ["kimi_k25", "Reachable", "", 1381590, "moonshotai/Kimi-K2.5"], ["kimi_k25_vision", "Unreachable", "no usable checkpoint", 26, "Aquiles-ai/MoonViT-3D"], ["kimi_linear", "Unreachable", "shrink fails", 214668, "moonshotai/Kimi-Linear-48B-A3B-Instruct"], ["kolibri1", "Reachable", "", 52602, "Aleph-Alpha/Kolibri-1"], ["kosmos-2", "Unreachable", "forward fails after shrink", 24555, "microsoft/kosmos-2-patch14-224"], ["kosmos-2.5", "Reachable", "", 7720, "microsoft/kosmos-2.5"], ["kosmos_2_5_text_model", "Unreachable", "component config only", 14, "Fireblossom/kosmos-2.5-4bit-text"], ["kosmos_2_5_vision_model", "Unreachable", "component config only", 15, "Fireblossom/kosmos-2.5-4bit-vision"], ["kosmos_2_text_model", "Unreachable", "component config only", 0, ""], ["kosmos_2_vision_model", "Unreachable", "component config only", 0, ""], ["kyutai_speech_to_text", "Unreachable", "forward fails after shrink", 8225, "kyutai/stt-2.6b-en-trfs"], ["laguna", "Done", "", 3313579, "poolside/Laguna-XS.2"], ["lasr_ctc", "Unreachable", "forward fails after shrink", 13038, "google/medasr"], ["lasr_encoder", "Unreachable", "no usable checkpoint", 40, "hf-tiny-v2/tiny-random-LasrEncoder"], ["layoutlm", "Done", "", 518133, "impira/layoutlm-document-qa"], ["layoutlmv2", "Unreachable", "model build fails", 440458, "microsoft/layoutxlm-base"], ["layoutlmv3", "Unreachable", "forward fails after shrink", 970269, "microsoft/layoutlmv3-base"], ["layoutxlm", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/layoutxlm-base is another type)", 0, ""], ["led", "Unreachable", "model build fails", 33694, "allenai/led-base-16384"], ["levit", "Reachable", "", 19387, "optimum-intel-internal-testing/tiny-random-LevitModel"], ["lfm2", "Done", "", 1255733, "LiquidAI/LFM2.5-230M-Base"], ["lfm2_moe", "Done", "", 688657, "lmstudio-community/LFM2-24B-A2B-MLX-4bit"], ["lfm2_vl", "Done", "", 199290, "LiquidAI/LFM2.5-VL-450M"], ["lightglue", "Unreachable", "forward fails after shrink", 22972, "ETH-CVG/lightglue_superpoint"], ["lighton_ocr", "Reachable", "", 606, "onnx-community/LightOnOCR-2-1B-ONNX"], ["lilt", "Unreachable", "forward fails after shrink", 81714, "SCUT-DLVCLab/lilt-roberta-en-base"], ["llama", "Done", "", 50186457, "meta-llama/Llama-3.1-8B-Instruct"], ["llama4", "Done", "", 373138, "meta-llama/Llama-4-Scout-17B-16E-Instruct"], ["llama4_text", "Reachable", "", 28667, "trl-internal-testing/tiny-Llama4ForCausalLM"], ["llama4_vision_model", "Unreachable", "checkpoint gated", 0, "meta-llama/Llama-4-Scout-17B-16E"], ["llava", "Done", "", 1766490, "llava-hf/llava-1.5-7b-hf"], ["llava_next", "Done", "", 667783, "llava-hf/llava-v1.6-mistral-7b-hf"], ["llava_next_video", "Done", "", 53334, "llava-hf/LLaVA-NeXT-Video-7B-hf"], ["llava_onevision", "Done", "", 262658, "llava-hf/llava-onevision-qwen2-0.5b-ov-hf"], ["longcat_flash", "Unreachable", "forward fails after shrink", 1566, "wangbo981016/longcat2.0-nvfp4"], ["longformer", "Done", "", 1478599, "allenai/longformer-base-4096"], ["longt5", "Done", "", 243258, "pszemraj/long-t5-tglobal-base-sci-simplify"], ["luke", "Done", "", 321239, "studio-ousia/luke-large-finetuned-tacred"], ["lw_detr", "Reachable", "", 13080, "AnnaZhang/lwdetr_small_60e_coco"], ["lw_detr_vit", "Unreachable", "backward fails", 57, "hf-tiny-v2/tiny-random-LwDetrViTBackbone"], ["lxmert", "Unreachable", "forward fails after shrink", 4977, "unc-nlp/lxmert-base-uncased"], ["m2m_100", "Done", "", 2425200, "facebook/nllb-200-distilled-600M"], ["mamba", "Done", "", 138353, "state-spaces/mamba-130m-hf"], ["mamba2", "Unreachable", "shrink fails", 1401, "limloop/whiff-mamba2-50M-v0.1"], ["marian", "Done", "", 5061374, "Helsinki-NLP/opus-mt-nl-en"], ["markuplm", "Reachable", "", 10766, "microsoft/markuplm-base"], ["mask2former", "Done", "", 1060987, "facebook/mask2former-swin-large-ade-semantic"], ["maskformer", "Reachable", "", 16159, "facebook/maskformer-swin-base-ade"], ["maskformer-swin", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/swin-tiny-patch4-window7-224 is another type)", 0, ""], ["mbart", "Done", "", 579384, "ai4bharat/IndicBART"], ["megatron-bert", "Reachable", "", 21960, "KBLab/robust-swedish-sentiment-multiclass"], ["mellum", "Reachable", "", 51107, "JetBrains/Mellum2-12B-A2.5B-Instruct"], ["metaclip_2", "Reachable", "", 92623, "facebook/metaclip-2-worldwide-b16"], ["metaclip_2_text_model", "Unreachable", "component config only", 0, ""], ["metaclip_2_vision_model", "Unreachable", "component config only", 0, ""], ["mgp-str", "Unreachable", "forward fails after shrink", 13719, "alibaba-damo/mgp-str-base"], ["mimi", "Unreachable", "forward fails after shrink", 184266, "kyutai/mimi"], ["mimi_decoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["mimi_encoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["mimo_v2_flash", "Reachable", "", 49711, "XiaomiMiMo/MiMo-V2-Flash"], ["minicpm3", "Done", "", 17559, "openbmb/MiniCPM3-4B"], ["minicpm_ocr", "Reachable", "", 2508, "openbmb/MiniCPM-OCR-1.0"], ["minicpm_ocr_vision", "Unreachable", "forward fails after shrink", 0, ""], ["minicpmv4_6", "Done", "", 372489, "openbmb/MiniCPM-V-4.6"], ["minicpmv4_6_vision", "Unreachable", "component config only", 0, ""], ["minicpmv4_7", "Unreachable", "no usable checkpoint", 775231, "openbmb/MiniCPM-V-4.7-35B-A3B"], ["minicpmv4_7_vision", "Unreachable", "no usable checkpoint", 28, "hf-tiny-v2/tiny-random-MiniCPMV4_7VisionModel"], ["minimax", "Reachable", "", 19547, "MiniMaxAI/MiniMax-Text-01-hf"], ["minimax_m2", "Done", "", 1886356, "MiniMaxAI/MiniMax-M2.5"], ["minimax_m3_vl", "Done", "", 384244, "MiniMaxAI/MiniMax-M3"], ["minimax_m3_vl_text", "Reachable", "", 314, "ProprietaryLegal/minimax-m3-legal-reap-0.25"], ["minimax_m3_vl_vision", "Unreachable", "component config only", 30, "immanuelpeter/MiniMax-M3-Vision"], ["ministral", "Unreachable", "no usable checkpoint", 1857, "saidutta69/Ministral-8B-Instruct-heretic"], ["ministral3", "Done", "", 500650, "nvidia/Nemotron-3-Embed-1B-BF16"], ["mistral", "Done", "", 10522940, "mistralai/Mistral-7B-Instruct-v0.2"], ["mistral3", "Done", "", 2910198, "lightonai/LightOnOCR-2-1B"], ["mistral4", "Reachable", "", 971, "onnx-internal-testing/tiny-random-Mistral4ForCausalLM"], ["mixtral", "Done", "", 1221154, "mistralai/Mixtral-8x7B-Instruct-v0.1"], ["mlcd", "Reachable", "", 6623, "DeepGlint-AI/mlcd-vit-bigG-patch14-336"], ["mlcd_vision_model", "Unreachable", "no usable checkpoint", 107, "hf-tiny-v2/tiny-random-MLCDVisionModel"], ["mllama", "Done", "", 128886, "meta-llama/Llama-3.2-11B-Vision-Instruct"], ["mllama_text_model", "Unreachable", "component config only", 637, "hf-tiny-v2/tiny-random-MllamaTextModel"], ["mllama_vision_model", "Unreachable", "component config only", 65, "hf-tiny-v2/tiny-random-MllamaVisionModel"], ["mm-grounding-dino", "Done", "", 77467, "openmmlab-community/mm_grounding_dino_large_all"], ["mobilebert", "Unreachable", "model build fails", 378729, "google/mobilebert-uncased"], ["mobilenet_v1", "Reachable", "", 22769, "optimum-intel-internal-testing/mobilenet_v1_0.75_192"], ["mobilenet_v2", "Done", "", 116493, "google/mobilenet_v2_1.0_224"], ["mobilevit", "Done", "", 62083, "apple/mobilevit-xx-small"], ["mobilevitv2", "Reachable", "", 4054, "apple/mobilevitv2-1.0-imagenet1k-256"], ["modernbert", "Done", "", 22009085, "ibm-granite/granite-embedding-small-english-r2"], ["modernbert-decoder", "Reachable", "", 13311, "blab-jhu/test-32m-dec"], ["modernvbert", "Reachable", "", 477, "naver/v-splade-quality"], ["molmo2", "Unreachable", "shrink fails", 124222, "allenai/Molmo2-8B"], ["molmo2_adapter", "No checkpoint", "no Hub checkpoint has this model_type (documented allenai/Molmo2-8B is another type)", 0, ""], ["molmo2_text", "Reachable", "", 0, ""], ["molmo2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented allenai/Molmo2-8B is another type)", 0, ""], ["moonshine", "Unreachable", "forward fails after shrink", 120877, "moonshine-ai/moonshine-tiny"], ["moonshine_streaming", "Unreachable", "forward fails after shrink", 167472, "moonshine-ai/moonshine-streaming-tiny"], ["moonshine_streaming_encoder", "Unreachable", "component config only", 0, ""], ["moshi", "Reachable", "", 8294, "kyutai/personaplex-rl-seamless"], ["moshi_depth", "Unreachable", "component config only", 0, ""], ["moss_transcribe_diarize", "Done", "", 139891, "OpenMOSS-Team/MOSS-Transcribe-Diarize"], ["moss_transcribe_diarize_audio", "Reachable", "", 0, ""], ["mpnet", "Done", "", 23509384, "sentence-transformers/all-mpnet-base-v2"], ["mpt", "Reachable", "", 84471, "echarlaix/tiny-mpt-random-remote-code"], ["mra", "Reachable", "", 3305, "uw-madison/mra-base-512-4"], ["mt5", "Done", "", 725133, "google/mt5-small"], ["muse_glimmer", "Done", "", 688700, "RadixArk/Muse-Glimmer-NVFP4"], ["muse_glimmer_assistant", "Unreachable", "synthetic inputs rejected", 220523, "meta-models/Muse-Glimmer-30B-assistant"], ["muse_glimmer_text", "Reachable", "", 210, "OpensourceWTF/Muse-Glimmer-30B-MLX-q4"], ["muse_glimmer_vision", "Unreachable", "forward fails after shrink", 54, "immanuelpeter/Muse-Glimmer-Vision"], ["muse_spark", "Reachable", "", 48269, "meta-models/Muse-Spark-Open-816B-A42B-NVFP4"], ["muse_spark_audio", "Unreachable", "checkpoint gated", 0, "meta-models/Muse-Spark-Open-816B-A42B"], ["muse_spark_text", "Unreachable", "checkpoint gated", 0, "meta-models/Muse-Spark-Open-816B-A42B"], ["muse_spark_vision", "Unreachable", "checkpoint gated", 0, "meta-models/Muse-Spark-Open-816B-A42B"], ["musicflamingo", "Reachable", "", 18191, "nvidia/music-flamingo-2601-hf"], ["musicgen", "Unreachable", "model build fails", 2150786, "facebook/musicgen-medium"], ["musicgen_decoder", "Unreachable", "component config only", 13, "hf-tiny-v2/tiny-random-MusicgenForCausalLM"], ["musicgen_melody", "Unreachable", "model build fails", 12068, "facebook/musicgen-melody"], ["musicgen_melody_decoder", "Unreachable", "component config only", 39, "ylacombe/musicgen-melody-bella-ciao"], ["mvp", "Reachable", "", 4701, "RUCAIBox/mvp"], ["nanochat", "Reachable", "", 5316, "l0ulan/chinese-babylm-nanochat-d22-step10000"], ["nemo_speechlm", "Unreachable", "shrink fails", 5799, "nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B"], ["nemotron", "Reachable", "", 48968, "nvidia/Nemotron-Mini-4B-Instruct"], ["nemotron3_5_asr", "Unreachable", "forward fails after shrink", 842520, "nvidia/nemotron-3.5-asr-streaming-0.6b"], ["nemotron3_5_transcribe_asr_audio", "Unreachable", "forward fails after shrink", 0, ""], ["nemotron3_5_transcribe_audio", "Unreachable", "shrink fails", 0, ""], ["nemotron3_5_transcribe_word_timestamp", "Unreachable", "synthetic inputs rejected", 0, ""], ["nemotron3_diarization", "Unreachable", "shrink fails", 146475, "nvidia/Nemotron-3-Diarization"], ["nemotron3_diarization_audio", "Unreachable", "forward fails after shrink", 34, "hf-tiny-v2/tiny-random-Nemotron3DiarizationAudioModel"], ["nemotron_asr_streaming", "Unreachable", "forward fails after shrink", 103140, "nvidia/nemotron-speech-streaming-en-0.6b"], ["nemotron_asr_streaming_encoder", "Unreachable", "forward fails after shrink", 6, "hf-tiny-v2/tiny-random-NemotronAsrStreamingEncoder"], ["nemotron_h", "Done", "", 12146358, "nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16"], ["nemotron_h_omni", "Unreachable", "no usable checkpoint", 113462, "nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16"], ["neomme", "Reachable", "", 6699, "Hcompany/NeoMME-260M"], ["neucodec", "Unreachable", "forward fails after shrink", 48553, "neuphonic/neucodec"], ["nllb-moe", "Reachable", "", 3303, "facebook/nllb-moe-54b"], ["nomic_bert", "Done", "", 14371774, "nomic-ai/nomic-embed-text-v1.5"], ["nougat", "Unreachable", "config load fails", 8, "zphilip48/nougat-middle-cn"], ["nystromformer", "Reachable", "", 5529, "uw-madison/nystromformer-512"], ["olmo", "Done", "", 125508, "optimum-intel-internal-testing/tiny-random-olmo-hf"], ["olmo2", "Done", "", 442211, "allenai/OLMo-2-0425-1B"], ["olmo3", "Done", "", 1055647, "allenai/Olmo-3-7B-Think"], ["olmo_hybrid", "Reachable", "", 19286, "allenai/Olmo-Hybrid-7B"], ["olmoe", "Done", "", 132339, "allenai/OLMoE-1B-7B-0924"], ["omdet-turbo", "Unreachable", "config load fails", 35542, "omlab/omdet-turbo-swin-tiny-hf"], ["oneformer", "Unreachable", "synthetic inputs rejected", 390959, "shi-labs/oneformer_cityscapes_swin_large"], ["openai-gpt", "Reachable", "", 85250, "openai-community/openai-gpt"], ["openai_privacy_filter", "Done", "", 634682, "OpenMed/privacy-filter-nemotron-v2"], ["opt", "Done", "", 5616003, "facebook/opt-125m"], ["ovis2", "Reachable", "", 6898, "thisisiron/Ovis2-1B-hf"], ["ovis2_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented thisisiron/Ovis2-1B-hf is another type)", 0, ""], ["owlv2", "Done", "", 1501028, "google/owlv2-base-patch16-ensemble"], ["owlv2_text_model", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["owlv2_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["owlvit", "Done", "", 220839, "google/owlvit-base-patch32"], ["owlvit_text_model", "Unreachable", "component config only", 0, ""], ["owlvit_vision_model", "Unreachable", "component config only", 0, ""], ["paddleocr_vl", "Done", "", 227189, "BDRC/tibetan-ocr"], ["paddleocr_vl_text", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PaddleOCR-VL is another type)", 0, ""], ["paddleocr_vl_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PaddleOCR-VL is another type)", 0, ""], ["paligemma", "Done", "", 490686, "google/paligemma-3b-pt-224"], ["parakeet_ctc", "Unreachable", "forward fails after shrink", 105992, "nvidia/parakeet-ctc-1.1b"], ["parakeet_encoder", "Unreachable", "forward fails after shrink", 125, "nithinraok/parakeet-tdt-v3-encoder"], ["parakeet_rnnt", "Unreachable", "forward fails after shrink", 40081, "nvidia/parakeet-rnnt-0.6b"], ["parakeet_tdt", "Unreachable", "forward fails after shrink", 694911, "nvidia/parakeet-tdt-0.6b-v3"], ["patchtsmixer", "Unreachable", "synthetic inputs rejected", 15970, "ibm-granite/granite-timeseries-patchtsmixer"], ["patchtst", "Unreachable", "synthetic inputs rejected", 28739, "GilpinLab/panda"], ["pe_audio", "Unreachable", "no usable checkpoint", 89572, "facebook/pe-a-frame-large"], ["pe_audio_encoder", "Unreachable", "forward fails after shrink", 31, "hf-tiny-v2/tiny-random-PeAudioEncoder"], ["pe_audio_video", "Unreachable", "forward fails after shrink", 21786, "facebook/pe-av-large"], ["pe_audio_video_encoder", "Unreachable", "forward fails after shrink", 21, "hf-tiny-v2/tiny-random-PeAudioVideoEncoder"], ["pe_video", "Unreachable", "no usable checkpoint", 25, "hf-tiny-v2/tiny-random-PeVideoModel"], ["pe_video_encoder", "Unreachable", "synthetic inputs rejected", 25, "hf-tiny-v2/tiny-random-PeVideoEncoder"], ["pegasus", "Done", "", 120276, "optimum-intel-internal-testing/tiny-random-pegasus"], ["pegasus_x", "Reachable", "", 6499, "google/pegasus-x-large"], ["perceiver", "Done", "", 41482, "optimum-intel-internal-testing/tiny-random-language_perceiver"], ["perception_lm", "Reachable", "", 1028, "facebook/Perception-LM-8B"], ["persimmon", "Reachable", "", 50277, "optimum-intel-internal-testing/tiny-random-PersimmonForCausalLM"], ["phi", "Done", "", 575882, "microsoft/phi-2"], ["phi3", "Done", "", 2440907, "microsoft/Phi-3.5-mini-instruct"], ["phi4_multimodal", "Done", "", 549, "huggingface-mic/Phi-4-multimodal-instruct-mic"], ["phi4_multimodal_audio", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/Phi-4-multimodal-instruct is another type)", 0, ""], ["phi4_multimodal_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/Phi-4-multimodal-instruct is another type)", 0, ""], ["phimoe", "Done", "", 169903, "optimum-intel-internal-testing/phi-3.5-moe-tiny-random"], ["pi0", "Unreachable", "synthetic inputs rejected", 181, "yinchenghust/openpi_libero_all"], ["pix2struct", "Unreachable", "synthetic inputs rejected", 116238, "optimum-intel-internal-testing/pix2struct-tiny-random"], ["pix2struct_text_model", "Unreachable", "component config only", 0, ""], ["pix2struct_vision_model", "Unreachable", "component config only", 22, "UiPath/pix2struct-vision-base"], ["pixio", "Reachable", "", 163, "LiheYoung/pixio-vith16"], ["pixtral", "Reachable", "", 940, "mlx-community/pixtral-12b-4bit"], ["plbart", "Reachable", "", 11136, "uclanlp/plbart-base"], ["pocket_tts", "Unreachable", "shrink fails", 2949, "mlx-community/pocket-tts"], ["pocket_tts_audio_tokenizer", "Unreachable", "forward fails after shrink", 471, "gradium/phonon-mimi"], ["pocket_tts_audio_tokenizer_decoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["pocket_tts_audio_tokenizer_encoder", "No checkpoint", "no Hub checkpoint has this model_type (none documented)", 0, ""], ["poolformer", "Reachable", "", 22043, "optimum-intel-internal-testing/tiny-random-PoolFormerModel"], ["pop2piano", "Reachable", "", 105, "hf-tiny-v2/tiny-random-Pop2PianoForConditionalGeneration"], ["pp_chart2table", "Reachable", "", 272, "PaddlePaddle/PP-Chart2Table_safetensors"], ["pp_chart2table_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["pp_doclayout_v2", "Unreachable", "forward fails after shrink", 585, "PaddlePaddle/PP-DocLayoutV2_safetensors"], ["pp_doclayout_v3", "Unreachable", "forward fails after shrink", 932166, "PaddlePaddle/PP-DocLayoutV3_safetensors"], ["pp_doclayout_v4", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PP-DocLayoutV4_safetensors does not exist)", 0, ""], ["pp_formulanet", "Unreachable", "forward fails after shrink", 3829, "PaddlePaddle/PP-FormulaNet_plus-L_safetensors"], ["pp_formulanet_text", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PP-FormulaNet_plus-L_safetensors is another type)", 0, ""], ["pp_formulanet_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/PP-FormulaNet_plus-L_safetensors is another type)", 0, ""], ["pp_lcnet", "Reachable", "", 865, "PaddlePaddle/PP-LCNet_x1_0_doc_ori_safetensors"], ["pp_lcnet_v3", "Unreachable", "no usable checkpoint", 112, "hf-tiny-v2/tiny-random-PPLCNetV3Backbone"], ["pp_lcnet_v4", "Unreachable", "no usable checkpoint", 21, "hf-tiny-v2/tiny-random-PPLCNetV4Backbone"], ["pp_ocrv5_mobile_det", "Reachable", "", 530, "PaddlePaddle/PP-OCRv5_mobile_det_safetensors"], ["pp_ocrv5_mobile_rec", "Reachable", "", 3902, "PaddlePaddle/latin_PP-OCRv5_mobile_rec_safetensors"], ["pp_ocrv5_server_det", "Reachable", "", 509, "PaddlePaddle/PP-OCRv5_server_det_safetensors"], ["pp_ocrv5_server_rec", "Reachable", "", 585, "PaddlePaddle/PP-OCRv5_server_rec_safetensors"], ["pp_ocrv6_medium_det", "Reachable", "", 2694, "PaddlePaddle/PP-OCRv6_medium_det_safetensors"], ["pp_ocrv6_small_det", "Reachable", "", 1617, "PaddlePaddle/PP-OCRv6_small_det_safetensors"], ["pp_ocrv6_small_rec", "Reachable", "", 3559, "PaddlePaddle/PP-OCRv6_medium_rec_safetensors"], ["pp_ocrv6_tiny_rec", "Reachable", "", 693, "PaddlePaddle/PP-OCRv6_tiny_rec_safetensors"], ["prompt_depth_anything", "Unreachable", "forward fails after shrink", 1530, "depth-anything/prompt-depth-anything-vits-hf"], ["prophetnet", "Reachable", "", 7258, "microsoft/prophetnet-large-uncased"], ["pvt", "Done", "", 8712, "Xrenya/pvt-tiny-224"], ["pvt_v2", "Reachable", "", 9824, "OpenGVLab/pvt_v2_b0"], ["qianfan_ocr", "Reachable", "", 38183, "baidu/Qianfan-OCR"], ["qianfan_ocr_vision", "Reachable", "", 77, "hf-tiny-v2/tiny-random-QianfanOCRVisionModel"], ["qwen2", "Done", "", 52739906, "Qwen/Qwen2.5-7B-Instruct"], ["qwen2_5_omni", "Done", "", 456813, "Qwen/Qwen2.5-Omni-7B"], ["qwen2_5_omni_audio_encoder", "Unreachable", "component config only", 23, "FlexiSLM/Qwen2_5-Omni-Audio_Encoder"], ["qwen2_5_omni_bigvgan", "Unreachable", "component config only", 24, "Ryan241005/scail-2"], ["qwen2_5_omni_dit", "Unreachable", "component config only", 0, ""], ["qwen2_5_omni_talker", "Unreachable", "component config only", 0, ""], ["qwen2_5_omni_text", "Unreachable", "component config only", 59, "AlexHung29629/Qwen-2.5-Omni-ThinkerTextModel"], ["qwen2_5_omni_thinker", "Done", "", 128795, "laion/voiceclap-large-v2"], ["qwen2_5_omni_token2wav", "Unreachable", "component config only", 0, ""], ["qwen2_5_omni_vision_encoder", "Unreachable", "component config only", 0, ""], ["qwen2_5_vl", "Done", "", 15195814, "Qwen/Qwen2.5-VL-7B-Instruct"], ["qwen2_5_vl_text", "Reachable", "", 83, "hf-tiny-v2/tiny-random-Qwen2_5_VLTextModel"], ["qwen2_5_vl_vision", "Unreachable", "no usable checkpoint", 22, "hf-tiny-v2/tiny-random-Qwen2_5_VisionTransformerPretrainedModel"], ["qwen2_audio", "Done", "", 199606, "Qwen/Qwen2-Audio-7B-Instruct"], ["qwen2_audio_encoder", "Reachable", "", 200, "Atotti/qwen2-audio-encoder"], ["qwen2_moe", "Done", "", 194932, "Qwen/Qwen1.5-MoE-A2.7B"], ["qwen2_vl", "Done", "", 3757282, "Qwen/Qwen2-VL-7B-Instruct-AWQ"], ["qwen2_vl_text", "Reachable", "", 72, "hf-tiny-v2/tiny-random-Qwen2VLTextModel"], ["qwen2_vl_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented Qwen/Qwen2-VL-7B-Instruct is another type)", 0, ""], ["qwen3", "Done", "", 108746675, "Qwen/Qwen3-0.6B"], ["qwen3_5", "Done", "", 92274452, "Qwen/Qwen3.8-27B"], ["qwen3_5_moe", "Done", "", 34351908, "Qwen/Qwen3.6-35B-A3B"], ["qwen3_5_moe_text", "Done", "", 717113, "Bahushruth/Qwen3.6-35B-A3B-abliterated-v4"], ["qwen3_5_moe_vision", "Unreachable", "no usable checkpoint", 16, "hf-tiny-v2/tiny-random-Qwen3_5MoeVisionModel"], ["qwen3_5_text", "Done", "", 2017293, "autotrust/JEV-9B"], ["qwen3_5_vision", "Unreachable", "forward fails after shrink", 135, "immanuelpeter/Qwen3.8-27B-Vision"], ["qwen3_asr", "Done", "", 4105092, "Qwen/Qwen3-ASR-1.7B"], ["qwen3_asr_encoder", "Unreachable", "forward fails after shrink", 17, "hf-tiny-v2/tiny-random-Qwen3ASREncoder"], ["qwen3_moe", "Done", "", 8500460, "Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8"], ["qwen3_next", "Done", "", 2235316, "Qwen/Qwen3-Coder-Next-FP8"], ["qwen3_omni_moe", "Unreachable", "forward fails after shrink", 1249447, "Qwen/Qwen3-Omni-30B-A3B-Instruct"], ["qwen3_omni_moe_audio_encoder", "Unreachable", "component config only", 1968, "Atotti/Qwen3-Omni-AudioTransformer"], ["qwen3_omni_moe_talker_code_predictor", "Unreachable", "component config only", 0, ""], ["qwen3_omni_moe_talker_text", "No checkpoint", "no Hub checkpoint has this model_type (documented Qwen/Qwen3-Omni-30B-A3B-Instruct is another type)", 0, ""], ["qwen3_omni_moe_text", "Unreachable", "component config only", 0, ""], ["qwen3_omni_moe_thinker", "Reachable", "", 298, "hf-tiny-v2/tiny-random-Qwen3OmniMoeThinkerForConditionalGeneration"], ["qwen3_omni_moe_vision_encoder", "Unreachable", "component config only", 0, ""], ["qwen3_vl", "Done", "", 19337885, "Qwen/Qwen3-VL-8B-Instruct"], ["qwen3_vl_moe", "Done", "", 1686056, "Qwen/Qwen3-VL-30B-A3B-Instruct"], ["qwen3_vl_moe_text", "Reachable", "", 63, "hf-tiny-v2/tiny-random-Qwen3VLMoeTextModel"], ["qwen3_vl_moe_vision", "Unreachable", "no usable checkpoint", 54, "hf-tiny-v2/tiny-random-Qwen3VLMoeVisionModel"], ["qwen3_vl_text", "Reachable", "", 587, "DuoNeural/Cosmos3-Nano-GPTQ-4bit"], ["qwen3_vl_vision", "Unreachable", "forward fails after shrink", 221, "madrisight/madrimed1.2-vision-embed"], ["qwen4_exp", "Done", "", 5763811, "Qwen/Qwen3.8-Flash-Next"], ["qwen4_exp_text", "Reachable", "", 17917, "Jab1718/qwen3.8-flash-coder-44gb-selective-int8"], ["qwen4_exp_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented Qwen/Qwen4-Exp does not exist)", 0, ""], ["radio", "Unreachable", "no usable checkpoint", 3, "hf-tiny-v2/tiny-random-RadioModel"], ["rag", "Unreachable", "no usable checkpoint", 3924, "facebook/rag-token-nq"], ["recurrent_gemma", "Done", "", 4729, "google/recurrentgemma-2b-it"], ["reformer", "Unreachable", "model build fails", 17043, "google/reformer-crime-and-punishment"], ["regnet", "Reachable", "", 7237, "facebook/regnet-y-040"], ["rembert", "Reachable", "", 11890, "google/rembert"], ["resnet", "Done", "", 1174379, "microsoft/resnet-18"], ["rf_detr", "Unreachable", "forward fails after shrink", 28618, "Roboflow/rf-detr-base"], ["rf_detr_dinov2", "Unreachable", "forward fails after shrink", 70, "hf-tiny-v2/tiny-random-RfDetrDinov2Backbone"], ["roberta", "Done", "", 46597117, "FacebookAI/roberta-large"], ["roberta-prelayernorm", "Reachable", "", 4655, "andreasmadsen/efficient_mlm_m0.40"], ["roc_bert", "Reachable", "", 12437, "weiweishi/roc-bert-base-zh"], ["roformer", "Reachable", "", 45159, "optimum-intel-internal-testing/tiny-random-roformer"], ["rt_detr", "Done", "", 1687337, "PekingU/rtdetr_r101vd_coco_o365"], ["rt_detr_resnet", "Unreachable", "no usable checkpoint", 54, "hf-tiny-v2/tiny-random-RTDetrResNetBackbone"], ["rt_detr_v2", "Done", "forward fails after shrink", 2526763, "docling-project/docling-layout-heron"], ["rwkv", "Reachable", "", 14762, "RWKV/rwkv-4-169m-pile"], ["sam", "Unreachable", "forward fails after shrink", 976537, "facebook/sam-vit-base"], ["sam2", "Done", "", 19375, "onnx-community/sam2.1-hiera-tiny-ONNX"], ["sam2_hiera_det_model", "Unreachable", "forward fails after shrink", 59, "hf-tiny-v2/tiny-random-Sam2HieraDetModel"], ["sam2_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2.1-hiera-tiny is another type)", 0, ""], ["sam2_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2.1-hiera-tiny is another type)", 0, ""], ["sam2_video", "Unreachable", "forward fails after shrink", 701203, "facebook/sam2.1-hiera-large"], ["sam2_video_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2_video.1-hiera-tiny does not exist)", 0, ""], ["sam2_video_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam2_video.1-hiera-tiny does not exist)", 0, ""], ["sam2_vision_model", "Unreachable", "forward fails after shrink", 57, "hf-tiny-v2/tiny-random-Sam2VisionModel"], ["sam3", "Unreachable", "forward fails after shrink", 3530, "unsloth/sam3.1"], ["sam3_detr_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_detr_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_geometry_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_lite_text", "Unreachable", "forward fails after shrink", 14775, "vil-uob/sam3-litetext-s0"], ["sam3_lite_text_detr_decoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_detr_encoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_geometry_encoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_mask_decoder", "Unreachable", "component config only", 0, ""], ["sam3_lite_text_text_model", "Unreachable", "forward fails after shrink", 39, "hf-tiny-v2/tiny-random-Sam3LiteTextTextModel"], ["sam3_mask_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker", "Unreachable", "no usable checkpoint", 1300, "onnx-community/sam3-tracker-ONNX"], ["sam3_tracker_mask_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_prompt_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_video", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_video_mask_decoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_tracker_video_prompt_encoder", "Unreachable", "checkpoint gated", 0, "facebook/sam3"], ["sam3_video", "Unreachable", "forward fails after shrink", 2087329, "facebook/sam3"], ["sam3_vision_model", "Unreachable", "no usable checkpoint", 38, "hf-tiny-v2/tiny-random-Sam3VisionModel"], ["sam3_vit_model", "Unreachable", "no usable checkpoint", 37, "hf-tiny-v2/tiny-random-Sam3ViTModel"], ["sam_hq", "Unreachable", "forward fails after shrink", 16860, "syscv-community/sam-hq-vit-base"], ["sam_hq_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented syscv-community/sam-hq-vit-base is another type)", 0, ""], ["sam_hq_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented syscv-community/sam-hq-vit-base is another type)", 0, ""], ["sam_hq_vision_model", "Reachable", "", 138, "hf-tiny-v2/tiny-random-SamHQVisionModel"], ["sam_mask_decoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["sam_prompt_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented facebook/sam-vit-huge is another type)", 0, ""], ["sam_vision_model", "Reachable", "", 138, "hf-tiny-v2/tiny-random-SamVisionModel"], ["sapiens2", "Unreachable", "forward fails after shrink", 103690, "facebook/sapiens2-pose-1b"], ["sapiens2_head", "Unreachable", "component config only", 0, ""], ["seamless_m4t", "Reachable", "", 13797, "facebook/hf-seamless-m4t-medium"], ["seamless_m4t_v2", "Done", "", 316927, "facebook/seamless-m4t-v2-large"], ["seed_oss", "Done", "", 128943, "ByteDance-Seed/Seed-OSS-36B-Instruct"], ["segformer", "Done", "", 2291917, "mattmdjaga/segformer_b2_clothes"], ["seggpt", "Unreachable", "forward fails after shrink", 21384, "BAAI/seggpt-vit-large"], ["sew", "Reachable", "", 15680, "optimum-intel-internal-testing/tiny-random-SEWModel"], ["sew-d", "Reachable", "", 18152, "optimum-intel-internal-testing/sew-d-tiny-100k-ft-ls100h"], ["shieldgemma2", "Reachable", "", 6627, "google/shieldgemma-2-4b-it"], ["siglip", "Done", "", 18018379, "prithivMLmods/Gender-Classifier-Mini"], ["siglip2", "Unreachable", "forward fails after shrink", 1622065, "google/siglip2-base-patch16-naflex"], ["siglip2_text_model", "Unreachable", "component config only", 56, "sortify-ai/siglip2-so400m-patch14-384-text"], ["siglip2_vision_model", "Unreachable", "forward fails after shrink", 58, "hf-tiny-v2/tiny-random-Siglip2VisionModel"], ["siglip_text_model", "Unreachable", "component config only", 19, "carsonpoole/binary-siglip-text"], ["siglip_vision_model", "Reachable", "", 1842, "zchuning/pi05-siglip-vision-tower"], ["slanet", "Reachable", "", 6957, "PaddlePaddle/SLANet_plus_safetensors"], ["slanext", "Reachable", "", 212, "PaddlePaddle/SLANeXt_wired_safetensors"], ["slanext_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented PaddlePaddle/SLANeXt_wired_safetensors is another type)", 0, ""], ["smollm3", "Done", "", 1135357, "HuggingFaceTB/SmolLM3-3B"], ["smolvlm", "Done", "", 1742362, "HuggingFaceTB/SmolVLM2-500M-Video-Instruct"], ["smolvlm_vision", "Unreachable", "forward fails after shrink", 5, "hf-tiny-v2/tiny-random-SmolVLMVisionTransformer"], ["solar_open", "Reachable", "", 32622, "upstage/Solar-Open-100B"], ["speech-encoder-decoder", "Unreachable", "synthetic inputs rejected", 5244, "hf-tiny-v2/tiny-random-SpeechEncoderDecoderModel-wav2vec2-bert"], ["speech_to_text", "Unreachable", "forward fails after shrink", 23544, "facebook/s2t-small-librispeech-asr"], ["speecht5", "Done", "", 194296, "optimum-intel-internal-testing/tiny-random-SpeechT5ForTextToSpeech"], ["speecht5_hifigan", "No checkpoint", "no Hub checkpoint has this model_type (documented microsoft/speecht5_asr is another type)", 0, ""], ["splinter", "Reachable", "", 3741, "tau/splinter-base"], ["squeezebert", "Unreachable", "model build fails", 33330, "optimum-intel-internal-testing/tiny-random-squeezebert"], ["stablelm", "Done", "", 169690, "optimum-intel-internal-testing/tiny-random-StableLmForCausalLM"], ["starcoder2", "Done", "", 152451, "optimum-intel-internal-testing/tiny-random-Starcoder2ForCausalLM"], ["step3p5", "Unreachable", "shrink fails", 266885, "stepfun-ai/Step-3.5-Flash"], ["step3p5_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented stepfun-ai/Step-3.7-Flash is another type)", 0, ""], ["step3p7", "Unreachable", "shrink fails", 26135, "stepfun-ai/Step-3.7-Flash"], ["superglue", "Unreachable", "forward fails after shrink", 8964, "magic-leap-community/superglue_indoor"], ["superpoint", "Unreachable", "backward fails", 9952, "magic-leap-community/superpoint"], ["swiftformer", "Reachable", "", 7504, "MBZUAI/swiftformer-xs"], ["swin", "Done", "", 238869, "Organika/sdxl-detector"], ["swin2sr", "Reachable", "", 56057, "caidas/swin2SR-classical-sr-x2-64"], ["swinv2", "Done", "", 434756, "microsoft/swinv2-tiny-patch4-window16-256"], ["switch_transformers", "Reachable", "", 10063, "google/switch-base-8"], ["t5", "Done", "", 94678669, "google-t5/t5-small"], ["t5_gemma_module", "Unreachable", "component config only", 70, "RE-N-Y/t5gemma-2b-2b-ul2-it"], ["t5gemma", "Unreachable", "no usable checkpoint", 176067, "google/t5gemma-2b-2b-ul2-it"], ["t5gemma2", "Unreachable", "forward fails after shrink", 114374, "google/t5gemma-2-270m-270m"], ["t5gemma2_decoder", "Unreachable", "component config only", 0, ""], ["t5gemma2_encoder", "Reachable", "", 149, "hf-tiny-v2/tiny-random-T5Gemma2Encoder"], ["t5gemma2_text", "Unreachable", "component config only", 30, "amanchoudhri/t5-gemma-2-270m-text-encoder"], ["table-transformer", "Done", "", 1872765, "microsoft/table-transformer-structure-recognition"], ["tapas", "Reachable", "", 37821, "google/tapas-base-finetuned-wtq"], ["textnet", "Unreachable", "forward fails after shrink", 7475, "czczup/textnet-base"], ["time_series_transformer", "Unreachable", "synthetic inputs rejected", 4534, "huggingface/time-series-transformer-tourism-monthly"], ["timesfm", "Done", "", 1854129, "google/timesfm-2.5-200m-pytorch"], ["timesfm2_5", "Unreachable", "synthetic inputs rejected", 37239, "google/timesfm-2.5-200m-transformers"], ["timesformer", "Unreachable", "forward fails after shrink", 25737, "facebook/timesformer-base-finetuned-k400"], ["timm_backbone", "Unreachable", "no usable checkpoint", 133, "hf-tiny-v2/tiny-random-TimmBackbone"], ["timm_wrapper", "Unreachable", "forward fails after shrink", 383060, "Freepik/nsfw_image_detector"], ["tipsv2", "Done", "", 289583, "google/tipsv2-so400m14"], ["tipsv2_dpt", "Unreachable", "forward fails after shrink", 8014, "google/tipsv2-b14-dpt"], ["tipsv2_text_model", "Reachable", "", 4, "hf-tiny-v2/tiny-random-Tipsv2TextModel"], ["tipsv2_vision_model", "Reachable", "", 8, "hf-tiny-v2/tiny-random-Tipsv2VisionBackbone"], ["trocr", "Reachable", "", 53, "sasmithasarasi/trocr-sinhala-handwriting"], ["tvp", "Unreachable", "forward fails after shrink", 13994, "Intel/tvp-base"], ["udop", "Unreachable", "forward fails after shrink", 9051, "microsoft/udop-large"], ["ultrabert", "Reachable", "", 29547, "dmis-lab/UltraBERT-380M"], ["ultravbert", "Reachable", "", 1133, "athrael-soju/ColUltraVBERT"], ["umt5", "Done", "", 5305, "huggingface-mic/umt5-small-mic"], ["unispeech", "Reachable", "", 24130, "optimum-intel-internal-testing/tiny-random-unispeech"], ["unispeech-sat", "Reachable", "", 38482, "optimum-intel-internal-testing/tiny-random-UnispeechSatModel"], ["univnet", "Unreachable", "forward fails after shrink", 5412, "dg845/univnet-dev"], ["unlimited_ocr", "Done", "", 3535621, "baidu/Unlimited-OCR"], ["unlimited_ocr_sam_vision_model", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["unlimited_ocr_text", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["unlimited_ocr_vision", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["unlimited_ocr_vision_encoder", "No checkpoint", "no Hub checkpoint has this model_type (documented baidu/Unlimited-OCR is another type)", 0, ""], ["upernet", "Unreachable", "forward fails after shrink", 46968, "openmmlab/upernet-convnext-small"], ["uvdoc", "Reachable", "", 431, "PaddlePaddle/UVDoc_safetensors"], ["uvdoc_backbone", "Unreachable", "backward fails", 39, "hf-tiny-v2/tiny-random-UVDocBackbone"], ["vaultgemma", "Reachable", "", 2790, "google/vaultgemma-1b"], ["vibevoice", "Done", "", 1070414, "microsoft/VibeVoice-ASR"], ["vibevoice_acoustic_tokenizer", "Unreachable", "forward fails after shrink", 2151, "microsoft/VibeVoice-AcousticTokenizer"], ["vibevoice_acoustic_tokenizer_decoder", "Unreachable", "no usable checkpoint", 33, "hf-tiny-v2/tiny-random-VibeVoiceAcousticTokenizerDecoderModel"], ["vibevoice_acoustic_tokenizer_encoder", "Unreachable", "no usable checkpoint", 49, "hf-tiny-v2/tiny-random-VibeVoiceAcousticTokenizerEncoderModel"], ["vibevoice_asr", "Reachable", "", 59115, "microsoft/VibeVoice-ASR-HF"], ["vibevoice_diffusion", "No checkpoint", "no Hub checkpoint has this model_type (documented vibevoice/VibeVoice-1.5B-hf is another type)", 0, ""], ["video_llama_3", "Reachable", "", 11087, "lkhl/VideoLLaMA3-2B-Image-HF"], ["video_llama_3_vision", "Unreachable", "forward fails after shrink", 32, "hf-tiny-v2/tiny-random-VideoLlama3VisionModel"], ["video_llava", "Reachable", "", 10045, "LanguageBind/Video-LLaVA-7B-hf"], ["videomae", "Unreachable", "forward fails after shrink", 475037, "MCG-NJU/videomae-base"], ["videomt", "Unreachable", "synthetic inputs rejected", 5328, "tue-mps/videomt-dinov2-small-ytvis2019"], ["videoprism", "Unreachable", "synthetic inputs rejected", 545, "MHRDYN7/videoprism-lvt-base-f16r288"], ["videoprism_text_model", "No checkpoint", "no Hub checkpoint has this model_type (documented google/videoprism-lvt-base-f16r288 has no config.json)", 0, ""], ["videoprism_vision_model", "Unreachable", "synthetic inputs rejected", 191, "MHRDYN7/videoprism-base-f16r288"], ["vilt", "Unreachable", "forward fails after shrink", 89943, "dandelin/vilt-b32-finetuned-vqa"], ["vipllava", "Reachable", "", 7898, "llava-hf/vip-llava-7b-hf"], ["vision-encoder-decoder", "Unreachable", "forward fails after shrink", 2482977, "kha-white/manga-ocr-base"], ["vision-text-dual-encoder", "Unreachable", "model build fails", 3250, "hf-tiny-v2/tiny-random-VisionTextDualEncoderModel-vit-bert"], ["visual_bert", "Reachable", "", 5237, "uclanlp/visualbert-vqa-coco-pre"], ["vit", "Done", "", 23571533, "google/vit-base-patch16-224"], ["vit_mae", "Reachable", "", 43367, "facebook/vit-mae-base"], ["vit_msn", "Reachable", "", 1782, "facebook/vit-msn-small"], ["vitdet", "Unreachable", "no usable checkpoint", 214, "hf-tiny-v2/tiny-random-VitDetBackbone"], ["vitmatte", "Done", "", 383080, "hustvl/vitmatte-small-composition-1k"], ["vitpose", "Unreachable", "forward fails after shrink", 392468, "usyd-community/vitpose-base-simple"], ["vitpose_backbone", "Unreachable", "forward fails after shrink", 124, "hf-tiny-v2/tiny-random-VitPoseBackbone"], ["vits", "Done", "", 455247, "facebook/mms-tts-eng"], ["vivit", "Unreachable", "forward fails after shrink", 89874, "google/vivit-b-16x2"], ["vjepa2", "Unreachable", "synthetic inputs rejected", 426419, "facebook/vjepa2-vitl-fpc64-256"], ["voxtral", "Done", "", 319265, "mzbac/voxtral-mini-3b-4bit-mixed"], ["voxtral_encoder", "Reachable", "", 108, "hf-tiny-v2/tiny-random-VoxtralEncoder"], ["voxtral_realtime", "Unreachable", "forward fails after shrink", 1948459, "mistralai/Voxtral-Mini-4B-Realtime-2602"], ["voxtral_realtime_encoder", "Reachable", "", 80, "hf-tiny-v2/tiny-random-VoxtralRealtimeEncoder"], ["voxtral_realtime_text", "Unreachable", "model build fails", 10, "hf-tiny-v2/tiny-random-VoxtralRealtimeTextModel"], ["wav2vec2", "Done", "", 100317588, "jonatasgrosman/wav2vec2-large-xlsr-53-japanese"], ["wav2vec2-bert", "Unreachable", "model build fails", 1320062, "facebook/w2v-bert-2.0"], ["wav2vec2-conformer", "Unreachable", "model build fails", 160365, "facebook/wav2vec2-conformer-rope-large-960h-ft"], ["wavlm", "Done", "", 1990346, "microsoft/wavlm-large"], ["whisper", "Done", "", 22332636, "openai/whisper-large-v3-turbo"], ["xclip", "Unreachable", "forward fails after shrink", 453798, "microsoft/xclip-base-patch16-zero-shot"], ["xclip_text_model", "Unreachable", "component config only", 0, ""], ["xclip_vision_model", "Unreachable", "component config only", 0, ""], ["xcodec", "Unreachable", "config load fails", 14050, "hf-audio/xcodec-hubert-general"], ["xcodec2", "Unreachable", "forward fails after shrink", 9249, "HKUSTAudio/xcodec2-hf"], ["xglm", "Reachable", "", 64850, "optimum-intel-internal-testing/tiny-random-XGLMForCausalLM"], ["xlm", "Reachable", "", 19454, "FacebookAI/xlm-mlm-en-2048"], ["xlm-roberta", "Done", "", 104687870, "BAAI/bge-m3"], ["xlm-roberta-xl", "Done", "", 101085, "facebook/xlm-roberta-xl"], ["xlnet", "Unreachable", "shrink fails", 166190, "xlnet/xlnet-base-cased"], ["xlstm", "Unreachable", "config load fails", 4922, "NX-AI/xLSTM-7b"], ["xmod", "Unreachable", "forward fails after shrink", 29093, "facebook/xmod-base"], ["yolos", "Done", "", 859794, "hustvl/yolos-small"], ["yoso", "Reachable", "", 5281, "uw-madison/yoso-4096"], ["youtu", "Reachable", "", 12971, "tencent/Youtu-LLM-2B"], ["zamba", "Unreachable", "model build fails", 5935, "Zyphra/Zamba-7B-v1"], ["zamba2", "Unreachable", "forward fails after shrink", 70083, "Zyphra/Zamba2-1.2B-instruct"], ["zaya", "Reachable", "", 8978, "Zyphra/ZAYA1-8B"], ["zoedepth", "Unreachable", "forward fails after shrink", 48545, "Intel/zoedepth-nyu-kitti"]], "meta": {"generated": "2026-11-01", "fleet": "666d54fcc782e473ff276ec9626df0b0a3f0cafd", "transformers": "420dca92a4cf7d96dfbc1cc46f00375b0681ff8f", "sweep": "41e1eb5257d6a189eee599418fa256c51e49c63a"}, "shrink": {"albert": [["albert-base-v2-mic", 0]], "apertus": [["Apertus-v1.1-0.5B-Instruct-mic", 0]], "audioflamingo3": [["audio-flamingo-3-hf-mic", 0]], "bart": [["bart-base-mic", 0]], "beit": [["beit-base-patch16-224-mic", 0]], "bert": [["bert-base-uncased-mic", 0], ["bert-mic", 0]], "big_bird": [["bigbird-roberta-base-mic", 0]], "blenderbot": [["blenderbot-400M-distill-mic", 0]], "bloom": [["bloomz-560m-mic", 0]], "camembert": [["camembert-base-mic", 0]], "canine": [["canine-c-mic", 0]], "chinese_clip": [["chinese-clip-vit-base-patch16-mic", 0]], "clap": [["clap-htsat-fused-mic", 0]], "clip": [["clip-mic", 0], ["clip-vit-base-patch32-mic", 0]], "clip_text_model": [["tiny-clip-text-2-mic", 2]], "clip_vision_model": [["clip-vit-base-patch32-vision-mic", 0]], "clipseg": [["clipseg-rd64-refined-mic", 0]], "cohere": [["aya-expanse-8b-ungated-mic", 0]], "cohere2": [["tiny-random-aya-base-mic", 2]], "convbert": [["conv-bert-base-mic", 0]], "convnext": [["convnext-tiny-224-mic", 0]], "convnextv2": [["convnextv2-nano-1k-224-mic", 0]], "cosmos3_omni": [["Cosmos3-Nano-mic", 0]], "csm": [["csm-1b-mic", 0]], "deberta": [["deberta-large-mnli-mic", 0]], "deberta-v2": [["mDeBERTa-v3-base-xnli-multilingual-nli-2mil7-mic", 0]], "deepseek_v2": [["DeepSeek-Coder-V2-Lite-Instruct-mic", 0]], "deepseek_v3": [["DeepSeek-R1-mic", 0]], "deepseek_v32": [["DeepSeek-V3.2-mic", 0]], "deepseek_v4": [["DeepSeek-V4-Flash-0731-mic", 0]], "detr": [["detr-mic", 1]], "dinov2": [["dinov2-small-mic", 0]], "dinov2_with_registers": [["dinov2-with-registers-base-mic", 0]], "dinov3_vit": [["dinov3-vitl16-pretrain-lvd1689m-mic", 0]], "distilbert": [["distilbert-base-uncased-mic", 0]], "dpt": [["dpt-large-mic", 0]], "electra": [["electra-small-discriminator-mic", 0]], "embedding_gemma2": [["embeddinggemma-2-mic", 0]], "encoder-decoder": [["bert2bert-cnn_dailymail-fp16-mic", 0]], "esm": [["esm2_t6_8M_UR50D-mic", 0]], "esmc": [["ESMC-300M-hf-mic", 0]], "eurobert": [["EuroBERT-210m-mic", 0]], "falcon": [["falcon-7b-mic", 0]], "falcon_h1": [["Falcon-H1-Tiny-90M-Instruct-mic", 0]], "florence2": [["Florence-2-base-mic", 0]], "fnet": [["fnet-base-mic", 0]], "gemma": [["gemma-2b-mic", 0]], "gemma2": [["gemma-2-2b-it-mic", 0]], "gemma3": [["gemma3-mic", 0]], "gemma3n": [["gemma-3n-E2B-it-mic", 0]], "gemma4": [["gemma-4-E2B-it-mic", 0]], "gemma4_unified": [["gemma-4-12B-it-mic", 0]], "glm4": [["GLM-4-9B-0414-mic", 0]], "glm4_moe": [["GLM-4.5-Air-mic", 0]], "glm4_moe_lite": [["GLM-4.7-Flash-mic", 0]], "glm5_next": [["GLM-5.3-Flash-mic", 1]], "glm_moe_dsa": [["GLM-5.3-mic", 1]], "glm_ocr": [["GLM-OCR-mic", 0]], "got_ocr2": [["got_ocr2-mic", 0]], "gpt2": [["gpt2-mic", 0]], "gpt_bigcode": [["tiny_starcoder_py-mic", 0]], "gpt_neox": [["pythia-160m-mic", 0]], "gpt_oss": [["gpt_oss-mic", 0]], "granite": [["granite-3.1-2b-instruct-mic", 0]], "granite4_vision": [["granite-vision-4.1-4b-mic", 0]], "granite_speech": [["granite-4.0-1b-speech-mic", 0]], "granitemoe": [["granite-3.0-1b-a400m-instruct-mic", 0]], "granitemoehybrid": [["granite-4.0-h-350m-mic", 0]], "grounding-dino": [["grounding-dino-tiny-mic", 0]], "hubert": [["hubert-large-ls960-ft-mic", 0]], "hunyuan_v1_dense": [["Hunyuan-0.5B-Instruct-mic", 0]], "hunyuan_vl": [["HunyuanOCR-mic", 0]], "hy_v3": [["Hy-MT2-30B-A3B-mic", 0]], "idefics3": [["granite-docling-258M-mic", 0]], "inkling_mm_model": [["Inkling-mic", 0]], "internvl": [["internvl-mic", 0]], "jamba": [["Jamba-tiny-dev-mic", 0]], "janus": [["Janus-Pro-1B-mic", 0]], "jina_embeddings_v3": [["jina-embeddings-v3-hf-mic", 0]], "laguna": [["Laguna-XS.2-mic", 0]], "layoutlm": [["layoutlm-base-uncased-mic", 0]], "lfm2": [["LFM2.5-230M-Base-mic", 0]], "lfm2_moe": [["LFM2-8B-A1B-mic", 0]], "lfm2_vl": [["LFM2.5-VL-450M-mic", 0]], "llama": [["llama-mic", 0], ["smollm-1.7b-mic", 0], ["smollm-mic", 0]], "llama4": [["Llama-4-Scout-17B-16E-Instruct-mic", 0]], "llava": [["llava-mic", 0]], "llava_next": [["llava-v1.6-mistral-7b-hf-mic", 0]], "llava_next_video": [["LLaVA-NeXT-Video-7B-hf-mic", 0]], "llava_onevision": [["llava-onevision-qwen2-0.5b-ov-hf-mic", 0]], "longformer": [["longformer-base-4096-mic", 0]], "longt5": [["long-t5-tglobal-base-mic", 0]], "luke": [["luke-base-mic", 0]], "m2m_100": [["nllb-200-distilled-600M-mic", 0]], "mamba": [["mamba-130m-hf-mic", 0]], "marian": [["opus-mt-fr-en-mic", 0]], "mask2former": [["mask2former-swin-tiny-coco-instance-mic", 0]], "mbart": [["mbart-large-50-mic", 0]], "minicpm3": [["MiniCPM3-4B-mic", 0]], "minicpmv4_6": [["MiniCPM-V-4.6-mic", 0]], "minimax_m2": [["MiniMax-M2.5-mic", 0]], "minimax_m3_vl": [["MiniMax-M3-MXFP8-mic", 0]], "ministral3": [["Nemotron-3-Embed-1B-BF16-mic", 0]], "mistral": [["Mistral-7B-Instruct-v0.2-mic", 1]], "mistral3": [["Ministral-3-3B-Instruct-2512-mic", 0], ["mistral3-mic", 0]], "mixtral": [["Mixtral-8x7B-Instruct-v0.1-mic", 0]], "mllama": [["Llama-3.2-11B-Vision-Instruct-mic", 0]], "mm-grounding-dino": [["mm_grounding_dino_tiny_o365v1_goldg_v3det-mic", 0]], "mobilenet_v2": [["mobilenet_v2_1.0_224-mic", 1]], "mobilevit": [["mobilevit-xx-small-mic", 1]], "modernbert": [["modernbert-mic", 0]], "moss_transcribe_diarize": [["MOSS-Transcribe-Diarize-HF-mic", 0]], "mpnet": [["mpnet-base-mic", 0]], "mt5": [["mt5-small-mic", 0]], "muse_glimmer": [["tiny-random-muse-glimmer-mic", 2]], "nemotron_h": [["NVIDIA-Nemotron-3-Nano-4B-BF16-mic", 0]], "nomic_bert": [["nomic-embed-text-v1.5-mic", 0]], "olmo": [["OLMo-1B-hf-mic", 0]], "olmo2": [["OLMo-2-0425-1B-mic", 0]], "olmo3": [["Olmo-3-7B-Instruct-mic", 0]], "olmoe": [["OLMoE-1B-7B-0924-mic", 0]], "openai_privacy_filter": [["privacy-filter-mic", 0]], "opt": [["opt-125m-mic", 0]], "owlv2": [["owlv2-base-patch16-ensemble-mic", 0]], "owlvit": [["owlvit-base-patch32-mic", 0]], "paddleocr_vl": [["tibetan-ocr-mic", 0]], "paligemma": [["paligemma-3b-pt-224-mic", 0]], "pegasus": [["pegasus-xsum-mic", 0]], "perceiver": [["language-perceiver-mic", 1]], "phi": [["phi-2-mic", 0]], "phi3": [["Phi-3.5-mini-instruct-mic", 0]], "phi4_multimodal": [["Phi-4-multimodal-instruct-mic", 0]], "phimoe": [["Phi-3.5-MoE-0.8B-A0.2B-mic", 0]], "pvt": [["pvt-tiny-224-mic", 0]], "qwen2": [["qwen2-mic", 0]], "qwen2_5_omni": [["qwen2_5_omni-mic", 0]], "qwen2_5_omni_thinker": [["LCO-Embedding-Omni-3B-mic", 0]], "qwen2_5_vl": [["qwen2_5_vl-mic", 0]], "qwen2_audio": [["qwen2_audio-mic", 0]], "qwen2_moe": [["Qwen1.5-MoE-A2.7B-mic", 0]], "qwen2_vl": [["Qwen2-VL-2B-Instruct-mic", 0]], "qwen3": [["Qwen3-0.6B-mic", 0]], "qwen3_5_moe": [["Qwen3.5-35B-A3B-mic", 0]], "qwen3_5_moe_text": [["Qwen3.8-1.0B-A0.6B-mic", 0]], "qwen3_5_text": [["Qwen3.5-0.8B-text-only-mic", 0]], "qwen3_asr": [["Qwen3-ASR-0.6B-hf-mic", 0]], "qwen3_moe": [["Qwen3-30B-A3B-mic", 0]], "qwen3_next": [["Qwen3-Next-80B-A3B-Instruct-mic", 0]], "qwen3_vl": [["Qwen3-VL-2B-Instruct-mic", 0]], "qwen3_vl_moe": [["Qwen3-VL-30B-A3B-Instruct-mic", 0]], "qwen4_exp": [["Qwen3.8-Flash-Next-mic", 0]], "recurrent_gemma": [["recurrentgemma-2b-mic", 0]], "resnet": [["resnet-18-mic", 0]], "roberta": [["roberta-base-mic", 0]], "rt_detr": [["rtdetr_r50vd-mic", 0]], "rt_detr_v2": [["docling-layout-heron-mic", 0]], "sam2": [["sam2.1-hiera-tiny-mic", 4]], "seamless_m4t_v2": [["seamless-m4t-v2-large-mic", 0]], "seed_oss": [["Seed-OSS-36B-Instruct-mic", 0]], "segformer": [["segformer-b0-finetuned-ade-512-512-mic", 0]], "siglip": [["siglip-base-patch16-224-mic", 0]], "smollm3": [["SmolLM3-3B-mic", 1]], "smolvlm": [["smolvlm-mic", 0]], "speecht5": [["speech_t5-mic", 0]], "stablelm": [["stablelm-2-1_6b-mic", 0]], "starcoder2": [["starcoder2-3b-mic", 0]], "swin": [["swin-tiny-patch4-window7-224-mic", 0]], "swinv2": [["swinv2-tiny-patch4-window16-256-mic", 0]], "t5": [["t5-mic", 0], ["t5-small-mic", 0]], "table-transformer": [["table_transformer-mic", 0]], "timesfm": [["timesfm-2.0-500m-pytorch-mic", 0]], "tipsv2": [["tipsv2-b14-mic", 0]], "umt5": [["umt5-small-mic", 0]], "unlimited_ocr": [["Unlimited-OCR-mic", 0]], "vibevoice": [["VibeVoice-1.5B-hf-mic", 0]], "vit": [["vit-mic", 0]], "vitmatte": [["vitmatte-small-composition-1k-mic", 0]], "vits": [["mms-tts-eng-mic", 0]], "voxtral": [["Voxtral-Mini-3B-2507-mic", 0]], "wav2vec2": [["wav2vec2-base-960h-mic", 0], ["wav2vec2-mic", 0]], "wavlm": [["wavlm-base-plus-mic", 0]], "whisper": [["whisper-mic", 0], ["whisper-tiny-mic", 0]], "xlm-roberta": [["xlm-roberta-base-mic", 0]], "xlm-roberta-xl": [["xlm-roberta-xl-mic", 0]], "yolos": [["yolos-small-mic", 0]]}, "gated": {"aya_vision": "auto", "chameleon": "manual", "chmv2": "manual", "cohere_asr": "auto", "csm_depth_decoder_model": "auto", "cwm": "manual", "dinov3_convnext": "manual", "gemma": "manual", "gemma2": "manual", "gemma3": "manual", "gemma3_text": "manual", "gemma3n": "manual", "jais2": "auto", "lasr_ctc": "auto", "llama4_vision_model": "manual", "moshi": "auto", "muse_spark": "manual", "muse_spark_audio": "manual", "muse_spark_text": "manual", "muse_spark_vision": "manual", "neucodec": "auto", "perception_lm": "manual", "recurrent_gemma": "manual", "sam3_detr_decoder": "manual", "sam3_detr_encoder": "manual", "sam3_geometry_encoder": "manual", "sam3_mask_decoder": "manual", "sam3_tracker_mask_decoder": "manual", "sam3_tracker_prompt_encoder": "manual", "sam3_tracker_video": "manual", "sam3_tracker_video_mask_decoder": "manual", "sam3_tracker_video_prompt_encoder": "manual", "sam3_video": "manual", "shieldgemma2": "manual", "t5gemma": "manual", "t5gemma2": "manual", "vaultgemma": "manual"}}
\ No newline at end of file
diff --git a/index.html b/index.html
index 337c73609d8c6cfbad51a0018e20ba354f140554..409b703960b8242a82b27700aa57dafb370abac0 100644
--- a/index.html
+++ b/index.html
@@ -174,7 +174,7 @@ a{color:var(--accent);text-decoration:none}a:hover{text-decoration:underline}
       const view = td('', 'top');
       if (r[1] === 'Done') for (const [page, findings] of shrink[r[0]] || []) {
         const a = document.createElement('a');
-        a.href = `shrink/${page}.html`; a.target = '_blank'; a.rel = 'noopener';
+        a.href = `shrink/${page}.html`;
         a.textContent = page.replace(/-mic$/, '') + (findings ? ` ⚠ ${findings}` : '');
         a.title = `${page}: ${findings ? findings + ' finding(s) to review' : 'same architecture, smaller'}`;
         if (view.childNodes.length) view.append(document.createElement('br'));
diff --git a/shrink/Apertus-v1.1-0.5B-Instruct-mic.html b/shrink/Apertus-v1.1-0.5B-Instruct-mic.html
index 4fd3339a2a94b51f88ff0ca5e16ead0d32cb153d..a66c8a62c6a77a2393ad3485e2a7158d2589c88f 100644
--- a/shrink/Apertus-v1.1-0.5B-Instruct-mic.html
+++ b/shrink/Apertus-v1.1-0.5B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>swiss-ai/Apertus-v1.1-0.5B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Apertus-v1.1-0.5B-Instruct-mic@55ae98cbccd9da534828bccc9b1f8bee4974731d ← swiss-ai/Apertus-v1.1-0.5B-Instruct@a140fd61fb57422c36a26301caea17edee799874 · ApertusForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>swiss-ai/Apertus-v1.1-0.5B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Apertus-v1.1-0.5B-Instruct-mic/tree/55ae98cbccd9da534828bccc9b1f8bee4974731d" title="huggingface-mic/Apertus-v1.1-0.5B-Instruct-mic@55ae98cbccd9da534828bccc9b1f8bee4974731d">huggingface-mic/Apertus-v1.1-0.5B-Instruct-mic@55ae98cbccd9</a> ← <a href="https://huggingface.co/swiss-ai/Apertus-v1.1-0.5B-Instruct/tree/a140fd61fb57422c36a26301caea17edee799874" title="swiss-ai/Apertus-v1.1-0.5B-Instruct@a140fd61fb57422c36a26301caea17edee799874">swiss-ai/Apertus-v1.1-0.5B-Instruct@a140fd61fb57</a> · ApertusForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">438.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 20</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>20</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.693809" title="model.layers: 304.1M (69.4%)"></i><i class="c1" style="flex:0.306189" title="model.embed_tokens: 134.2M (30.6%)"></i><i class="c2" style="flex:0.000002" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006828" title="model.layers: 57.7K (0.7%)"></i><i class="c1" style="flex:0.993165" title="model.embed_tokens: 8.4M (99.3%)"></i><i class="c2" style="flex:0.000008" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>304.1M</td><td>69.4%</td><td>57.7K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>134.2M</td><td>30.6%</td><td>8.4M</td><td>99.3%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>linear</td><td>linear</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>apertus</td><td>apertus</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>xielu</td><td>xielu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Apertus omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>121</td><td>13</td></tr><tr class=""><td>RMSNorm</td><td>81</td><td>9</td></tr><tr class=""><td>Attention</td><td>20</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>20</td><td>2</td></tr><tr class=""><td>MLP</td><td>20</td><td>2</td></tr><tr class=""><td>XIELUActivation</td><td>20</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>6144</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>20</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Apertus-v1.1-0.5B-Instruct-mic.json b/shrink/Apertus-v1.1-0.5B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..21a56608cc6ed5c9a2fe73f7d682c0c8a18d6170
--- /dev/null
+++ b/shrink/Apertus-v1.1-0.5B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Apertus-v1.1-0.5B-Instruct-mic@55ae98cbccd9da534828bccc9b1f8bee4974731d", "findings": []}
\ No newline at end of file
diff --git a/shrink/Cosmos3-Nano-mic.html b/shrink/Cosmos3-Nano-mic.html
index a10a1fa285f918e7bcf0524f30c407c21d5ae070..be4e75a5194707f4bdaa7e5bbf7121141f7b2bf9 100644
--- a/shrink/Cosmos3-Nano-mic.html
+++ b/shrink/Cosmos3-Nano-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>nvidia/Cosmos3-Nano: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Cosmos3-Nano-mic@4c4a117fd8d6b13993505f86a24b15580e8b25d6 ← nvidia/Cosmos3-Nano@e59a53c25979a090fa8706c9acc0c254a6e89b92 · Cosmos3OmniForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>nvidia/Cosmos3-Nano</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Cosmos3-Nano-mic/tree/4c4a117fd8d6b13993505f86a24b15580e8b25d6" title="huggingface-mic/Cosmos3-Nano-mic@4c4a117fd8d6b13993505f86a24b15580e8b25d6">huggingface-mic/Cosmos3-Nano-mic@4c4a117fd8d6</a> ← <a href="https://huggingface.co/nvidia/Cosmos3-Nano/tree/e59a53c25979a090fa8706c9acc0c254a6e89b92" title="nvidia/Cosmos3-Nano@e59a53c25979a090fa8706c9acc0c254a6e89b92">nvidia/Cosmos3-Nano@e59a53c25979</a> · Cosmos3OmniForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">8.8B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">20.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: Qwen3VLVisionBlock"></i><i class="c0" title="layer 1: Qwen3VLVisionBlock"></i><i class="c0" title="layer 2: Qwen3VLVisionBlock"></i><i class="c0" title="layer 3: Qwen3VLVisionBlock"></i><i class="c0" title="layer 4: Qwen3VLVisionBlock"></i><i class="c0" title="layer 5: Qwen3VLVisionBlock"></i><i class="c0" title="layer 6: Qwen3VLVisionBlock"></i><i class="c0" title="layer 7: Qwen3VLVisionBlock"></i><i class="c0" title="layer 8: Qwen3VLVisionBlock"></i><i class="c0" title="layer 9: Qwen3VLVisionBlock"></i><i class="c0" title="layer 10: Qwen3VLVisionBlock"></i><i class="c0" title="layer 11: Qwen3VLVisionBlock"></i><i class="c0" title="layer 12: Qwen3VLVisionBlock"></i><i class="c0" title="layer 13: Qwen3VLVisionBlock"></i><i class="c0" title="layer 14: Qwen3VLVisionBlock"></i><i class="c0" title="layer 15: Qwen3VLVisionBlock"></i><i class="c0" title="layer 16: Qwen3VLVisionBlock"></i><i class="c0" title="layer 17: Qwen3VLVisionBlock"></i><i class="c0" title="layer 18: Qwen3VLVisionBlock"></i><i class="c0" title="layer 19: Qwen3VLVisionBlock"></i><i class="c0" title="layer 20: Qwen3VLVisionBlock"></i><i class="c0" title="layer 21: Qwen3VLVisionBlock"></i><i class="c0" title="layer 22: Qwen3VLVisionBlock"></i><i class="c0" title="layer 23: Qwen3VLVisionBlock"></i><i class="c0" title="layer 24: Qwen3VLVisionBlock"></i><i class="c0" title="layer 25: Qwen3VLVisionBlock"></i><i class="c0" title="layer 26: Qwen3VLVisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen3VLVisionBlock"></i><i class="c0" title="layer 1: Qwen3VLVisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen3VLVisionBlock</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.visual.deepstack_merger_list</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: GELU"></i><i class="c0" title="layer 1: GELU"></i><i class="c0" title="layer 2: GELU"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: GELU"></i><i class="c0" title="layer 1: GELU"></i><i class="c0" title="layer 2: GELU"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>GELU</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 36</span><div class="cells"><i class="c0" title="layer 0: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 1: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 2: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 3: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 4: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 5: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 6: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 7: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 8: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 9: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 10: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 11: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 12: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 13: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 14: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 15: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 16: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 17: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 18: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 19: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 20: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 21: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 22: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 23: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 24: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 25: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 26: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 27: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 28: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 29: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 30: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 31: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 32: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 33: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 34: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 35: Qwen3VLTextDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen3VLTextDecoderLayer"></i><i class="c0" title="layer 1: Qwen3VLTextDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen3VLTextDecoderLayer</td><td>36</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.792286" title="model.language_model.layers: 6.9B (79.2%)"></i><i class="c1" style="flex:0.070985" title="model.language_model: 622.3M (7.1%)"></i><i class="c2" style="flex:0.070984" title="lm_head: 622.3M (7.1%)"></i><i class="c3" style="flex:0.046933" title="model.visual.blocks: 411.5M (4.7%)"></i><i class="c4" style="flex:0.013731" title="model.visual.deepstack_merger_list: 120.4M (1.4%)"></i><i class="c5" style="flex:0.005081" title="model.visual: 44.5M (0.5%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003672" title="model.language_model.layers: 74.0K (0.4%)"></i><i class="c1" style="flex:0.482213" title="model.language_model: 9.7M (48.2%)"></i><i class="c2" style="flex:0.482209" title="lm_head: 9.7M (48.2%)"></i><i class="c3" style="flex:0.003320" title="model.visual.blocks: 66.9K (0.3%)"></i><i class="c4" style="flex:0.012311" title="model.visual.deepstack_merger_list: 248.3K (1.2%)"></i><i class="c5" style="flex:0.016275" title="model.visual: 328.2K (1.6%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>6.9B</td><td>79.2%</td><td>74.0K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>622.3M</td><td>7.1%</td><td>9.7M</td><td>48.2%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>622.3M</td><td>7.1%</td><td>9.7M</td><td>48.2%</td></tr><tr><td><i class="c3"></i>model.visual.blocks</td><td>411.5M</td><td>4.7%</td><td>66.9K</td><td>0.3%</td></tr><tr><td><i class="c4"></i>model.visual.deepstack_merger_list</td><td>120.4M</td><td>1.4%</td><td>248.3K</td><td>1.2%</td></tr><tr><td><i class="c5"></i>model.visual</td><td>44.5M</td><td>0.5%</td><td>328.2K</td><td>1.6%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>cosmos3_omni</td><td>cosmos3_omni</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3_vl_text</td><td>qwen3_vl_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen3_vl_vision</td><td>qwen3_vl_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Cosmos3Omni omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>369</td><td>31</td></tr><tr class=""><td>Qwen3VLTextRMSNorm</td><td>145</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>58</td><td>8</td></tr><tr class=""><td>Qwen3VLTextAttention</td><td>36</td><td>2</td></tr><tr class=""><td>Qwen3VLTextDecoderLayer</td><td>36</td><td>2</td></tr><tr class=""><td>Qwen3VLTextMLP</td><td>36</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>36</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>27</td><td>2</td></tr><tr class=""><td>Qwen3VLVisionAttention</td><td>27</td><td>2</td></tr><tr class=""><td>Qwen3VLVisionBlock</td><td>27</td><td>2</td></tr><tr class=""><td>Qwen3VLVisionMLP</td><td>27</td><td>2</td></tr><tr class=""><td>GELU</td><td>4</td><td>4</td></tr><tr class=""><td>Qwen3VLVisionPatchMerger</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3VLTextModel</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3VLTextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3VLVisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3VLVisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3VLVisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 54 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>4096</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>12288</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>36</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [24, 20, 20], &quot;rope_type&quot;: &quot;defa…</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [3, 3, 2], &quot;rope_theta&quot;: 5000000…</td></tr><tr><td>vision_config.depth</td><td>27</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>4096</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Cosmos3-Nano-mic.json b/shrink/Cosmos3-Nano-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a57792ed1b74625ab347bf615c8f9a313acecb3
--- /dev/null
+++ b/shrink/Cosmos3-Nano-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Cosmos3-Nano-mic@4c4a117fd8d6b13993505f86a24b15580e8b25d6", "findings": []}
\ No newline at end of file
diff --git a/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.html b/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.html
index b1d36db1515329f0f631d12d5af9a942508dadef..d03286abb34737c3fdf92b925f9f736258759703 100644
--- a/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.html
+++ b/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/DeepSeek-Coder-V2-Lite-Instruct-mic@90601ffd1320fd0db167a3ade1c6cd0ed35c2229 ← deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct@e434a23f91ba5b4923cf6c9d9a238eb4a08e3a11 · DeepseekV2ForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">15.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">13.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 2: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 3: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 4: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 5: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 6: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 7: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 8: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 9: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 10: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 11: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 12: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 13: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 14: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 15: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 16: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 17: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 18: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 19: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 20: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 21: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 22: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 23: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 24: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 25: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 26: Experts, Moe, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, Moe, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, Moe, TopkRouter</td><td>26</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.973296" title="model.layers: 15.3B (97.3%)"></i><i class="c1" style="flex:0.013352" title="lm_head: 209.7M (1.3%)"></i><i class="c2" style="flex:0.013352" title="model.embed_tokens: 209.7M (1.3%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006792" title="model.layers: 89.6K (0.7%)"></i><i class="c1" style="flex:0.496602" title="lm_head: 6.6M (49.7%)"></i><i class="c2" style="flex:0.496602" title="model.embed_tokens: 6.6M (49.7%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>15.3B</td><td>97.3%</td><td>89.6K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>209.7M</td><td>1.3%</td><td>6.6M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>209.7M</td><td>1.3%</td><td>6.6M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-6 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>deepseek_v2</td><td>deepseek_v2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>102400</td><td>102400</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>163840</td><td>163840</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DeepseekV2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>190</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>82</td><td>7</td></tr><tr class=""><td>SiLUActivation</td><td>53</td><td>3</td></tr><tr class=""><td>Attention</td><td>27</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>MLP</td><td>27</td><td>2</td></tr><tr class=""><td>Experts</td><td>26</td><td>1</td></tr><tr class=""><td>Moe</td><td>26</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>26</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 39 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>8</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>10944</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>16</td></tr><tr><td>moe_intermediate_size</td><td>1408</td><td>32</td></tr><tr><td>n_routed_experts</td><td>64</td><td>4</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>6</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>27</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr><tr><td>qk_nope_head_dim</td><td>128</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>v_head_dim</td><td>128</td><td>16</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/DeepSeek-Coder-V2-Lite-Instruct-mic/tree/90601ffd1320fd0db167a3ade1c6cd0ed35c2229" title="huggingface-mic/DeepSeek-Coder-V2-Lite-Instruct-mic@90601ffd1320fd0db167a3ade1c6cd0ed35c2229">huggingface-mic/DeepSeek-Coder-V2-Lite-Instruct-mic@90601ffd1320</a> ← <a href="https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct/tree/e434a23f91ba5b4923cf6c9d9a238eb4a08e3a11" title="deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct@e434a23f91ba5b4923cf6c9d9a238eb4a08e3a11">deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct@e434a23f91ba</a> · DeepseekV2ForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">15.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">13.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 2: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 3: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 4: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 5: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 6: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 7: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 8: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 9: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 10: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 11: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 12: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 13: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 14: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 15: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 16: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 17: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 18: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 19: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 20: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 21: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 22: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 23: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 24: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 25: Experts, Moe, TopkRouter"></i><i class="c1" title="layer 26: Experts, Moe, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, Moe, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, Moe, TopkRouter</td><td>26</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.973296" title="model.layers: 15.3B (97.3%)"></i><i class="c1" style="flex:0.013352" title="model.embed_tokens: 209.7M (1.3%)"></i><i class="c2" style="flex:0.013352" title="lm_head: 209.7M (1.3%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006792" title="model.layers: 89.6K (0.7%)"></i><i class="c1" style="flex:0.496602" title="model.embed_tokens: 6.6M (49.7%)"></i><i class="c2" style="flex:0.496602" title="lm_head: 6.6M (49.7%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>15.3B</td><td>97.3%</td><td>89.6K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>209.7M</td><td>1.3%</td><td>6.6M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>209.7M</td><td>1.3%</td><td>6.6M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-6 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>deepseek_v2</td><td>deepseek_v2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>102400</td><td>102400</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>163840</td><td>163840</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DeepseekV2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>190</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>82</td><td>7</td></tr><tr class=""><td>SiLUActivation</td><td>53</td><td>3</td></tr><tr class=""><td>Attention</td><td>27</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>MLP</td><td>27</td><td>2</td></tr><tr class=""><td>Experts</td><td>26</td><td>1</td></tr><tr class=""><td>Moe</td><td>26</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>26</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 39 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>8</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>10944</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>16</td></tr><tr><td>moe_intermediate_size</td><td>1408</td><td>32</td></tr><tr><td>n_routed_experts</td><td>64</td><td>4</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>6</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>27</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr><tr><td>qk_nope_head_dim</td><td>128</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>v_head_dim</td><td>128</td><td>16</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.json b/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..05f4975d43f898f2c19efc6cb539511c8112404e
--- /dev/null
+++ b/shrink/DeepSeek-Coder-V2-Lite-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/DeepSeek-Coder-V2-Lite-Instruct-mic@90601ffd1320fd0db167a3ade1c6cd0ed35c2229", "findings": []}
\ No newline at end of file
diff --git a/shrink/DeepSeek-R1-mic.html b/shrink/DeepSeek-R1-mic.html
index a28e1ba5e0c5dcf28a13a1e4cbb65b6efb121f92..600511fc9be25f89164d04292ca92004df74b98d 100644
--- a/shrink/DeepSeek-R1-mic.html
+++ b/shrink/DeepSeek-R1-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>deepseek-ai/DeepSeek-R1: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/DeepSeek-R1-mic@89eddb3d8f9419e63b667eb7d5244fec8c768ed9 ← deepseek-ai/DeepSeek-R1@56d4cbbb4d29f4355bab4b9a39ccb717a14ad5ad · DeepseekV3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>deepseek-ai/DeepSeek-R1</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/DeepSeek-R1-mic/tree/89eddb3d8f9419e63b667eb7d5244fec8c768ed9" title="huggingface-mic/DeepSeek-R1-mic@89eddb3d8f9419e63b667eb7d5244fec8c768ed9">huggingface-mic/DeepSeek-R1-mic@89eddb3d8f94</a> ← <a href="https://huggingface.co/deepseek-ai/DeepSeek-R1/tree/56d4cbbb4d29f4355bab4b9a39ccb717a14ad5ad" title="deepseek-ai/DeepSeek-R1@56d4cbbb4d29f4355bab4b9a39ccb717a14ad5ad">deepseek-ai/DeepSeek-R1@56d4cbbb4d29</a> · DeepseekV3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">671.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 61</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c0" title="layer 1: no Experts"></i><i class="c0" title="layer 2: no Experts"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 6: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 10: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 14: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 18: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 22: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 26: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 30: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 34: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 38: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 42: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 46: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 47: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 48: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 49: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 50: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 51: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 52: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 53: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 54: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 55: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 56: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 57: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 58: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 59: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 60: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>3</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>58</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.997238" title="model.layers: 669.2B (99.7%)"></i><i class="c1" style="flex:0.001381" title="model.embed_tokens: 926.7M (0.1%)"></i><i class="c2" style="flex:0.001381" title="lm_head: 926.7M (0.1%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 7.2K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004593" title="model.layers: 76.4K (0.5%)"></i><i class="c1" style="flex:0.497702" title="model.embed_tokens: 8.3M (49.8%)"></i><i class="c2" style="flex:0.497702" title="lm_head: 8.3M (49.8%)"></i><i class="c3" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>669.2B</td><td>99.7%</td><td>76.4K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>926.7M</td><td>0.1%</td><td>8.3M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>926.7M</td><td>0.1%</td><td>8.3M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>7.2K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (128 heads, 128 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>deepseek_v3</td><td>deepseek_v3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>129280</td><td>129280</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>163840</td><td>163840</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DeepseekV3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>489</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>245</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>119</td><td>3</td></tr><tr class=""><td>Attention</td><td>61</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>61</td><td>2</td></tr><tr class=""><td>MLP</td><td>61</td><td>2</td></tr><tr class=""><td>Experts</td><td>58</td><td>1</td></tr><tr class=""><td>MoE</td><td>58</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>58</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">18 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>first_k_dense_replace</td><td>3</td><td>1</td></tr><tr><td>head_dim</td><td>64</td><td>8</td></tr><tr><td>hidden_size</td><td>7168</td><td>64</td></tr><tr><td>intermediate_size</td><td>18432</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>16</td></tr><tr><td>moe_intermediate_size</td><td>2048</td><td>32</td></tr><tr><td>n_group</td><td>8</td><td>1</td></tr><tr><td>n_routed_experts</td><td>256</td><td>4</td></tr><tr><td>num_attention_heads</td><td>128</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>61</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>128</td><td>4</td></tr><tr><td>q_lora_rank</td><td>1536</td><td>16</td></tr><tr><td>qk_nope_head_dim</td><td>128</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;e4m3&quot;, &quot;quant_method&quot;: &quot;fp8&quot;, &quot;weigh…</td><td>—</td></tr><tr><td>topk_group</td><td>4</td><td>1</td></tr><tr><td>v_head_dim</td><td>128</td><td>16</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/DeepSeek-R1-mic.json b/shrink/DeepSeek-R1-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..6712ed6cfb33da476e1f671495ffdef65243fcc8
--- /dev/null
+++ b/shrink/DeepSeek-R1-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/DeepSeek-R1-mic@89eddb3d8f9419e63b667eb7d5244fec8c768ed9", "findings": []}
\ No newline at end of file
diff --git a/shrink/DeepSeek-V3.2-mic.html b/shrink/DeepSeek-V3.2-mic.html
index 6440a2c04744163aabcaf8f9176593ed5082c171..84c75680f674077cc0391bc974dff0658543ff5b 100644
--- a/shrink/DeepSeek-V3.2-mic.html
+++ b/shrink/DeepSeek-V3.2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>deepseek-ai/DeepSeek-V3.2: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/DeepSeek-V3.2-mic@2156444e97b761624604df50de0ac677d402f289 ← deepseek-ai/DeepSeek-V3.2@a7e62ac04ecb2c0a54d736dc46601c5606cf10a6 · DeepseekV32ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>deepseek-ai/DeepSeek-V3.2</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/DeepSeek-V3.2-mic/tree/2156444e97b761624604df50de0ac677d402f289" title="huggingface-mic/DeepSeek-V3.2-mic@2156444e97b761624604df50de0ac677d402f289">huggingface-mic/DeepSeek-V3.2-mic@2156444e97b7</a> ← <a href="https://huggingface.co/deepseek-ai/DeepSeek-V3.2/tree/a7e62ac04ecb2c0a54d736dc46601c5606cf10a6" title="deepseek-ai/DeepSeek-V3.2@a7e62ac04ecb2c0a54d736dc46601c5606cf10a6">deepseek-ai/DeepSeek-V3.2@a7e62ac04ecb</a> · DeepseekV32ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">671.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 61</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c0" title="layer 1: no Experts"></i><i class="c0" title="layer 2: no Experts"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 6: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 10: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 14: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 18: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 22: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 26: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 30: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 34: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 38: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 42: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 46: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 47: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 48: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 49: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 50: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 51: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 52: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 53: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 54: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 55: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 56: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 57: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 58: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 59: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 60: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>3</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>58</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.997242" title="model.layers: 670.0B (99.7%)"></i><i class="c1" style="flex:0.001379" title="model.embed_tokens: 926.7M (0.1%)"></i><i class="c2" style="flex:0.001379" title="lm_head: 926.7M (0.1%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 7.2K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.021499" title="model.layers: 363.6K (2.1%)"></i><i class="c1" style="flex:0.489248" title="model.embed_tokens: 8.3M (48.9%)"></i><i class="c2" style="flex:0.489248" title="lm_head: 8.3M (48.9%)"></i><i class="c3" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>670.0B</td><td>99.7%</td><td>363.6K</td><td>2.1%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>926.7M</td><td>0.1%</td><td>8.3M</td><td>48.9%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>926.7M</td><td>0.1%</td><td>8.3M</td><td>48.9%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>7.2K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (128 heads, 128 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>deepseek_v32</td><td>deepseek_v32</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>129280</td><td>129280</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>163840</td><td>163840</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DeepseekV32 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>672</td><td>23</td></tr><tr class=""><td>RMSNorm</td><td>245</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>119</td><td>3</td></tr><tr class=""><td>Attention</td><td>61</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>61</td><td>2</td></tr><tr class=""><td>Indexer</td><td>61</td><td>2</td></tr><tr class=""><td>MLP</td><td>61</td><td>2</td></tr><tr class=""><td>LayerNorm</td><td>61</td><td>2</td></tr><tr class=""><td>Experts</td><td>58</td><td>1</td></tr><tr class=""><td>MoE</td><td>58</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>58</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">20 fields changed, 36 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>first_k_dense_replace</td><td>3</td><td>1</td></tr><tr><td>head_dim</td><td>64</td><td>8</td></tr><tr><td>hidden_size</td><td>7168</td><td>64</td></tr><tr><td>intermediate_size</td><td>18432</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>16</td></tr><tr><td>layer_types</td><td>61 entries: 61× indexed_attention</td><td>[&quot;indexed_attention&quot;, &quot;indexed_attention&quot;]</td></tr><tr><td>mlp_layer_types</td><td>61 entries: 58× sparse, 3× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>2048</td><td>32</td></tr><tr><td>n_group</td><td>8</td><td>1</td></tr><tr><td>n_routed_experts</td><td>256</td><td>4</td></tr><tr><td>num_attention_heads</td><td>128</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>61</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>128</td><td>4</td></tr><tr><td>q_lora_rank</td><td>1536</td><td>16</td></tr><tr><td>qk_nope_head_dim</td><td>128</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;e4m3&quot;, &quot;quant_method&quot;: &quot;fp8&quot;, &quot;scale…</td><td>—</td></tr><tr><td>topk_group</td><td>4</td><td>1</td></tr><tr><td>v_head_dim</td><td>128</td><td>16</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/DeepSeek-V3.2-mic.json b/shrink/DeepSeek-V3.2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..bb2976d4c6b19fc9ca40b43d50806f017abf1f63
--- /dev/null
+++ b/shrink/DeepSeek-V3.2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/DeepSeek-V3.2-mic@2156444e97b761624604df50de0ac677d402f289", "findings": []}
\ No newline at end of file
diff --git a/shrink/DeepSeek-V4-Flash-0731-mic.html b/shrink/DeepSeek-V4-Flash-0731-mic.html
index 54de7c82eb286e6437a5bab3e5b53a007cb3beda..c094ab18afefb55d1012475e3112ead1235e820b 100644
--- a/shrink/DeepSeek-V4-Flash-0731-mic.html
+++ b/shrink/DeepSeek-V4-Flash-0731-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>deepseek-ai/DeepSeek-V4-Flash-0731: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/DeepSeek-V4-Flash-0731-mic@7c36fb38ddaf52e554d784d9e6a2a7ed435b8311 ← deepseek-ai/DeepSeek-V4-Flash-0731@7872f01b1d1fe23eabc4c98b48bffcef5a386062 · DeepseekV4ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>deepseek-ai/DeepSeek-V4-Flash-0731</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/DeepSeek-V4-Flash-0731-mic/tree/7c36fb38ddaf52e554d784d9e6a2a7ed435b8311" title="huggingface-mic/DeepSeek-V4-Flash-0731-mic@7c36fb38ddaf52e554d784d9e6a2a7ed435b8311">huggingface-mic/DeepSeek-V4-Flash-0731-mic@7c36fb38ddaf</a> ← <a href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731/tree/7872f01b1d1fe23eabc4c98b48bffcef5a386062" title="deepseek-ai/DeepSeek-V4-Flash-0731@7872f01b1d1fe23eabc4c98b48bffcef5a386062">deepseek-ai/DeepSeek-V4-Flash-0731@7872f01b1d1f</a> · DeepseekV4ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">284.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">20.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 43</span><div class="cells"><i class="c0" title="layer 0: sliding_attention, HashRouter"></i><i class="c0" title="layer 1: sliding_attention, HashRouter"></i><i class="c1" title="layer 2: compressed_sparse_attention, CSACompressor, HashRouter …"></i><i class="c2" title="layer 3: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 4: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 5: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 6: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 7: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 8: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 9: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 10: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 11: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 12: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 13: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 14: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 15: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 16: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 17: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 18: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 19: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 20: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 21: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 22: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 23: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 24: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 25: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 26: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 27: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 28: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 29: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 30: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 31: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 32: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 33: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 34: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 35: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 36: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 37: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 38: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 39: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 40: compressed_sparse_attention, CSACompressor, Indexer …"></i><i class="c2" title="layer 41: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 42: compressed_sparse_attention, CSACompressor, Indexer …"></i></div></div><div class="row"><span class="lab">Fixture · 5</span><div class="cells"><i class="c0" title="layer 0: sliding_attention, HashRouter"></i><i class="c0" title="layer 1: sliding_attention, HashRouter"></i><i class="c1" title="layer 2: compressed_sparse_attention, CSACompressor, HashRouter …"></i><i class="c2" title="layer 3: heavily_compressed_attention, HCACompressor, RotaryEmbedding …"></i><i class="c3" title="layer 4: compressed_sparse_attention, CSACompressor, Indexer …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention, HashRouter</td><td>2</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>compressed_sparse_attention, CSACompressor, HashRouter …</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>heavily_compressed_attention, HCACompressor, RotaryEmbedding …</td><td>20</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c3"></i>compressed_sparse_attention, CSACompressor, Indexer …</td><td>20</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.996275" title="model.layers: 283.3B (99.6%)"></i><i class="c1" style="flex:0.001862" title="lm_head: 529.5M (0.2%)"></i><i class="c2" style="flex:0.001862" title="model.embed_tokens: 529.5M (0.2%)"></i><i class="c3" style="flex:0.000000" title="model.hc_head: 65.5K (0.0%)"></i><i class="c4" style="flex:0.000000" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.175915" title="model.layers: 3.5M (17.6%)"></i><i class="c1" style="flex:0.412015" title="lm_head: 8.3M (41.2%)"></i><i class="c2" style="flex:0.412015" title="model.embed_tokens: 8.3M (41.2%)"></i><i class="c3" style="flex:0.000051" title="model.hc_head: 1.0K (0.0%)"></i><i class="c4" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>283.3B</td><td>99.6%</td><td>3.5M</td><td>17.6%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>529.5M</td><td>0.2%</td><td>8.3M</td><td>41.2%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>529.5M</td><td>0.2%</td><td>8.3M</td><td>41.2%</td></tr><tr><td><i class="c3"></i>model.hc_head</td><td>65.5K</td><td>0.0%</td><td>1.0K</td><td>0.0%</td></tr><tr><td><i class="c4"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MQA (64 heads, 1 KV)</td><td>MQA (4 heads, 1 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-6 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>deepseek_v4</td><td>deepseek_v4</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>129280</td><td>129280</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>1048576</td><td>1048576</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>rope_theta</td><td>10000</td><td>10000</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DeepseekV4 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>468</td><td>50</td></tr><tr class=""><td>RMSNorm</td><td>235</td><td>26</td></tr><tr class=""><td>UnweightedRMSNorm</td><td>130</td><td>16</td></tr><tr class=""><td>HyperConnection</td><td>86</td><td>10</td></tr><tr class=""><td>SiLUActivation</td><td>86</td><td>10</td></tr><tr class=""><td>RotaryEmbedding</td><td>63</td><td>6</td></tr><tr class=""><td>Attention</td><td>43</td><td>5</td></tr><tr class=""><td>DecoderLayer</td><td>43</td><td>5</td></tr><tr class=""><td>Experts</td><td>43</td><td>5</td></tr><tr class=""><td>GroupedLinear</td><td>43</td><td>5</td></tr><tr class=""><td>MLP</td><td>43</td><td>5</td></tr><tr class=""><td>SparseMoeBlock</td><td>43</td><td>5</td></tr><tr class=""><td>SqrtSoftplusActivation</td><td>43</td><td>5</td></tr><tr class=""><td>TopKRouter</td><td>40</td><td>2</td></tr><tr class=""><td>CSACompressor</td><td>21</td><td>2</td></tr><tr class=""><td>Indexer</td><td>21</td><td>2</td></tr><tr class=""><td>IndexerScorer</td><td>21</td><td>2</td></tr><tr class=""><td>HCACompressor</td><td>20</td><td>1</td></tr><tr class=""><td>HashRouter</td><td>3</td><td>3</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>HyperHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 52 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>512</td><td>16</td></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>layer_types</td><td>43 entries: 21× compressed_sparse_attention, 20× heavily_compressed_attention, 2× sliding_attention</td><td>[&quot;sliding_attention&quot;, &quot;sliding_attention&quot;, &quot;compressed_sparse_attention&quot;, &quot;he…</td></tr><tr><td>mlp_layer_types</td><td>43 entries: 40× moe, 3× hash_moe</td><td>[&quot;hash_moe&quot;, &quot;hash_moe&quot;, &quot;hash_moe&quot;, &quot;moe&quot;, &quot;moe&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>2048</td><td>32</td></tr><tr><td>n_routed_experts</td><td>256</td><td>4</td></tr><tr><td>num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>6</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>43</td><td>5</td></tr><tr><td>q_lora_rank</td><td>1024</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>2</td></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;e4m3&quot;, &quot;quant_method&quot;: &quot;fp8&quot;, &quot;scale…</td><td>—</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/DeepSeek-V4-Flash-0731-mic.json b/shrink/DeepSeek-V4-Flash-0731-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1b6b539a7e6479f4ce56f23983342a6bbb30d2dc
--- /dev/null
+++ b/shrink/DeepSeek-V4-Flash-0731-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/DeepSeek-V4-Flash-0731-mic@7c36fb38ddaf52e554d784d9e6a2a7ed435b8311", "findings": []}
\ No newline at end of file
diff --git a/shrink/ESMC-300M-hf-mic.html b/shrink/ESMC-300M-hf-mic.html
index 721bfab75bcdfa8e823a16f6f3406e06f631b264..41c3782c115da9a6c4b4f5d51063a0d4867c4c9f 100644
--- a/shrink/ESMC-300M-hf-mic.html
+++ b/shrink/ESMC-300M-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>biohub/ESMC-300M-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/ESMC-300M-hf-mic@7c11c22ceffcbddb7fb1bb3216255d34ff38a43c ← biohub/ESMC-300M-hf@5554e4b3ab19eb66f04d7ff8e8c2d36c69b52fa0 · EsmcForMaskedLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>biohub/ESMC-300M-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/ESMC-300M-hf-mic/tree/7c11c22ceffcbddb7fb1bb3216255d34ff38a43c" title="huggingface-mic/ESMC-300M-hf-mic@7c11c22ceffcbddb7fb1bb3216255d34ff38a43c">huggingface-mic/ESMC-300M-hf-mic@7c11c22ceffc</a> ← <a href="https://huggingface.co/biohub/ESMC-300M-hf/tree/5554e4b3ab19eb66f04d7ff8e8c2d36c69b52fa0" title="biohub/ESMC-300M-hf@5554e4b3ab19eb66f04d7ff8e8c2d36c69b52fa0">biohub/ESMC-300M-hf@5554e4b3ab19</a> · EsmcForMaskedLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">333.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">95.3K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>esmc.layers</h3><div class="row"><span class="lab">Original · 30</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i><i class="c0" title="layer 24: Layer"></i><i class="c0" title="layer 25: Layer"></i><i class="c0" title="layer 26: Layer"></i><i class="c0" title="layer 27: Layer"></i><i class="c0" title="layer 28: Layer"></i><i class="c0" title="layer 29: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>30</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.996852" title="esmc.layers: 331.9M (99.7%)"></i><i class="c1" style="flex:0.002770" title="lm_head.dense: 922.6K (0.3%)"></i><i class="c2" style="flex:0.000185" title="lm_head.decoder: 61.5K (0.0%)"></i><i class="c3" style="flex:0.000185" title="esmc.embed_tokens: 61.4K (0.0%)"></i><i class="c4" style="flex:0.000006" title="lm_head.layer_norm: 1.9K (0.0%)"></i><i class="c5" style="flex:0.000003" title="esmc.norm: 960 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.867696" title="esmc.layers: 82.7K (86.8%)"></i><i class="c1" style="flex:0.043653" title="lm_head.dense: 4.2K (4.4%)"></i><i class="c2" style="flex:0.043653" title="lm_head.decoder: 4.2K (4.4%)"></i><i class="c3" style="flex:0.042982" title="esmc.embed_tokens: 4.1K (4.3%)"></i><i class="c4" style="flex:0.001343" title="lm_head.layer_norm: 128 (0.1%)"></i><i class="c5" style="flex:0.000672" title="esmc.norm: 64 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>esmc.layers</td><td>331.9M</td><td>99.7%</td><td>82.7K</td><td>86.8%</td></tr><tr><td><i class="c1"></i>lm_head.dense</td><td>922.6K</td><td>0.3%</td><td>4.2K</td><td>4.4%</td></tr><tr><td><i class="c2"></i>lm_head.decoder</td><td>61.5K</td><td>0.0%</td><td>4.2K</td><td>4.4%</td></tr><tr><td><i class="c3"></i>esmc.embed_tokens</td><td>61.4K</td><td>0.0%</td><td>4.1K</td><td>4.3%</td></tr><tr><td><i class="c4"></i>lm_head.layer_norm</td><td>1.9K</td><td>0.0%</td><td>128</td><td>0.1%</td></tr><tr><td><i class="c5"></i>esmc.norm</td><td>960</td><td>0.0%</td><td>64</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (15 heads, 15 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>esmc</td><td>esmc</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>64</td><td>64</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Esmc omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>212</td><td>16</td></tr><tr class=""><td>LayerNorm</td><td>121</td><td>9</td></tr><tr class=""><td>Attention</td><td>30</td><td>2</td></tr><tr class=""><td>Layer</td><td>30</td><td>2</td></tr><tr class=""><td>MLP</td><td>30</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>30</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>MaskedLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>LayerNorm</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 23 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>960</td><td>64</td></tr><tr><td>intermediate_size</td><td>2560</td><td>128</td></tr><tr><td>num_attention_heads</td><td>15</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>30</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>15</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/ESMC-300M-hf-mic.json b/shrink/ESMC-300M-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..54914ff8924d68409bd6c0ee7f9c34284af671de
--- /dev/null
+++ b/shrink/ESMC-300M-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/ESMC-300M-hf-mic@7c11c22ceffcbddb7fb1bb3216255d34ff38a43c", "findings": []}
\ No newline at end of file
diff --git a/shrink/EuroBERT-210m-mic.html b/shrink/EuroBERT-210m-mic.html
index b433e9c2f65392126f9f3cdf3b63be778144f9f1..0d796f547fb79918ec46d9c34ef088c4fb1f63de 100644
--- a/shrink/EuroBERT-210m-mic.html
+++ b/shrink/EuroBERT-210m-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>EuroBERT/EuroBERT-210m: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/EuroBERT-210m-mic@514cea1b679f118f051124ef88272fb809f25f16 ← EuroBERT/EuroBERT-210m@39b51e15dd1f1a06f58b5cbf6a8a188cec60bd0e · EuroBertForMaskedLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>EuroBERT/EuroBERT-210m</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/EuroBERT-210m-mic/tree/514cea1b679f118f051124ef88272fb809f25f16" title="huggingface-mic/EuroBERT-210m-mic@514cea1b679f118f051124ef88272fb809f25f16">huggingface-mic/EuroBERT-210m-mic@514cea1b679f</a> ← <a href="https://huggingface.co/EuroBERT/EuroBERT-210m/tree/39b51e15dd1f1a06f58b5cbf6a8a188cec60bd0e" title="EuroBERT/EuroBERT-210m@39b51e15dd1f1a06f58b5cbf6a8a188cec60bd0e">EuroBERT/EuroBERT-210m@39b51e15dd1f</a> · EuroBertForMaskedLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">310.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.365056" title="model.layers: 113.3M (36.5%)"></i><i class="c1" style="flex:0.317471" title="model.embed_tokens: 98.5M (31.7%)"></i><i class="c2" style="flex:0.317471" title="lm_head: 98.5M (31.7%)"></i><i class="c3" style="flex:0.000002" title="model.norm: 768 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004981" title="model.layers: 82.2K (0.5%)"></i><i class="c1" style="flex:0.497508" title="model.embed_tokens: 8.2M (49.8%)"></i><i class="c2" style="flex:0.497508" title="lm_head: 8.2M (49.8%)"></i><i class="c3" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>113.3M</td><td>36.5%</td><td>82.2K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>98.5M</td><td>31.7%</td><td>8.2M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>98.5M</td><td>31.7%</td><td>8.2M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>768</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (12 heads, 12 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>eurobert</td><td>eurobert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>128256</td><td>128256</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix EuroBert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>85</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>25</td><td>5</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>MLP</td><td>12</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 33 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>12</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/EuroBERT-210m-mic.json b/shrink/EuroBERT-210m-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a9ccc25f1a123aee6499b55f3eaf9a032adc1d86
--- /dev/null
+++ b/shrink/EuroBERT-210m-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/EuroBERT-210m-mic@514cea1b679f118f051124ef88272fb809f25f16", "findings": []}
\ No newline at end of file
diff --git a/shrink/Falcon-H1-Tiny-90M-Instruct-mic.html b/shrink/Falcon-H1-Tiny-90M-Instruct-mic.html
index 21ad67604569a33dc7996ea08a76fd55052ed4cb..ec8552d6560ccb1688612fd23372604400b00496 100644
--- a/shrink/Falcon-H1-Tiny-90M-Instruct-mic.html
+++ b/shrink/Falcon-H1-Tiny-90M-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>tiiuae/Falcon-H1-Tiny-90M-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Falcon-H1-Tiny-90M-Instruct-mic@0db75b996c3e47fea9cc1a9c213983988b24495a ← tiiuae/Falcon-H1-Tiny-90M-Instruct@e6389502a0b12cd8da894b395ba5bf7436873b16 · FalconH1ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>tiiuae/Falcon-H1-Tiny-90M-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Falcon-H1-Tiny-90M-Instruct-mic/tree/0db75b996c3e47fea9cc1a9c213983988b24495a" title="huggingface-mic/Falcon-H1-Tiny-90M-Instruct-mic@0db75b996c3e47fea9cc1a9c213983988b24495a">huggingface-mic/Falcon-H1-Tiny-90M-Instruct-mic@0db75b996c3e</a> ← <a href="https://huggingface.co/tiiuae/Falcon-H1-Tiny-90M-Instruct/tree/e6389502a0b12cd8da894b395ba5bf7436873b16" title="tiiuae/Falcon-H1-Tiny-90M-Instruct@e6389502a0b12cd8da894b395ba5bf7436873b16">tiiuae/Falcon-H1-Tiny-90M-Instruct@e6389502a0b1</a> · FalconH1ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">91.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.815895" title="model.layers: 74.4M (81.6%)"></i><i class="c1" style="flex:0.184100" title="model.embed_tokens: 16.8M (18.4%)"></i><i class="c2" style="flex:0.000006" title="model.final_layernorm: 512 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.154103" title="model.layers: 382.1K (15.4%)"></i><i class="c1" style="flex:0.845871" title="model.embed_tokens: 2.1M (84.6%)"></i><i class="c2" style="flex:0.000026" title="model.final_layernorm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>74.4M</td><td>81.6%</td><td>382.1K</td><td>15.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>16.8M</td><td>18.4%</td><td>2.1M</td><td>84.6%</td></tr><tr><td><i class="c2"></i>model.final_layernorm</td><td>512</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>falcon_h1</td><td>falcon_h1</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32768</td><td>32768</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix FalconH1 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>217</td><td>19</td></tr><tr class=""><td>RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>48</td><td>4</td></tr><tr class=""><td>Conv1d</td><td>24</td><td>2</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>Mixer</td><td>24</td><td>2</td></tr><tr class=""><td>Identity</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 53 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>512</td><td>64</td></tr><tr><td>intermediate_size</td><td>768</td><td>128</td></tr><tr><td>mamba_chunk_size</td><td>128</td><td>8</td></tr><tr><td>mamba_d_head</td><td>32</td><td>96</td></tr><tr><td>mamba_d_state</td><td>64</td><td>16</td></tr><tr><td>mamba_n_heads</td><td>24</td><td>8</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Falcon-H1-Tiny-90M-Instruct-mic.json b/shrink/Falcon-H1-Tiny-90M-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..cc83226aa0d3f54b566e38ee9e9d63a63a23a9f7
--- /dev/null
+++ b/shrink/Falcon-H1-Tiny-90M-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Falcon-H1-Tiny-90M-Instruct-mic@0db75b996c3e47fea9cc1a9c213983988b24495a", "findings": []}
\ No newline at end of file
diff --git a/shrink/Florence-2-base-mic.html b/shrink/Florence-2-base-mic.html
index 641104dd1814b250c090b8dc9688793f10681607..393591f7e480632a1b9dbf026dc54df52849d611 100644
--- a/shrink/Florence-2-base-mic.html
+++ b/shrink/Florence-2-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>florence-community/Florence-2-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Florence-2-base-mic@85c1b12e1d9565763f6f2b9466669c58bc7aac33 ← florence-community/Florence-2-base@00921df66db728a9ceb750f5eca43e5c203a2051 · Florence2ForConditionalGeneration · overfit weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>florence-community/Florence-2-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Florence-2-base-mic/tree/85c1b12e1d9565763f6f2b9466669c58bc7aac33" title="huggingface-mic/Florence-2-base-mic@85c1b12e1d9565763f6f2b9466669c58bc7aac33">huggingface-mic/Florence-2-base-mic@85c1b12e1d95</a> ← <a href="https://huggingface.co/florence-community/Florence-2-base/tree/00921df66db728a9ceb750f5eca43e5c203a2051" title="florence-community/Florence-2-base@00921df66db728a9ceb750f5eca43e5c203a2051">florence-community/Florence-2-base@00921df66db7</a> · Florence2ForConditionalGeneration · overfit weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">231.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">6.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">4</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.convs</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i><i class="c0" title="layer 3: Conv2d"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i><i class="c0" title="layer 3: Conv2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv2d</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vision_tower.blocks</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.encoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: BartEncoderLayer"></i><i class="c0" title="layer 1: BartEncoderLayer"></i><i class="c0" title="layer 2: BartEncoderLayer"></i><i class="c0" title="layer 3: BartEncoderLayer"></i><i class="c0" title="layer 4: BartEncoderLayer"></i><i class="c0" title="layer 5: BartEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: BartEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BartEncoderLayer</td><td>6</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: BartDecoderLayer"></i><i class="c0" title="layer 1: BartDecoderLayer"></i><i class="c0" title="layer 2: BartDecoderLayer"></i><i class="c0" title="layer 3: BartDecoderLayer"></i><i class="c0" title="layer 4: BartDecoderLayer"></i><i class="c0" title="layer 5: BartDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: BartDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BartDecoderLayer</td><td>6</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.363596" title="model.vision_tower.blocks: 84.2M (36.4%)"></i><i class="c1" style="flex:0.245030" title="model.language_model.decoder.layers: 56.7M (24.5%)"></i><i class="c2" style="flex:0.183747" title="model.language_model.encoder.layers: 42.5M (18.4%)"></i><i class="c3" style="flex:0.177144" title="model.language_model: 41.0M (17.7%)"></i><i class="c4" style="flex:0.026857" title="model.vision_tower.convs: 6.2M (2.7%)"></i><i class="c5" style="flex:0.003626" title="model.multi_modal_projector: 839.2K (0.4%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.391378" title="model.vision_tower.blocks: 2.5M (39.1%)"></i><i class="c1" style="flex:0.007796" title="model.language_model.decoder.layers: 50.2K (0.8%)"></i><i class="c2" style="flex:0.005194" title="model.language_model.encoder.layers: 33.5K (0.5%)"></i><i class="c3" style="flex:0.530140" title="model.language_model: 3.4M (53.0%)"></i><i class="c4" style="flex:0.060945" title="model.vision_tower.convs: 392.8K (6.1%)"></i><i class="c5" style="flex:0.004548" title="model.multi_modal_projector: 29.3K (0.5%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.vision_tower.blocks</td><td>84.2M</td><td>36.4%</td><td>2.5M</td><td>39.1%</td></tr><tr><td><i class="c1"></i>model.language_model.decoder.layers</td><td>56.7M</td><td>24.5%</td><td>50.2K</td><td>0.8%</td></tr><tr><td><i class="c2"></i>model.language_model.encoder.layers</td><td>42.5M</td><td>18.4%</td><td>33.5K</td><td>0.5%</td></tr><tr><td><i class="c3"></i>model.language_model</td><td>41.0M</td><td>17.7%</td><td>3.4M</td><td>53.0%</td></tr><tr><td><i class="c4"></i>model.vision_tower.convs</td><td>6.2M</td><td>2.7%</td><td>392.8K</td><td>6.1%</td></tr><tr><td><i class="c5"></i>model.multi_modal_projector</td><td>839.2K</td><td>0.4%</td><td>29.3K</td><td>0.5%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>florence2</td><td>florence2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>bart</td><td>bart</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>51328</td><td>51328</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>1024</td><td>1024</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>florence_vision</td><td>florence_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>max_position_embeddings</td><td>50</td><td>50</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>[7, 3, 3, 3]</td><td>[7, 3, 3, 3]</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Florence2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>194</td><td>58</td></tr><tr class=""><td>LayerNorm</td><td>85</td><td>32</td></tr><tr class=""><td>Conv2d</td><td>52</td><td>24</td></tr><tr class=""><td>DropPath</td><td>46</td><td>18</td></tr><tr class=""><td>GELUActivation</td><td>36</td><td>12</td></tr><tr class=""><td>VisionMLP</td><td>24</td><td>10</td></tr><tr class=""><td>BartAttention</td><td>18</td><td>3</td></tr><tr class=""><td>VisionBlock</td><td>12</td><td>5</td></tr><tr class=""><td>VisionChannelAttention</td><td>12</td><td>5</td></tr><tr class=""><td>VisionChannelBlock</td><td>12</td><td>5</td></tr><tr class=""><td>VisionSpatialBlock</td><td>12</td><td>5</td></tr><tr class=""><td>VisionWindowAttention</td><td>12</td><td>5</td></tr><tr class=""><td>ModuleList</td><td>8</td><td>8</td></tr><tr class=""><td>BartDecoderLayer</td><td>6</td><td>1</td></tr><tr class=""><td>BartEncoderLayer</td><td>6</td><td>1</td></tr><tr class=""><td>VisionConvEmbed</td><td>4</td><td>4</td></tr><tr class=""><td>BartScaledWordEmbedding</td><td>3</td><td>3</td></tr><tr class=""><td>BartLearnedPositionalEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Identity</td><td>2</td><td>2</td></tr><tr class=""><td>BartDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>BartEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>BartModel</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>VisionBackbone</td><td>1</td><td>1</td></tr><tr class=""><td>VisionLearnedAbsolutePositionEmbedding2D</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPositionalEmbeddingCosine1D</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 72 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.d_model</td><td>768</td><td>64</td></tr><tr><td>text_config.decoder_attention_heads</td><td>12</td><td>2</td></tr><tr><td>text_config.decoder_ffn_dim</td><td>3072</td><td>128</td></tr><tr><td>text_config.decoder_layers</td><td>6</td><td>1</td></tr><tr><td>text_config.encoder_attention_heads</td><td>12</td><td>2</td></tr><tr><td>text_config.encoder_ffn_dim</td><td>3072</td><td>128</td></tr><tr><td>text_config.encoder_layers</td><td>6</td><td>1</td></tr><tr><td>vision_config.depths</td><td>[1, 1, 9, 1]</td><td>[1, 1, 2, 1]</td></tr><tr><td>vision_config.embed_dim</td><td>[128, 256, 512, 1024]</td><td>[32, 64, 128, 256]</td></tr><tr><td>vision_config.num_groups</td><td>[4, 8, 16, 32]</td><td>[2, 4, 8, 16]</td></tr><tr><td>vision_config.num_heads</td><td>[4, 8, 16, 32]</td><td>[2, 4, 8, 16]</td></tr><tr><td>vision_config.projection_dim</td><td>768</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Florence-2-base-mic.json b/shrink/Florence-2-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..36bf256fa3de9da0a8de7bcedcf32a5e3d28309e
--- /dev/null
+++ b/shrink/Florence-2-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Florence-2-base-mic@85c1b12e1d9565763f6f2b9466669c58bc7aac33", "findings": []}
\ No newline at end of file
diff --git a/shrink/GLM-4-9B-0414-mic.html b/shrink/GLM-4-9B-0414-mic.html
index 1b1fc2dfe2de73cb2964dffacb00a96e16a4b991..ce82d703079edfd1bb10e1e6e5bf362fc6920051 100644
--- a/shrink/GLM-4-9B-0414-mic.html
+++ b/shrink/GLM-4-9B-0414-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>zai-org/GLM-4-9B-0414: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/GLM-4-9B-0414-mic@140a76e1cd7001bc3d0aed3b715a2a902214ece6 ← zai-org/GLM-4-9B-0414@645b8482494e31b6b752272bf7f7f273ef0f3caf · Glm4ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>zai-org/GLM-4-9B-0414</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/GLM-4-9B-0414-mic/tree/140a76e1cd7001bc3d0aed3b715a2a902214ece6" title="huggingface-mic/GLM-4-9B-0414-mic@140a76e1cd7001bc3d0aed3b715a2a902214ece6">huggingface-mic/GLM-4-9B-0414-mic@140a76e1cd70</a> ← <a href="https://huggingface.co/zai-org/GLM-4-9B-0414/tree/645b8482494e31b6b752272bf7f7f273ef0f3caf" title="zai-org/GLM-4-9B-0414@645b8482494e31b6b752272bf7f7f273ef0f3caf">zai-org/GLM-4-9B-0414@645b8482494e</a> · Glm4ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">9.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>40</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.867928" title="model.layers: 8.2B (86.8%)"></i><i class="c1" style="flex:0.066036" title="model.embed_tokens: 620.8M (6.6%)"></i><i class="c2" style="flex:0.066036" title="lm_head: 620.8M (6.6%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003826" title="model.layers: 74.5K (0.4%)"></i><i class="c1" style="flex:0.498086" title="model.embed_tokens: 9.7M (49.8%)"></i><i class="c2" style="flex:0.498086" title="lm_head: 9.7M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>8.2B</td><td>86.8%</td><td>74.5K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>620.8M</td><td>6.6%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>620.8M</td><td>6.6%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>glm4</td><td>glm4</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151552</td><td>151552</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Glm4 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>241</td><td>13</td></tr><tr class=""><td>RMSNorm</td><td>161</td><td>9</td></tr><tr class=""><td>Attention</td><td>40</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>MLP</td><td>40</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>40</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 23 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>13696</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>40</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/GLM-4-9B-0414-mic.json b/shrink/GLM-4-9B-0414-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..364fe894bf4fdb8e2c63324b66bf77cc4b1c58ed
--- /dev/null
+++ b/shrink/GLM-4-9B-0414-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/GLM-4-9B-0414-mic@140a76e1cd7001bc3d0aed3b715a2a902214ece6", "findings": []}
\ No newline at end of file
diff --git a/shrink/GLM-4.5-Air-mic.html b/shrink/GLM-4.5-Air-mic.html
index 2a87324fcfbf8e6088e9e59bbbd08ff249123167..27cc7ba09223c9d208fb1f25dce604f0137cd0eb 100644
--- a/shrink/GLM-4.5-Air-mic.html
+++ b/shrink/GLM-4.5-Air-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>zai-org/GLM-4.5-Air: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/GLM-4.5-Air-mic@abc7ea97d6d633936cfd54900edce3b610755f86 ← zai-org/GLM-4.5-Air@a24ceef6ce4f3536971efe9b778bdaa1bab18daa · Glm4MoeForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">106.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 46</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 2: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 6: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 10: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 14: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 18: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 22: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 26: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 30: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 34: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 38: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 42: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>45</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.988381" title="model.layers: 105.6B (98.8%)"></i><i class="c1" style="flex:0.005809" title="model.embed_tokens: 620.8M (0.6%)"></i><i class="c2" style="flex:0.005809" title="lm_head: 620.8M (0.6%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004140" title="model.layers: 80.6K (0.4%)"></i><i class="c1" style="flex:0.497928" title="model.embed_tokens: 9.7M (49.8%)"></i><i class="c2" style="flex:0.497928" title="lm_head: 9.7M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>105.6B</td><td>98.8%</td><td>80.6K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>620.8M</td><td>0.6%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>620.8M</td><td>0.6%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (96 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 128 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>glm4_moe</td><td>glm4_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151552</td><td>151552</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Glm4Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>323</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>93</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>91</td><td>3</td></tr><tr class=""><td>Attention</td><td>46</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>46</td><td>2</td></tr><tr class=""><td>MLP</td><td>46</td><td>2</td></tr><tr class=""><td>Experts</td><td>45</td><td>1</td></tr><tr class=""><td>MoE</td><td>45</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>45</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>10944</td><td>128</td></tr><tr><td>moe_intermediate_size</td><td>1408</td><td>32</td></tr><tr><td>n_routed_experts</td><td>128</td><td>4</td></tr><tr><td>num_attention_heads</td><td>96</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>46</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>zai-org/GLM-4.5-Air</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/GLM-4.5-Air-mic/tree/abc7ea97d6d633936cfd54900edce3b610755f86" title="huggingface-mic/GLM-4.5-Air-mic@abc7ea97d6d633936cfd54900edce3b610755f86">huggingface-mic/GLM-4.5-Air-mic@abc7ea97d6d6</a> ← <a href="https://huggingface.co/zai-org/GLM-4.5-Air/tree/a24ceef6ce4f3536971efe9b778bdaa1bab18daa" title="zai-org/GLM-4.5-Air@a24ceef6ce4f3536971efe9b778bdaa1bab18daa">zai-org/GLM-4.5-Air@a24ceef6ce4f</a> · Glm4MoeForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">106.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 46</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 2: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 6: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 10: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 14: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 18: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 22: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 26: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 30: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 34: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 38: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 42: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>45</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.988381" title="model.layers: 105.6B (98.8%)"></i><i class="c1" style="flex:0.005809" title="lm_head: 620.8M (0.6%)"></i><i class="c2" style="flex:0.005809" title="model.embed_tokens: 620.8M (0.6%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004140" title="model.layers: 80.6K (0.4%)"></i><i class="c1" style="flex:0.497928" title="lm_head: 9.7M (49.8%)"></i><i class="c2" style="flex:0.497928" title="model.embed_tokens: 9.7M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>105.6B</td><td>98.8%</td><td>80.6K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>620.8M</td><td>0.6%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>620.8M</td><td>0.6%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (96 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 128 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>glm4_moe</td><td>glm4_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151552</td><td>151552</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Glm4Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>323</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>93</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>91</td><td>3</td></tr><tr class=""><td>Attention</td><td>46</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>46</td><td>2</td></tr><tr class=""><td>MLP</td><td>46</td><td>2</td></tr><tr class=""><td>Experts</td><td>45</td><td>1</td></tr><tr class=""><td>MoE</td><td>45</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>45</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>10944</td><td>128</td></tr><tr><td>moe_intermediate_size</td><td>1408</td><td>32</td></tr><tr><td>n_routed_experts</td><td>128</td><td>4</td></tr><tr><td>num_attention_heads</td><td>96</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>46</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/GLM-4.5-Air-mic.json b/shrink/GLM-4.5-Air-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..49f97270cb3ca3dc052f5da36eacd457f9cde14b
--- /dev/null
+++ b/shrink/GLM-4.5-Air-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/GLM-4.5-Air-mic@abc7ea97d6d633936cfd54900edce3b610755f86", "findings": []}
\ No newline at end of file
diff --git a/shrink/GLM-4.7-Flash-mic.html b/shrink/GLM-4.7-Flash-mic.html
index 1e5d95f5bb7d82316a71c509a78503c1f3f5c1fc..b6b4ef23a35cde4083347571d0412ad689a637c1 100644
--- a/shrink/GLM-4.7-Flash-mic.html
+++ b/shrink/GLM-4.7-Flash-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>zai-org/GLM-4.7-Flash: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/GLM-4.7-Flash-mic@8b4c9af3fc9e6a8325e3dade3558514edd2cd694 ← zai-org/GLM-4.7-Flash@7dd20894a642a0aa287e9827cb1a1f7f91386b67 · Glm4MoeLiteForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>zai-org/GLM-4.7-Flash</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/GLM-4.7-Flash-mic/tree/8b4c9af3fc9e6a8325e3dade3558514edd2cd694" title="huggingface-mic/GLM-4.7-Flash-mic@8b4c9af3fc9e6a8325e3dade3558514edd2cd694">huggingface-mic/GLM-4.7-Flash-mic@8b4c9af3fc9e</a> ← <a href="https://huggingface.co/zai-org/GLM-4.7-Flash/tree/7dd20894a642a0aa287e9827cb1a1f7f91386b67" title="zai-org/GLM-4.7-Flash@7dd20894a642a0aa287e9827cb1a1f7f91386b67">zai-org/GLM-4.7-Flash@7dd20894a642</a> · Glm4MoeLiteForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">29.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 47</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 2: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 6: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 10: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 14: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 18: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 22: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 26: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 30: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 34: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 38: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 42: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 46: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>46</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.978814" title="model.layers: 29.3B (97.9%)"></i><i class="c1" style="flex:0.010593" title="model.embed_tokens: 317.2M (1.1%)"></i><i class="c2" style="flex:0.010593" title="lm_head: 317.2M (1.1%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003837" title="model.layers: 76.4K (0.4%)"></i><i class="c1" style="flex:0.498080" title="model.embed_tokens: 9.9M (49.8%)"></i><i class="c2" style="flex:0.498080" title="lm_head: 9.9M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>29.3B</td><td>97.9%</td><td>76.4K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>317.2M</td><td>1.1%</td><td>9.9M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>317.2M</td><td>1.1%</td><td>9.9M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (20 heads, 20 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-4 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>glm4_moe_lite</td><td>glm4_moe_lite</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>154880</td><td>154880</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>202752</td><td>202752</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Glm4MoeLite omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>377</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>189</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>93</td><td>3</td></tr><tr class=""><td>Attention</td><td>47</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>47</td><td>2</td></tr><tr class=""><td>MLP</td><td>47</td><td>2</td></tr><tr class=""><td>Experts</td><td>46</td><td>1</td></tr><tr class=""><td>MoE</td><td>46</td><td>1</td></tr><tr class=""><td>TopkRouter</td><td>46</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">14 fields changed, 34 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>10240</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>16</td></tr><tr><td>mlp_layer_types</td><td>47 entries: 46× sparse, 1× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>1536</td><td>32</td></tr><tr><td>n_routed_experts</td><td>64</td><td>4</td></tr><tr><td>num_attention_heads</td><td>20</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>4</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>47</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>20</td><td>4</td></tr><tr><td>q_lora_rank</td><td>768</td><td>16</td></tr><tr><td>qk_nope_head_dim</td><td>192</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>v_head_dim</td><td>256</td><td>16</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/GLM-4.7-Flash-mic.json b/shrink/GLM-4.7-Flash-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..0569b46cd1fc321d82a81039f5929c2c0f440fe0
--- /dev/null
+++ b/shrink/GLM-4.7-Flash-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/GLM-4.7-Flash-mic@8b4c9af3fc9e6a8325e3dade3558514edd2cd694", "findings": []}
\ No newline at end of file
diff --git a/shrink/GLM-5.3-Flash-mic.html b/shrink/GLM-5.3-Flash-mic.html
index 195d98ddc0e2053fc602b6a54d8ca81e1787b4d5..88316fcdacb6ae6b5d3ff0bbdce1008000f18bd5 100644
--- a/shrink/GLM-5.3-Flash-mic.html
+++ b/shrink/GLM-5.3-Flash-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>zai-org/GLM-5.3-Flash: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/GLM-5.3-Flash-mic@53f3e5bd10a67d2c75a170a3c2a28009e338c256 ← zai-org/GLM-5.3-Flash@46a050bec13002ad29ee795c16e61cf4fa3c4a9c · Glm5NextForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>zai-org/GLM-5.3-Flash</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/GLM-5.3-Flash-mic/tree/53f3e5bd10a67d2c75a170a3c2a28009e338c256" title="huggingface-mic/GLM-5.3-Flash-mic@53f3e5bd10a67d2c75a170a3c2a28009e338c256">huggingface-mic/GLM-5.3-Flash-mic@53f3e5bd10a6</a> ← <a href="https://huggingface.co/zai-org/GLM-5.3-Flash/tree/46a050bec13002ad29ee795c16e61cf4fa3c4a9c" title="zai-org/GLM-5.3-Flash@46a050bec13002ad29ee795c16e61cf4fa3c4a9c">zai-org/GLM-5.3-Flash@46a050bec130</a> · Glm5NextForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">313.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">40.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>vision_config: image_size 448 became 56</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 45</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, TextForgetGate …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, TextForgetGate …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, TextForgetGate …"></i><i class="c1" title="layer 3: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 4: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 5: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 6: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 7: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 8: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 9: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 10: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 11: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 12: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 13: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 14: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 15: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 16: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 17: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 18: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 19: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 20: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 21: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 22: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 23: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 24: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 25: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 26: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 27: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 28: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 29: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 30: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 31: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 32: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 33: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 34: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 35: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 36: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 37: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 38: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 39: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 40: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 41: linear_attention, Conv1d, TextExperts …"></i><i class="c2" title="layer 42: linear_attention, Conv1d, TextExperts …"></i><i class="c1" title="layer 43: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 44: linear_attention, Conv1d, TextExperts …"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, TextForgetGate …"></i><i class="c1" title="layer 1: is_causal=True, TextAttention, TextExperts …"></i><i class="c2" title="layer 2: linear_attention, Conv1d, TextExperts …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, TextForgetGate …</td><td>3</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, TextAttention, TextExperts …</td><td>11</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>linear_attention, Conv1d, TextExperts …</td><td>31</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.994162" title="model.language_model.layers: 312.1B (99.4%)"></i><i class="c1" style="flex:0.002021" title="model.language_model: 634.4M (0.2%)"></i><i class="c2" style="flex:0.002021" title="lm_head: 634.4M (0.2%)"></i><i class="c3" style="flex:0.001284" title="model.visual.blocks: 403.0M (0.1%)"></i><i class="c4" style="flex:0.000512" title="model.visual: 160.6M (0.1%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.019919" title="model.language_model.layers: 811.0K (2.0%)"></i><i class="c1" style="flex:0.486881" title="model.language_model: 19.8M (48.7%)"></i><i class="c2" style="flex:0.486877" title="lm_head: 19.8M (48.7%)"></i><i class="c3" style="flex:0.002048" title="model.visual.blocks: 83.4K (0.2%)"></i><i class="c4" style="flex:0.004275" title="model.visual: 174.1K (0.4%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>312.1B</td><td>99.4%</td><td>811.0K</td><td>2.0%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>634.4M</td><td>0.2%</td><td>19.8M</td><td>48.7%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>634.4M</td><td>0.2%</td><td>19.8M</td><td>48.7%</td></tr><tr><td><i class="c3"></i>model.visual.blocks</td><td>403.0M</td><td>0.1%</td><td>83.4K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>model.visual</td><td>160.6M</td><td>0.1%</td><td>174.1K</td><td>0.4%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>glm5_next</td><td>glm5_next</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MHA (64 heads, 64 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-8 of 288 experts</td><td>top-2 of 8 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>glm5_next_text</td><td>glm5_next_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>154880</td><td>154880</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>1048576</td><td>1048576</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>glm5_next_vision</td><td>glm5_next_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class="bad"><td>vision_config</td><td>image_size</td><td>448</td><td>56</td><td>⚠ changed</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Glm5Next omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>654</td><td>50</td></tr><tr class=""><td>TextRMSNorm</td><td>113</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>97</td><td>9</td></tr><tr class=""><td>TextHyperConnection</td><td>90</td><td>6</td></tr><tr class=""><td>TextUnweightedRMSNorm</td><td>90</td><td>6</td></tr><tr class=""><td>SiLUActivation</td><td>70</td><td>6</td></tr><tr class=""><td>TextDecoderLayer</td><td>45</td><td>3</td></tr><tr class=""><td>TextMLP</td><td>45</td><td>3</td></tr><tr class=""><td>TextExperts</td><td>42</td><td>2</td></tr><tr class=""><td>TextMoE</td><td>42</td><td>2</td></tr><tr class=""><td>TextTopkRouter</td><td>42</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>34</td><td>2</td></tr><tr class=""><td>TextForgetGate</td><td>34</td><td>2</td></tr><tr class=""><td>TextLinearAttention</td><td>34</td><td>2</td></tr><tr class=""><td>TextRMSNormGated</td><td>34</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>24</td><td>2</td></tr><tr class=""><td>VisionBlock</td><td>24</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>24</td><td>2</td></tr><tr class=""><td>LayerNorm</td><td>12</td><td>2</td></tr><tr class=""><td>TextAttention</td><td>11</td><td>1</td></tr><tr class=""><td>TextIndexer</td><td>11</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextHyperHead</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">33 fields changed, 86 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;e4m3&quot;, &quot;modules_to_not_convert&quot;: [&quot;a…</td><td>—</td></tr><tr><td>text_config.first_k_dense_replace</td><td>3</td><td>1</td></tr><tr><td>text_config.hidden_size</td><td>4096</td><td>128</td></tr><tr><td>text_config.index_head_dim</td><td>128</td><td>32</td></tr><tr><td>text_config.index_n_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.index_topk</td><td>2048</td><td>64</td></tr><tr><td>text_config.indexer_types</td><td>45 entries: 45× full</td><td>[&quot;full&quot;, &quot;full&quot;, &quot;full&quot;]</td></tr><tr><td>text_config.intermediate_size</td><td>12288</td><td>128</td></tr><tr><td>text_config.kv_lora_rank</td><td>512</td><td>64</td></tr><tr><td>text_config.layer_types</td><td>45 entries: 34× linear_attention, 11× indexed_attention</td><td>[&quot;linear_attention&quot;, &quot;indexed_attention&quot;, &quot;linear_attention&quot;]</td></tr><tr><td>text_config.linear_attn_config.full_attn_layers</td><td>11 entries</td><td>[1]</td></tr><tr><td>text_config.linear_attn_config.head_dim</td><td>128</td><td>32</td></tr><tr><td>text_config.linear_attn_config.kda_layers</td><td>34 entries</td><td>[0, 2]</td></tr><tr><td>text_config.linear_attn_config.num_heads</td><td>64</td><td>4</td></tr><tr><td>text_config.linear_head_dim</td><td>128</td><td>32</td></tr><tr><td>text_config.linear_num_heads</td><td>64</td><td>4</td></tr><tr><td>text_config.mlp_layer_types</td><td>45 entries: 42× sparse, 3× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;, &quot;sparse&quot;]</td></tr><tr><td>text_config.moe_intermediate_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.n_routed_experts</td><td>288</td><td>8</td></tr><tr><td>text_config.num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>45</td><td>3</td></tr><tr><td>text_config.num_key_value_heads</td><td>64</td><td>4</td></tr><tr><td>text_config.q_lora_rank</td><td>1536</td><td>64</td></tr><tr><td>text_config.qk_nope_head_dim</td><td>256</td><td>32</td></tr><tr><td>text_config.v_head_dim</td><td>256</td><td>32</td></tr><tr><td>vision_config.depth</td><td>24</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.image_size</td><td>448</td><td>56</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.projection_intermediate_size</td><td>10240</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/GLM-5.3-Flash-mic.json b/shrink/GLM-5.3-Flash-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..03e6d797089af667e676e7f07e8e3189186a431e
--- /dev/null
+++ b/shrink/GLM-5.3-Flash-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/GLM-5.3-Flash-mic@53f3e5bd10a67d2c75a170a3c2a28009e338c256", "findings": ["vision_config: image_size 448 became 56"]}
\ No newline at end of file
diff --git a/shrink/GLM-5.3-mic.html b/shrink/GLM-5.3-mic.html
index 88a9d0280a9276b0b57dd2be2f2ea3f074d17eb2..56ff056133487c03d0b3ebefbf40559a6e17e174 100644
--- a/shrink/GLM-5.3-mic.html
+++ b/shrink/GLM-5.3-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>zai-org/GLM-5.3: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/GLM-5.3-mic@1088c4f5a18bf3ba097240a67f6c391be8072006 ← zai-org/GLM-5.3@ba32f8c19c8738e2c57c71798c601d477a7bf844 · GlmMoeDsaForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">743.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">41.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>model: max_position_embeddings 1048576 became 512</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 78</span><div class="cells"><i class="c0" title="layer 0: Indexer, LayerNorm"></i><i class="c0" title="layer 1: Indexer, LayerNorm"></i><i class="c0" title="layer 2: Indexer, LayerNorm"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 6: Experts, Indexer, MoE …"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 10: Experts, Indexer, MoE …"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 14: Experts, Indexer, MoE …"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 18: Experts, Indexer, MoE …"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 22: Experts, Indexer, MoE …"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 26: Experts, Indexer, MoE …"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 30: Experts, Indexer, MoE …"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 34: Experts, Indexer, MoE …"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 38: Experts, Indexer, MoE …"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 42: Experts, Indexer, MoE …"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 46: Experts, Indexer, MoE …"></i><i class="c1" title="layer 47: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 48: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 49: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 50: Experts, Indexer, MoE …"></i><i class="c1" title="layer 51: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 52: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 53: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 54: Experts, Indexer, MoE …"></i><i class="c1" title="layer 55: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 56: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 57: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 58: Experts, Indexer, MoE …"></i><i class="c1" title="layer 59: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 60: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 61: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 62: Experts, Indexer, MoE …"></i><i class="c1" title="layer 63: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 64: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 65: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 66: Experts, Indexer, MoE …"></i><i class="c1" title="layer 67: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 68: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 69: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 70: Experts, Indexer, MoE …"></i><i class="c1" title="layer 71: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 72: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 73: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 74: Experts, Indexer, MoE …"></i><i class="c1" title="layer 75: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 76: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 77: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 7</span><div class="cells"><i class="c0" title="layer 0: Indexer, LayerNorm"></i><i class="c0" title="layer 1: Indexer, LayerNorm"></i><i class="c0" title="layer 2: Indexer, LayerNorm"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 6: Experts, Indexer, MoE …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Indexer, LayerNorm</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>57</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>Experts, Indexer, MoE …</td><td>18</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.997440" title="model.layers: 741.5B (99.7%)"></i><i class="c1" style="flex:0.001280" title="lm_head: 951.6M (0.1%)"></i><i class="c2" style="flex:0.001280" title="model.embed_tokens: 951.6M (0.1%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 6.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.033438" title="model.layers: 1.4M (3.3%)"></i><i class="c1" style="flex:0.483279" title="lm_head: 19.8M (48.3%)"></i><i class="c2" style="flex:0.483279" title="model.embed_tokens: 19.8M (48.3%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>741.5B</td><td>99.7%</td><td>1.4M</td><td>3.3%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>951.6M</td><td>0.1%</td><td>19.8M</td><td>48.3%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>951.6M</td><td>0.1%</td><td>19.8M</td><td>48.3%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>6.1K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (64 heads, 64 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 8 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>glm_moe_dsa</td><td>glm_moe_dsa</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>154880</td><td>154880</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class="bad"><td>model</td><td>max_position_embeddings</td><td>1048576</td><td>512</td><td>⚠ changed</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GlmMoeDsa omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>688</td><td>69</td></tr><tr class=""><td>RMSNorm</td><td>313</td><td>29</td></tr><tr class=""><td>SiLUActivation</td><td>153</td><td>11</td></tr><tr class=""><td>Attention</td><td>78</td><td>7</td></tr><tr class=""><td>DecoderLayer</td><td>78</td><td>7</td></tr><tr class=""><td>MLP</td><td>78</td><td>7</td></tr><tr class=""><td>Experts</td><td>75</td><td>4</td></tr><tr class=""><td>MoE</td><td>75</td><td>4</td></tr><tr class=""><td>TopkRouter</td><td>75</td><td>4</td></tr><tr class=""><td>Indexer</td><td>21</td><td>4</td></tr><tr class=""><td>LayerNorm</td><td>21</td><td>4</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">23 fields changed, 42 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>first_k_dense_replace</td><td>3</td><td>1</td></tr><tr><td>head_dim</td><td>64</td><td>8</td></tr><tr><td>hidden_size</td><td>6144</td><td>128</td></tr><tr><td>index_head_dim</td><td>128</td><td>16</td></tr><tr><td>index_n_heads</td><td>32</td><td>2</td></tr><tr><td>index_topk</td><td>2048</td><td>8</td></tr><tr><td>indexer_types</td><td>78 entries: 57× shared, 21× full</td><td>7 entries: 4× full, 3× shared</td></tr><tr><td>intermediate_size</td><td>12288</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>32</td></tr><tr><td>layer_types</td><td>78 entries: 78× indexed_attention</td><td>7 entries: 7× indexed_attention</td></tr><tr><td>max_position_embeddings</td><td>1048576</td><td>512</td></tr><tr><td>mlp_layer_types</td><td>78 entries: 75× sparse, 3× dense</td><td>7 entries: 4× sparse, 3× dense</td></tr><tr><td>moe_intermediate_size</td><td>2048</td><td>64</td></tr><tr><td>n_routed_experts</td><td>256</td><td>8</td></tr><tr><td>num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>78</td><td>7</td></tr><tr><td>num_key_value_heads</td><td>64</td><td>4</td></tr><tr><td>q_lora_rank</td><td>2048</td><td>64</td></tr><tr><td>qk_nope_head_dim</td><td>192</td><td>24</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;e4m3&quot;, &quot;quant_method&quot;: &quot;fp8&quot;, &quot;weigh…</td><td>—</td></tr><tr><td>v_head_dim</td><td>256</td><td>32</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>zai-org/GLM-5.3</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/GLM-5.3-mic/tree/1088c4f5a18bf3ba097240a67f6c391be8072006" title="huggingface-mic/GLM-5.3-mic@1088c4f5a18bf3ba097240a67f6c391be8072006">huggingface-mic/GLM-5.3-mic@1088c4f5a18b</a> ← <a href="https://huggingface.co/zai-org/GLM-5.3/tree/ba32f8c19c8738e2c57c71798c601d477a7bf844" title="zai-org/GLM-5.3@ba32f8c19c8738e2c57c71798c601d477a7bf844">zai-org/GLM-5.3@ba32f8c19c87</a> · GlmMoeDsaForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">743.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">41.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>model: max_position_embeddings 1048576 became 512</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 78</span><div class="cells"><i class="c0" title="layer 0: Indexer, LayerNorm"></i><i class="c0" title="layer 1: Indexer, LayerNorm"></i><i class="c0" title="layer 2: Indexer, LayerNorm"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 6: Experts, Indexer, MoE …"></i><i class="c1" title="layer 7: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 10: Experts, Indexer, MoE …"></i><i class="c1" title="layer 11: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 14: Experts, Indexer, MoE …"></i><i class="c1" title="layer 15: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 18: Experts, Indexer, MoE …"></i><i class="c1" title="layer 19: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 22: Experts, Indexer, MoE …"></i><i class="c1" title="layer 23: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 26: Experts, Indexer, MoE …"></i><i class="c1" title="layer 27: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 30: Experts, Indexer, MoE …"></i><i class="c1" title="layer 31: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 34: Experts, Indexer, MoE …"></i><i class="c1" title="layer 35: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 38: Experts, Indexer, MoE …"></i><i class="c1" title="layer 39: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 42: Experts, Indexer, MoE …"></i><i class="c1" title="layer 43: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 46: Experts, Indexer, MoE …"></i><i class="c1" title="layer 47: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 48: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 49: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 50: Experts, Indexer, MoE …"></i><i class="c1" title="layer 51: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 52: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 53: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 54: Experts, Indexer, MoE …"></i><i class="c1" title="layer 55: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 56: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 57: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 58: Experts, Indexer, MoE …"></i><i class="c1" title="layer 59: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 60: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 61: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 62: Experts, Indexer, MoE …"></i><i class="c1" title="layer 63: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 64: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 65: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 66: Experts, Indexer, MoE …"></i><i class="c1" title="layer 67: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 68: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 69: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 70: Experts, Indexer, MoE …"></i><i class="c1" title="layer 71: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 72: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 73: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 74: Experts, Indexer, MoE …"></i><i class="c1" title="layer 75: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 76: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 77: Experts, MoE, TopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 7</span><div class="cells"><i class="c0" title="layer 0: Indexer, LayerNorm"></i><i class="c0" title="layer 1: Indexer, LayerNorm"></i><i class="c0" title="layer 2: Indexer, LayerNorm"></i><i class="c1" title="layer 3: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopkRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopkRouter"></i><i class="c2" title="layer 6: Experts, Indexer, MoE …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Indexer, LayerNorm</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopkRouter</td><td>57</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>Experts, Indexer, MoE …</td><td>18</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.997440" title="model.layers: 741.5B (99.7%)"></i><i class="c1" style="flex:0.001280" title="model.embed_tokens: 951.6M (0.1%)"></i><i class="c2" style="flex:0.001280" title="lm_head: 951.6M (0.1%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 6.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.033438" title="model.layers: 1.4M (3.3%)"></i><i class="c1" style="flex:0.483279" title="model.embed_tokens: 19.8M (48.3%)"></i><i class="c2" style="flex:0.483279" title="lm_head: 19.8M (48.3%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>741.5B</td><td>99.7%</td><td>1.4M</td><td>3.3%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>951.6M</td><td>0.1%</td><td>19.8M</td><td>48.3%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>951.6M</td><td>0.1%</td><td>19.8M</td><td>48.3%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>6.1K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (64 heads, 64 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 8 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>glm_moe_dsa</td><td>glm_moe_dsa</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>154880</td><td>154880</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class="bad"><td>model</td><td>max_position_embeddings</td><td>1048576</td><td>512</td><td>⚠ changed</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GlmMoeDsa omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>688</td><td>69</td></tr><tr class=""><td>RMSNorm</td><td>313</td><td>29</td></tr><tr class=""><td>SiLUActivation</td><td>153</td><td>11</td></tr><tr class=""><td>Attention</td><td>78</td><td>7</td></tr><tr class=""><td>DecoderLayer</td><td>78</td><td>7</td></tr><tr class=""><td>MLP</td><td>78</td><td>7</td></tr><tr class=""><td>Experts</td><td>75</td><td>4</td></tr><tr class=""><td>MoE</td><td>75</td><td>4</td></tr><tr class=""><td>TopkRouter</td><td>75</td><td>4</td></tr><tr class=""><td>Indexer</td><td>21</td><td>4</td></tr><tr class=""><td>LayerNorm</td><td>21</td><td>4</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">23 fields changed, 42 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>first_k_dense_replace</td><td>3</td><td>1</td></tr><tr><td>head_dim</td><td>64</td><td>8</td></tr><tr><td>hidden_size</td><td>6144</td><td>128</td></tr><tr><td>index_head_dim</td><td>128</td><td>16</td></tr><tr><td>index_n_heads</td><td>32</td><td>2</td></tr><tr><td>index_topk</td><td>2048</td><td>8</td></tr><tr><td>indexer_types</td><td>78 entries: 57× shared, 21× full</td><td>7 entries: 4× full, 3× shared</td></tr><tr><td>intermediate_size</td><td>12288</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>512</td><td>32</td></tr><tr><td>layer_types</td><td>78 entries: 78× indexed_attention</td><td>7 entries: 7× indexed_attention</td></tr><tr><td>max_position_embeddings</td><td>1048576</td><td>512</td></tr><tr><td>mlp_layer_types</td><td>78 entries: 75× sparse, 3× dense</td><td>7 entries: 4× sparse, 3× dense</td></tr><tr><td>moe_intermediate_size</td><td>2048</td><td>64</td></tr><tr><td>n_routed_experts</td><td>256</td><td>8</td></tr><tr><td>num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>78</td><td>7</td></tr><tr><td>num_key_value_heads</td><td>64</td><td>4</td></tr><tr><td>q_lora_rank</td><td>2048</td><td>64</td></tr><tr><td>qk_nope_head_dim</td><td>192</td><td>24</td></tr><tr><td>qk_rope_head_dim</td><td>64</td><td>8</td></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;e4m3&quot;, &quot;quant_method&quot;: &quot;fp8&quot;, &quot;weigh…</td><td>—</td></tr><tr><td>v_head_dim</td><td>256</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/GLM-5.3-mic.json b/shrink/GLM-5.3-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..008151574d968583ff76376ac53370a89db2f37d
--- /dev/null
+++ b/shrink/GLM-5.3-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/GLM-5.3-mic@1088c4f5a18bf3ba097240a67f6c391be8072006", "findings": ["model: max_position_embeddings 1048576 became 512"]}
\ No newline at end of file
diff --git a/shrink/GLM-OCR-mic.html b/shrink/GLM-OCR-mic.html
index 97e3991ccd975528145b818ba0735d030a36aa0d..7ee38168fe7e1745e786f9e388b4e08001134a6b 100644
--- a/shrink/GLM-OCR-mic.html
+++ b/shrink/GLM-OCR-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>zai-org/GLM-OCR: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/GLM-OCR-mic@08eddf7ae838d5d3b6b135b984c0e3ec53c28441 ← zai-org/GLM-OCR@2e85a62840ccac27daa451df36c736c4636b8628 · GlmOcrForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>zai-org/GLM-OCR</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/GLM-OCR-mic/tree/08eddf7ae838d5d3b6b135b984c0e3ec53c28441" title="huggingface-mic/GLM-OCR-mic@08eddf7ae838d5d3b6b135b984c0e3ec53c28441">huggingface-mic/GLM-OCR-mic@08eddf7ae838</a> ← <a href="https://huggingface.co/zai-org/GLM-OCR/tree/2e85a62840ccac27daa451df36c736c4636b8628" title="zai-org/GLM-OCR@2e85a62840ccac27daa451df36c736c4636b8628">zai-org/GLM-OCR@2e85a62840cc</a> · GlmOcrForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 24</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>24</td><td>24</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i><i class="c0" title="layer 2: TextDecoderLayer"></i><i class="c0" title="layer 3: TextDecoderLayer"></i><i class="c0" title="layer 4: TextDecoderLayer"></i><i class="c0" title="layer 5: TextDecoderLayer"></i><i class="c0" title="layer 6: TextDecoderLayer"></i><i class="c0" title="layer 7: TextDecoderLayer"></i><i class="c0" title="layer 8: TextDecoderLayer"></i><i class="c0" title="layer 9: TextDecoderLayer"></i><i class="c0" title="layer 10: TextDecoderLayer"></i><i class="c0" title="layer 11: TextDecoderLayer"></i><i class="c0" title="layer 12: TextDecoderLayer"></i><i class="c0" title="layer 13: TextDecoderLayer"></i><i class="c0" title="layer 14: TextDecoderLayer"></i><i class="c0" title="layer 15: TextDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextDecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.443227" title="model.language_model.layers: 490.8M (44.3%)"></i><i class="c1" style="flex:0.363936" title="model.visual.blocks: 403.0M (36.4%)"></i><i class="c2" style="flex:0.082380" title="model.language_model: 91.2M (8.2%)"></i><i class="c3" style="flex:0.082378" title="lm_head: 91.2M (8.2%)"></i><i class="c4" style="flex:0.028079" title="model.visual: 31.1M (2.8%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008427" title="model.language_model.layers: 74.2K (0.8%)"></i><i class="c1" style="flex:0.113586" title="model.visual.blocks: 1.0M (11.4%)"></i><i class="c2" style="flex:0.431453" title="model.language_model: 3.8M (43.1%)"></i><i class="c3" style="flex:0.431446" title="lm_head: 3.8M (43.1%)"></i><i class="c4" style="flex:0.015088" title="model.visual: 132.9K (1.5%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>490.8M</td><td>44.3%</td><td>74.2K</td><td>0.8%</td></tr><tr><td><i class="c1"></i>model.visual.blocks</td><td>403.0M</td><td>36.4%</td><td>1.0M</td><td>11.4%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>91.2M</td><td>8.2%</td><td>3.8M</td><td>43.1%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>91.2M</td><td>8.2%</td><td>3.8M</td><td>43.1%</td></tr><tr><td><i class="c4"></i>model.visual</td><td>31.1M</td><td>2.8%</td><td>132.9K</td><td>1.5%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>glm_ocr</td><td>glm_ocr</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>glm_ocr_text</td><td>glm_ocr_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>59392</td><td>59392</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>glm_ocr_vision</td><td>glm_ocr_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>336</td><td>336</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GlmOcr omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>221</td><td>137</td></tr><tr class=""><td>RMSNorm</td><td>162</td><td>106</td></tr><tr class=""><td>SiLUActivation</td><td>41</td><td>27</td></tr><tr class=""><td>VisionAttention</td><td>24</td><td>24</td></tr><tr class=""><td>VisionBlock</td><td>24</td><td>24</td></tr><tr class=""><td>VisionMlp</td><td>24</td><td>24</td></tr><tr class=""><td>TextAttention</td><td>16</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>TextMLP</td><td>16</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>LayerNorm</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 59 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>4608</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_section&quot;: [16, 24, 24], &quot;partial_rotary_factor&quot;: 1.0, &quot;rope_theta&quot;: 1…</td><td>{&quot;mrope_section&quot;: [2, 3, 3], &quot;partial_rotary_factor&quot;: 1.0, &quot;rope_theta&quot;: 1000…</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>1536</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/GLM-OCR-mic.json b/shrink/GLM-OCR-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..6ba45ffdedf9b78228fe4a004e1cc90415758ea8
--- /dev/null
+++ b/shrink/GLM-OCR-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/GLM-OCR-mic@08eddf7ae838d5d3b6b135b984c0e3ec53c28441", "findings": []}
\ No newline at end of file
diff --git a/shrink/Hunyuan-0.5B-Instruct-mic.html b/shrink/Hunyuan-0.5B-Instruct-mic.html
index 0a1e7f36839602ac2a1e659d0b511d9cabe807d0..ecfc640836fbec86626b902a73b1ee71d25605b1 100644
--- a/shrink/Hunyuan-0.5B-Instruct-mic.html
+++ b/shrink/Hunyuan-0.5B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>tencent/Hunyuan-0.5B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Hunyuan-0.5B-Instruct-mic@d7fe47dc3288bba155076d3747959365159073b0 ← tencent/Hunyuan-0.5B-Instruct@2359fb220c010e9d6d62c62d466f0eda179c2cf3 · HunYuanDenseV1ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>tencent/Hunyuan-0.5B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Hunyuan-0.5B-Instruct-mic/tree/d7fe47dc3288bba155076d3747959365159073b0" title="huggingface-mic/Hunyuan-0.5B-Instruct-mic@d7fe47dc3288bba155076d3747959365159073b0">huggingface-mic/Hunyuan-0.5B-Instruct-mic@d7fe47dc3288</a> ← <a href="https://huggingface.co/tencent/Hunyuan-0.5B-Instruct/tree/2359fb220c010e9d6d62c62d466f0eda179c2cf3" title="tencent/Hunyuan-0.5B-Instruct@2359fb220c010e9d6d62c62d466f0eda179c2cf3">tencent/Hunyuan-0.5B-Instruct@2359fb220c01</a> · HunYuanDenseV1ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">539.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">7.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.770471" title="model.layers: 415.3M (77.0%)"></i><i class="c1" style="flex:0.229528" title="model.embed_tokens: 123.7M (23.0%)"></i><i class="c2" style="flex:0.000002" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.009485" title="model.layers: 74.0K (0.9%)"></i><i class="c1" style="flex:0.990506" title="model.embed_tokens: 7.7M (99.1%)"></i><i class="c2" style="flex:0.000008" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>415.3M</td><td>77.0%</td><td>74.0K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>123.7M</td><td>23.0%</td><td>7.7M</td><td>99.1%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>dynamic</td><td>dynamic</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>hunyuan_v1_dense</td><td>hunyuan_v1_dense</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>120818</td><td>120818</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix HunYuanDenseV1 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>169</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>97</td><td>9</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 43 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>attention_head_dim</td><td>128</td><td>16</td></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>3584</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Hunyuan-0.5B-Instruct-mic.json b/shrink/Hunyuan-0.5B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..68928533a3ff5561a688f50b8c617c29b9ce861f
--- /dev/null
+++ b/shrink/Hunyuan-0.5B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Hunyuan-0.5B-Instruct-mic@d7fe47dc3288bba155076d3747959365159073b0", "findings": []}
\ No newline at end of file
diff --git a/shrink/HunyuanOCR-mic.html b/shrink/HunyuanOCR-mic.html
index 831f37f958e911bd57b2b04ac5972438408c987b..6274bfe9b85d84b7f024584866a0a3f95f65962b 100644
--- a/shrink/HunyuanOCR-mic.html
+++ b/shrink/HunyuanOCR-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>tencent/HunyuanOCR: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/HunyuanOCR-mic@2b69a85316ef4335c59c1a82d624555a1dfd7e71 ← tencent/HunyuanOCR@47644ecc4fc854efa4f505155158831f36773ee4 · HunYuanVLForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>tencent/HunyuanOCR</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/HunyuanOCR-mic/tree/2b69a85316ef4335c59c1a82d624555a1dfd7e71" title="huggingface-mic/HunyuanOCR-mic@2b69a85316ef4335c59c1a82d624555a1dfd7e71">huggingface-mic/HunyuanOCR-mic@2b69a85316ef</a> ← <a href="https://huggingface.co/tencent/HunyuanOCR/tree/47644ecc4fc854efa4f505155158831f36773ee4" title="tencent/HunyuanOCR@47644ecc4fc854efa4f505155158831f36773ee4">tencent/HunyuanOCR@47644ecc4fc8</a> · HunYuanVLForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">996.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">9.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: DenseV1DecoderLayer"></i><i class="c0" title="layer 1: DenseV1DecoderLayer"></i><i class="c0" title="layer 2: DenseV1DecoderLayer"></i><i class="c0" title="layer 3: DenseV1DecoderLayer"></i><i class="c0" title="layer 4: DenseV1DecoderLayer"></i><i class="c0" title="layer 5: DenseV1DecoderLayer"></i><i class="c0" title="layer 6: DenseV1DecoderLayer"></i><i class="c0" title="layer 7: DenseV1DecoderLayer"></i><i class="c0" title="layer 8: DenseV1DecoderLayer"></i><i class="c0" title="layer 9: DenseV1DecoderLayer"></i><i class="c0" title="layer 10: DenseV1DecoderLayer"></i><i class="c0" title="layer 11: DenseV1DecoderLayer"></i><i class="c0" title="layer 12: DenseV1DecoderLayer"></i><i class="c0" title="layer 13: DenseV1DecoderLayer"></i><i class="c0" title="layer 14: DenseV1DecoderLayer"></i><i class="c0" title="layer 15: DenseV1DecoderLayer"></i><i class="c0" title="layer 16: DenseV1DecoderLayer"></i><i class="c0" title="layer 17: DenseV1DecoderLayer"></i><i class="c0" title="layer 18: DenseV1DecoderLayer"></i><i class="c0" title="layer 19: DenseV1DecoderLayer"></i><i class="c0" title="layer 20: DenseV1DecoderLayer"></i><i class="c0" title="layer 21: DenseV1DecoderLayer"></i><i class="c0" title="layer 22: DenseV1DecoderLayer"></i><i class="c0" title="layer 23: DenseV1DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DenseV1DecoderLayer"></i><i class="c0" title="layer 1: DenseV1DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DenseV1DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vision_tower.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.416872" title="model.language_model.layers: 415.3M (41.7%)"></i><i class="c1" style="flex:0.413033" title="model.vision_tower.layers: 411.5M (41.3%)"></i><i class="c2" style="flex:0.124190" title="model.language_model: 123.7M (12.4%)"></i><i class="c3" style="flex:0.045906" title="model.vision_tower: 45.7M (4.6%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008178" title="model.language_model.layers: 74.0K (0.8%)"></i><i class="c1" style="flex:0.007394" title="model.vision_tower.layers: 66.9K (0.7%)"></i><i class="c2" style="flex:0.854014" title="model.language_model: 7.7M (85.4%)"></i><i class="c3" style="flex:0.130414" title="model.vision_tower: 1.2M (13.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>415.3M</td><td>41.7%</td><td>74.0K</td><td>0.8%</td></tr><tr><td><i class="c1"></i>model.vision_tower.layers</td><td>411.5M</td><td>41.3%</td><td>66.9K</td><td>0.7%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>123.7M</td><td>12.4%</td><td>7.7M</td><td>85.4%</td></tr><tr><td><i class="c3"></i>model.vision_tower</td><td>45.7M</td><td>4.6%</td><td>1.2M</td><td>13.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>hunyuan_vl</td><td>hunyuan_vl</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>dynamic</td><td>dynamic</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>hunyuan_vl_text</td><td>hunyuan_vl_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>120818</td><td>120818</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>hunyuan_vl_vision</td><td>hunyuan_vl_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix HunYuanVL omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>332</td><td>28</td></tr><tr class=""><td>RMSNorm</td><td>99</td><td>11</td></tr><tr class=""><td>LayerNorm</td><td>54</td><td>4</td></tr><tr class=""><td>GELUActivation</td><td>28</td><td>3</td></tr><tr class=""><td>VisionAttention</td><td>27</td><td>2</td></tr><tr class=""><td>VisionBlock</td><td>27</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>27</td><td>2</td></tr><tr class=""><td>DenseV1Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DenseV1DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>VisionTransformer</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">14 fields changed, 68 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3584</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;alpha&quot;: 1000.0, &quot;beta_fast&quot;: 32, &quot;beta_slow&quot;: 1, &quot;factor&quot;: 1.0, &quot;mscale&quot;: 1…</td><td>{&quot;alpha&quot;: 1000.0, &quot;beta_fast&quot;: 32, &quot;beta_slow&quot;: 1, &quot;factor&quot;: 1.0, &quot;mrope_sect…</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>27</td><td>2</td></tr><tr><td>vision_config.num_key_value_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.text_hidden_size</td><td>1024</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/HunyuanOCR-mic.json b/shrink/HunyuanOCR-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..2697c3a7622678b901f45134a2abbe2982bdc4a8
--- /dev/null
+++ b/shrink/HunyuanOCR-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/HunyuanOCR-mic@2b69a85316ef4335c59c1a82d624555a1dfd7e71", "findings": []}
\ No newline at end of file
diff --git a/shrink/Hy-MT2-30B-A3B-mic.html b/shrink/Hy-MT2-30B-A3B-mic.html
index 3b8dd981ebed320cd3ecd6b12a0b01152b226d50..6d1bc41d99fda41fc1649eec968e3cc66e3dcb07 100644
--- a/shrink/Hy-MT2-30B-A3B-mic.html
+++ b/shrink/Hy-MT2-30B-A3B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>tencent/Hy-MT2-30B-A3B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Hy-MT2-30B-A3B-mic@8357a369b5aa965d0f486cbe4fe63329d0a1d00e ← tencent/Hy-MT2-30B-A3B@d3ead4dba61c09aac60a261a96ad1df3e705febb · HYV3ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>tencent/Hy-MT2-30B-A3B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Hy-MT2-30B-A3B-mic/tree/8357a369b5aa965d0f486cbe4fe63329d0a1d00e" title="huggingface-mic/Hy-MT2-30B-A3B-mic@8357a369b5aa965d0f486cbe4fe63329d0a1d00e">huggingface-mic/Hy-MT2-30B-A3B-mic@8357a369b5aa</a> ← <a href="https://huggingface.co/tencent/Hy-MT2-30B-A3B/tree/d3ead4dba61c09aac60a261a96ad1df3e705febb" title="tencent/Hy-MT2-30B-A3B@d3ead4dba61c09aac60a261a96ad1df3e705febb">tencent/Hy-MT2-30B-A3B@d3ead4dba61c</a> · HYV3ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">30.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">15.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 2: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 3: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 4: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 5: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 6: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 7: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 8: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 9: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 10: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 11: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 12: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 13: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 14: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 15: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 16: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 17: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 18: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 19: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 20: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 21: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 22: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 23: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 24: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 25: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 26: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 27: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 28: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 29: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 30: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 31: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 32: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 33: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 34: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 35: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 36: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 37: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 38: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 39: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 40: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 41: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 42: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 43: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 44: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 45: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 46: Experts, MoE, TopKRouter"></i><i class="c1" title="layer 47: Experts, MoE, TopKRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, MoE, TopKRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, MoE, TopKRouter</td><td>47</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.983538" title="model.layers: 29.6B (98.4%)"></i><i class="c1" style="flex:0.008231" title="lm_head: 247.5M (0.8%)"></i><i class="c2" style="flex:0.008231" title="model.embed_tokens: 247.5M (0.8%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.005175" title="model.layers: 80.4K (0.5%)"></i><i class="c1" style="flex:0.497411" title="lm_head: 7.7M (49.7%)"></i><i class="c2" style="flex:0.497411" title="model.embed_tokens: 7.7M (49.7%)"></i><i class="c3" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>29.6B</td><td>98.4%</td><td>80.4K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>247.5M</td><td>0.8%</td><td>7.7M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>247.5M</td><td>0.8%</td><td>7.7M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 128 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>hy_v3</td><td>hy_v3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>120832</td><td>120832</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix HYV3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>337</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>193</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>95</td><td>3</td></tr><tr class=""><td>Attention</td><td>48</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>48</td><td>2</td></tr><tr class=""><td>MLP</td><td>48</td><td>2</td></tr><tr class=""><td>Experts</td><td>47</td><td>1</td></tr><tr class=""><td>MoE</td><td>47</td><td>1</td></tr><tr class=""><td>TopKRouter</td><td>47</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 39 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>6912</td><td>128</td></tr><tr><td>mlp_layer_types</td><td>48 entries: 47× sparse, 1× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>768</td><td>32</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_experts</td><td>128</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>48</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Hy-MT2-30B-A3B-mic.json b/shrink/Hy-MT2-30B-A3B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..01a54d165acf225168004db53458d777caae44b6
--- /dev/null
+++ b/shrink/Hy-MT2-30B-A3B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Hy-MT2-30B-A3B-mic@8357a369b5aa965d0f486cbe4fe63329d0a1d00e", "findings": []}
\ No newline at end of file
diff --git a/shrink/Inkling-mic.html b/shrink/Inkling-mic.html
index 6b7a2b5d8dcfc6648b3fffa3c0bd6ce90dda0f6b..649fefd512f35c61dc0541be5dbaefc562c59af8 100644
--- a/shrink/Inkling-mic.html
+++ b/shrink/Inkling-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>thinkingmachines/Inkling: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Inkling-mic@a7c61810959524051418bc1a49a70ac559307c0b ← thinkingmachines/Inkling@828496eeae4c243ff1a22f7f28ff83694f2f7bc9 · InklingForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>thinkingmachines/Inkling</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Inkling-mic/tree/a7c61810959524051418bc1a49a70ac559307c0b" title="huggingface-mic/Inkling-mic@a7c61810959524051418bc1a49a70ac559307c0b">huggingface-mic/Inkling-mic@a7c618109595</a> ← <a href="https://huggingface.co/thinkingmachines/Inkling/tree/828496eeae4c243ff1a22f7f28ff83694f2f7bc9" title="thinkingmachines/Inkling@828496eeae4c243ff1a22f7f28ff83694f2f7bc9">thinkingmachines/Inkling@828496eeae4c</a> · InklingForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">947.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">33.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 66</span><div class="cells"><i class="c0" title="layer 0: hybrid_sliding, MLP"></i><i class="c0" title="layer 1: hybrid_sliding, MLP"></i><i class="c1" title="layer 2: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 3: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 4: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 5: hybrid, Experts, MoE …"></i><i class="c1" title="layer 6: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 7: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 8: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 9: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 10: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 11: hybrid, Experts, MoE …"></i><i class="c1" title="layer 12: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 13: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 14: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 15: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 16: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 17: hybrid, Experts, MoE …"></i><i class="c1" title="layer 18: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 19: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 20: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 21: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 22: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 23: hybrid, Experts, MoE …"></i><i class="c1" title="layer 24: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 25: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 26: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 27: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 28: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 29: hybrid, Experts, MoE …"></i><i class="c1" title="layer 30: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 31: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 32: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 33: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 34: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 35: hybrid, Experts, MoE …"></i><i class="c1" title="layer 36: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 37: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 38: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 39: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 40: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 41: hybrid, Experts, MoE …"></i><i class="c1" title="layer 42: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 43: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 44: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 45: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 46: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 47: hybrid, Experts, MoE …"></i><i class="c1" title="layer 48: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 49: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 50: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 51: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 52: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 53: hybrid, Experts, MoE …"></i><i class="c1" title="layer 54: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 55: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 56: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 57: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 58: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 59: hybrid, Experts, MoE …"></i><i class="c1" title="layer 60: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 61: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 62: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 63: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 64: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 65: hybrid, Experts, MoE …"></i></div></div><div class="row"><span class="lab">Fixture · 6</span><div class="cells"><i class="c0" title="layer 0: hybrid_sliding, MLP"></i><i class="c0" title="layer 1: hybrid_sliding, MLP"></i><i class="c1" title="layer 2: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 3: hybrid_sliding, Experts, MoE …"></i><i class="c1" title="layer 4: hybrid_sliding, Experts, MoE …"></i><i class="c2" title="layer 5: hybrid, Experts, MoE …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>hybrid_sliding, MLP</td><td>2</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>hybrid_sliding, Experts, MoE …</td><td>53</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>hybrid, Experts, MoE …</td><td>11</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vision_tower.encoder_layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: RMSNorm"></i><i class="c0" title="layer 1: RMSNorm"></i><i class="c0" title="layer 2: RMSNorm"></i><i class="c1" title="layer 3: no RMSNorm"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: RMSNorm"></i><i class="c1" title="layer 1: no RMSNorm"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>RMSNorm</td><td>3</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>no RMSNorm</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.997295" title="model.language_model.layers: 944.6B (99.7%)"></i><i class="c1" style="flex:0.001304" title="model.language_model: 1.2B (0.1%)"></i><i class="c2" style="flex:0.001304" title="lm_head: 1.2B (0.1%)"></i><i class="c3" style="flex:0.000088" title="model.vision_tower.encoder_layers: 83.7M (0.0%)"></i><i class="c4" style="flex:0.000008" title="model.audio_tower: 7.9M (0.0%)"></i><i class="c5" style="flex:0.000000" title="model.vision_tower: 6.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.214181" title="model.language_model.layers: 7.2M (21.4%)"></i><i class="c1" style="flex:0.380580" title="model.language_model: 12.9M (38.1%)"></i><i class="c2" style="flex:0.380576" title="lm_head: 12.9M (38.1%)"></i><i class="c3" style="flex:0.022235" title="model.vision_tower.encoder_layers: 751.7K (2.2%)"></i><i class="c4" style="flex:0.002425" title="model.audio_tower: 82.0K (0.2%)"></i><i class="c5" style="flex:0.000002" title="model.vision_tower: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>944.6B</td><td>99.7%</td><td>7.2M</td><td>21.4%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>1.2B</td><td>0.1%</td><td>12.9M</td><td>38.1%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>1.2B</td><td>0.1%</td><td>12.9M</td><td>38.1%</td></tr><tr><td><i class="c3"></i>model.vision_tower.encoder_layers</td><td>83.7M</td><td>0.0%</td><td>751.7K</td><td>2.2%</td></tr><tr><td><i class="c4"></i>model.audio_tower</td><td>7.9M</td><td>0.0%</td><td>82.0K</td><td>0.2%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>6.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>inkling_mm_model</td><td>inkling_mm_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (64 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-6 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>inkling_text</td><td>inkling_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>201024</td><td>201024</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>1048576</td><td>1048576</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>inkling_audio</td><td>inkling_audio</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>inkling_vision</td><td>inkling_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>40</td><td>40</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Inkling omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>341</td><td>39</td></tr><tr class=""><td>RMSNorm</td><td>271</td><td>29</td></tr><tr class=""><td>Conv1d</td><td>264</td><td>24</td></tr><tr class=""><td>ShortConvolution</td><td>264</td><td>24</td></tr><tr class=""><td>SiLUActivation</td><td>130</td><td>10</td></tr><tr class=""><td>Attention</td><td>66</td><td>6</td></tr><tr class=""><td>DecoderLayer</td><td>66</td><td>6</td></tr><tr class=""><td>RelativeLogits</td><td>66</td><td>6</td></tr><tr class=""><td>Experts</td><td>64</td><td>4</td></tr><tr class=""><td>MoE</td><td>64</td><td>4</td></tr><tr class=""><td>SharedExperts</td><td>64</td><td>4</td></tr><tr class=""><td>TopkRouter</td><td>64</td><td>4</td></tr><tr class=""><td>VisionEncoderLayer</td><td>4</td><td>2</td></tr><tr class=""><td>MLP</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>AudioModelEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>NormedEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">16 fields changed, 102 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.text_hidden_size</td><td>6144</td><td>64</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>6144</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>24576</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>66 entries: 55× hybrid_sliding, 11× hybrid</td><td>[&quot;hybrid_sliding&quot;, &quot;hybrid_sliding&quot;, &quot;hybrid_sliding&quot;, &quot;hybrid_sliding&quot;, &quot;hyb…</td></tr><tr><td>text_config.mlp_layer_types</td><td>66 entries: 64× sparse, 2× dense</td><td>[&quot;dense&quot;, &quot;dense&quot;, &quot;sparse&quot;, &quot;sparse&quot;, &quot;sparse&quot;, &quot;sparse&quot;]</td></tr><tr><td>text_config.moe_intermediate_size</td><td>3072</td><td>32</td></tr><tr><td>text_config.n_routed_experts</td><td>256</td><td>4</td></tr><tr><td>text_config.num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>6</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>66</td><td>6</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.n_layers</td><td>4</td><td>2</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.text_hidden_size</td><td>6144</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Inkling-mic.json b/shrink/Inkling-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..23d234d2978b274b4321692515d57c868e20df1c
--- /dev/null
+++ b/shrink/Inkling-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Inkling-mic@a7c61810959524051418bc1a49a70ac559307c0b", "findings": []}
\ No newline at end of file
diff --git a/shrink/Jamba-tiny-dev-mic.html b/shrink/Jamba-tiny-dev-mic.html
index 76e4e350a283e5338fc278f6b8231f7b3d46e247..670e2000c17b7407dff05945fb194ee388999f3d 100644
--- a/shrink/Jamba-tiny-dev-mic.html
+++ b/shrink/Jamba-tiny-dev-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ai21labs/Jamba-tiny-dev: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Jamba-tiny-dev-mic@d03ef8cc4d1f81b8af2ae68030ecaa9bc1d51a6d ← ai21labs/Jamba-tiny-dev@ed303361004ac875426a61675edecf8e9d976882 · JambaForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ai21labs/Jamba-tiny-dev</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Jamba-tiny-dev-mic/tree/d03ef8cc4d1f81b8af2ae68030ecaa9bc1d51a6d" title="huggingface-mic/Jamba-tiny-dev-mic@d03ef8cc4d1f81b8af2ae68030ecaa9bc1d51a6d">huggingface-mic/Jamba-tiny-dev-mic@d03ef8cc4d1f</a> ← <a href="https://huggingface.co/ai21labs/Jamba-tiny-dev/tree/ed303361004ac875426a61675edecf8e9d976882" title="ai21labs/Jamba-tiny-dev@ed303361004ac875426a61675edecf8e9d976882">ai21labs/Jamba-tiny-dev@ed303361004a</a> · JambaForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">318.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 1: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c0" title="layer 2: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 3: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c2" title="layer 4: is_causal=True, Attention, AttentionDecoderLayer …"></i><i class="c1" title="layer 5: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c0" title="layer 6: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 7: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c0" title="layer 8: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 9: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c0" title="layer 10: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 11: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c2" title="layer 12: is_causal=True, Attention, AttentionDecoderLayer …"></i><i class="c1" title="layer 13: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c0" title="layer 14: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 15: Conv1d, Experts, MambaDecoderLayer …"></i></div></div><div class="row"><span class="lab">Fixture · 5</span><div class="cells"><i class="c0" title="layer 0: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 1: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c0" title="layer 2: Conv1d, MLP, MambaDecoderLayer …"></i><i class="c1" title="layer 3: Conv1d, Experts, MambaDecoderLayer …"></i><i class="c2" title="layer 4: is_causal=True, Attention, AttentionDecoderLayer …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv1d, MLP, MambaDecoderLayer …</td><td>6</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Conv1d, Experts, MambaDecoderLayer …</td><td>8</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>is_causal=True, Attention, AttentionDecoderLayer …</td><td>2</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.789420" title="model.layers: 251.6M (78.9%)"></i><i class="c1" style="flex:0.105289" title="lm_head: 33.6M (10.5%)"></i><i class="c2" style="flex:0.105289" title="model.embed_tokens: 33.6M (10.5%)"></i><i class="c3" style="flex:0.000002" title="model.final_layernorm: 512 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.050189" title="model.layers: 443.3K (5.0%)"></i><i class="c1" style="flex:0.474902" title="lm_head: 4.2M (47.5%)"></i><i class="c2" style="flex:0.474902" title="model.embed_tokens: 4.2M (47.5%)"></i><i class="c3" style="flex:0.000007" title="model.final_layernorm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>251.6M</td><td>78.9%</td><td>443.3K</td><td>5.0%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>33.6M</td><td>10.5%</td><td>4.2M</td><td>47.5%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>33.6M</td><td>10.5%</td><td>4.2M</td><td>47.5%</td></tr><tr><td><i class="c3"></i>model.final_layernorm</td><td>512</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 8 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>jamba</td><td>jamba</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>65536</td><td>65536</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Jamba omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>89</td><td>30</td></tr><tr class=""><td>RMSNorm</td><td>75</td><td>23</td></tr><tr class=""><td>SiLUActivation</td><td>30</td><td>9</td></tr><tr class=""><td>Conv1d</td><td>14</td><td>4</td></tr><tr class=""><td>MambaDecoderLayer</td><td>14</td><td>4</td></tr><tr class=""><td>MambaMixer</td><td>14</td><td>4</td></tr><tr class=""><td>Experts</td><td>8</td><td>2</td></tr><tr class=""><td>MLP</td><td>8</td><td>3</td></tr><tr class=""><td>SparseMoeBlock</td><td>8</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>8</td><td>2</td></tr><tr class=""><td>Attention</td><td>2</td><td>1</td></tr><tr class=""><td>AttentionDecoderLayer</td><td>2</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 39 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>512</td><td>64</td></tr><tr><td>intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_experts</td><td>8</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>5</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Jamba-tiny-dev-mic.json b/shrink/Jamba-tiny-dev-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..546197a20cf0eaa2ff38ab53eb1f94545d62eb0d
--- /dev/null
+++ b/shrink/Jamba-tiny-dev-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Jamba-tiny-dev-mic@d03ef8cc4d1f81b8af2ae68030ecaa9bc1d51a6d", "findings": []}
\ No newline at end of file
diff --git a/shrink/Janus-Pro-1B-mic.html b/shrink/Janus-Pro-1B-mic.html
index fdb32f404e33a7823f04dd1acc41cce989693d96..b038d93f7dd7b6d131c06a1c70e7d7058aca71a0 100644
--- a/shrink/Janus-Pro-1B-mic.html
+++ b/shrink/Janus-Pro-1B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>deepseek-community/Janus-Pro-1B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Janus-Pro-1B-mic@5d0bfc05499ea83be331a434ea79583ca970581a ← deepseek-community/Janus-Pro-1B@1655280bb75959cc1cb85529a2a8b26e7016072e · JanusForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>deepseek-community/Janus-Pro-1B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Janus-Pro-1B-mic/tree/5d0bfc05499ea83be331a434ea79583ca970581a" title="huggingface-mic/Janus-Pro-1B-mic@5d0bfc05499ea83be331a434ea79583ca970581a">huggingface-mic/Janus-Pro-1B-mic@5d0bfc05499e</a> ← <a href="https://huggingface.co/deepseek-community/Janus-Pro-1B/tree/1655280bb75959cc1cb85529a2a8b26e7016072e" title="deepseek-community/Janus-Pro-1B@1655280bb75959cc1cb85529a2a8b26e7016072e">deepseek-community/Janus-Pro-1B@1655280bb759</a> · JanusForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">81.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">4</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i><i class="c0" title="layer 8: VisionEncoderLayer"></i><i class="c0" title="layer 9: VisionEncoderLayer"></i><i class="c0" title="layer 10: VisionEncoderLayer"></i><i class="c0" title="layer 11: VisionEncoderLayer"></i><i class="c0" title="layer 12: VisionEncoderLayer"></i><i class="c0" title="layer 13: VisionEncoderLayer"></i><i class="c0" title="layer 14: VisionEncoderLayer"></i><i class="c0" title="layer 15: VisionEncoderLayer"></i><i class="c0" title="layer 16: VisionEncoderLayer"></i><i class="c0" title="layer 17: VisionEncoderLayer"></i><i class="c0" title="layer 18: VisionEncoderLayer"></i><i class="c0" title="layer 19: VisionEncoderLayer"></i><i class="c0" title="layer 20: VisionEncoderLayer"></i><i class="c0" title="layer 21: VisionEncoderLayer"></i><i class="c0" title="layer 22: VisionEncoderLayer"></i><i class="c0" title="layer 23: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vqmodel.encoder.down</h3><div class="row"><span class="lab">Original · 5</span><div class="cells"><i class="c0" title="layer 0: VQVAEConvDownsample"></i><i class="c0" title="layer 1: VQVAEConvDownsample"></i><i class="c0" title="layer 2: VQVAEConvDownsample"></i><i class="c0" title="layer 3: VQVAEConvDownsample"></i><i class="c1" title="layer 4: VQVAEAttnBlock"></i></div></div><div class="row"><span class="lab">Fixture · 5</span><div class="cells"><i class="c0" title="layer 0: VQVAEConvDownsample"></i><i class="c0" title="layer 1: VQVAEConvDownsample"></i><i class="c0" title="layer 2: VQVAEConvDownsample"></i><i class="c0" title="layer 3: VQVAEConvDownsample"></i><i class="c1" title="layer 4: VQVAEAttnBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VQVAEConvDownsample</td><td>4</td><td>4</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>VQVAEAttnBlock</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vqmodel.decoder.up</h3><div class="row"><span class="lab">Original · 5</span><div class="cells"><i class="c0" title="layer 0: VQVAEAttnBlock, VQVAEConvUpsample"></i><i class="c1" title="layer 1: VQVAEConvUpsample"></i><i class="c1" title="layer 2: VQVAEConvUpsample"></i><i class="c1" title="layer 3: VQVAEConvUpsample"></i><i class="c2" title="layer 4: no VQVAEAttnBlock"></i></div></div><div class="row"><span class="lab">Fixture · 5</span><div class="cells"><i class="c0" title="layer 0: VQVAEAttnBlock, VQVAEConvUpsample"></i><i class="c1" title="layer 1: VQVAEConvUpsample"></i><i class="c1" title="layer 2: VQVAEConvUpsample"></i><i class="c1" title="layer 3: VQVAEConvUpsample"></i><i class="c2" title="layer 4: no VQVAEAttnBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VQVAEAttnBlock, VQVAEConvUpsample</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>VQVAEConvUpsample</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>no VQVAEAttnBlock</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i><i class="c0" title="layer 2: LlamaDecoderLayer"></i><i class="c0" title="layer 3: LlamaDecoderLayer"></i><i class="c0" title="layer 4: LlamaDecoderLayer"></i><i class="c0" title="layer 5: LlamaDecoderLayer"></i><i class="c0" title="layer 6: LlamaDecoderLayer"></i><i class="c0" title="layer 7: LlamaDecoderLayer"></i><i class="c0" title="layer 8: LlamaDecoderLayer"></i><i class="c0" title="layer 9: LlamaDecoderLayer"></i><i class="c0" title="layer 10: LlamaDecoderLayer"></i><i class="c0" title="layer 11: LlamaDecoderLayer"></i><i class="c0" title="layer 12: LlamaDecoderLayer"></i><i class="c0" title="layer 13: LlamaDecoderLayer"></i><i class="c0" title="layer 14: LlamaDecoderLayer"></i><i class="c0" title="layer 15: LlamaDecoderLayer"></i><i class="c0" title="layer 16: LlamaDecoderLayer"></i><i class="c0" title="layer 17: LlamaDecoderLayer"></i><i class="c0" title="layer 18: LlamaDecoderLayer"></i><i class="c0" title="layer 19: LlamaDecoderLayer"></i><i class="c0" title="layer 20: LlamaDecoderLayer"></i><i class="c0" title="layer 21: LlamaDecoderLayer"></i><i class="c0" title="layer 22: LlamaDecoderLayer"></i><i class="c0" title="layer 23: LlamaDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>LlamaDecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.660565" title="model.language_model.layers: 1.2B (66.1%)"></i><i class="c1" style="flex:0.161929" title="model.vision_model.encoder.layers: 302.3M (16.2%)"></i><i class="c2" style="flex:0.112333" title="model.language_model: 209.7M (11.2%)"></i><i class="c3" style="flex:0.020230" title="model.generation_head: 37.8M (2.0%)"></i><i class="c4" style="flex:0.016482" title="model.vqmodel.decoder.up: 30.8M (1.6%)"></i><i class="c5" style="flex:0.012584" title="model.vqmodel: 23.5M (1.3%)"></i><i class="c6" style="flex:0.009437" title="model.vqmodel.encoder.down: 17.6M (0.9%)"></i><i class="c7" style="flex:0.006439" title="4 others: 12.0M (0.6%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.001003" title="model.language_model.layers: 82.2K (0.1%)"></i><i class="c1" style="flex:0.001220" title="model.vision_model.encoder.layers: 100.0K (0.1%)"></i><i class="c2" style="flex:0.080008" title="model.language_model: 6.6M (8.0%)"></i><i class="c3" style="flex:0.014602" title="model.generation_head: 1.2M (1.5%)"></i><i class="c4" style="flex:0.375654" title="model.vqmodel.decoder.up: 30.8M (37.6%)"></i><i class="c5" style="flex:0.298366" title="model.vqmodel: 24.4M (29.8%)"></i><i class="c6" style="flex:0.215090" title="model.vqmodel.encoder.down: 17.6M (21.5%)"></i><i class="c7" style="flex:0.014057" title="4 others: 1.2M (1.4%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>1.2B</td><td>66.1%</td><td>82.2K</td><td>0.1%</td></tr><tr><td><i class="c1"></i>model.vision_model.encoder.layers</td><td>302.3M</td><td>16.2%</td><td>100.0K</td><td>0.1%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>209.7M</td><td>11.2%</td><td>6.6M</td><td>8.0%</td></tr><tr><td><i class="c3"></i>model.generation_head</td><td>37.8M</td><td>2.0%</td><td>1.2M</td><td>1.5%</td></tr><tr><td><i class="c4"></i>model.vqmodel.decoder.up</td><td>30.8M</td><td>1.6%</td><td>30.8M</td><td>37.6%</td></tr><tr><td><i class="c5"></i>model.vqmodel</td><td>23.5M</td><td>1.3%</td><td>24.4M</td><td>29.8%</td></tr><tr><td><i class="c6"></i>model.vqmodel.encoder.down</td><td>17.6M</td><td>0.9%</td><td>17.6M</td><td>21.5%</td></tr><tr><td><i class="c7"></i>4 others</td><td>12.0M</td><td>0.6%</td><td>1.2M</td><td>1.4%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>janus</td><td>janus</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>llama</td><td>llama</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>102400</td><td>102400</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>16384</td><td>16384</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>janus_vision_model</td><td>janus_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>384</td><td>384</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>vq_config</td><td>model_type</td><td>janus_vqgan</td><td>janus_vqgan</td><td>✓ kept</td></tr><tr class=""><td>vq_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Janus omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>319</td><td>33</td></tr><tr class=""><td>Conv2d</td><td>105</td><td>105</td></tr><tr class=""><td>Dropout</td><td>77</td><td>33</td></tr><tr class=""><td>Identity</td><td>72</td><td>6</td></tr><tr class=""><td>GroupNorm</td><td>67</td><td>67</td></tr><tr class=""><td>LayerNorm</td><td>49</td><td>5</td></tr><tr class=""><td>LlamaRMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>VQVAEResnetBlock</td><td>29</td><td>29</td></tr><tr class=""><td>GELUActivation</td><td>27</td><td>5</td></tr><tr class=""><td>ModuleList</td><td>26</td><td>26</td></tr><tr class=""><td>VisionAttention</td><td>24</td><td>2</td></tr><tr class=""><td>VisionEncoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>24</td><td>2</td></tr><tr class=""><td>LlamaAttention</td><td>24</td><td>2</td></tr><tr class=""><td>LlamaDecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>LlamaMLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Module</td><td>10</td><td>10</td></tr><tr class=""><td>VQVAEAttnBlock</td><td>7</td><td>7</td></tr><tr class=""><td>Embedding</td><td>4</td><td>4</td></tr><tr class=""><td>VQVAEConvDownsample</td><td>4</td><td>4</td></tr><tr class=""><td>VQVAEConvUpsample</td><td>4</td><td>4</td></tr><tr class=""><td>VQVAEMidBlock</td><td>2</td><td>2</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>VQVAE</td><td>1</td><td>1</td></tr><tr class=""><td>VQVAEAlignerMLP</td><td>1</td><td>1</td></tr><tr class=""><td>VQVAEDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>VQVAEEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VQVAEHead</td><td>1</td><td>1</td></tr><tr class=""><td>VQVAEVectorQuantizer</td><td>1</td><td>1</td></tr><tr class=""><td>VisionAlignerMLP</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaModel</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 82 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>5632</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>2048</td><td>64</td></tr><tr><td>vq_config.embed_dim</td><td>8</td><td>64</td></tr><tr><td>vq_config.projection_dim</td><td>2048</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Janus-Pro-1B-mic.json b/shrink/Janus-Pro-1B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..58ffcf8cb3fa919dbea57479696172f6485392af
--- /dev/null
+++ b/shrink/Janus-Pro-1B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Janus-Pro-1B-mic@5d0bfc05499ea83be331a434ea79583ca970581a", "findings": []}
\ No newline at end of file
diff --git a/shrink/LCO-Embedding-Omni-3B-mic.html b/shrink/LCO-Embedding-Omni-3B-mic.html
index 2fbe0f857e38f829102a4bda829c8fea63d372a4..117d468f44428bdc7836d4c433ecd2944c6db567 100644
--- a/shrink/LCO-Embedding-Omni-3B-mic.html
+++ b/shrink/LCO-Embedding-Omni-3B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>LCO-Embedding/LCO-Embedding-Omni-3B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/LCO-Embedding-Omni-3B-mic@1e770f7015e573f2c4287893af7c6eb0c1d45c33 ← LCO-Embedding/LCO-Embedding-Omni-3B@73434565c33d032551639362f7fc91fa3c6f065f · Qwen2_5OmniThinkerForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>LCO-Embedding/LCO-Embedding-Omni-3B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/LCO-Embedding-Omni-3B-mic/tree/1e770f7015e573f2c4287893af7c6eb0c1d45c33" title="huggingface-mic/LCO-Embedding-Omni-3B-mic@1e770f7015e573f2c4287893af7c6eb0c1d45c33">huggingface-mic/LCO-Embedding-Omni-3B-mic@1e770f7015e5</a> ← <a href="https://huggingface.co/LCO-Embedding/LCO-Embedding-Omni-3B/tree/73434565c33d032551639362f7fc91fa3c6f065f" title="LCO-Embedding/LCO-Embedding-Omni-3B@73434565c33d032551639362f7fc91fa3c6f065f">LCO-Embedding/LCO-Embedding-Omni-3B@73434565c33d</a> · Qwen2_5OmniThinkerForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">21.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.audio_tower.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 1: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 2: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 3: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 4: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 5: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 6: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 7: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 8: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 9: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 10: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 11: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 12: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 13: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 14: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 15: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 16: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 17: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 18: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 19: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 20: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 21: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 22: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 23: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 24: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 25: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 26: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 27: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 28: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 29: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 30: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 31: Qwen2_5OmniAudioEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2_5OmniAudioEncoderLayer"></i><i class="c0" title="layer 1: Qwen2_5OmniAudioEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2_5OmniAudioEncoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 1: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 2: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 3: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 4: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 5: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 6: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 7: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 8: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 9: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 10: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 11: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 12: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 13: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 14: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 15: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 16: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 17: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 18: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 19: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 20: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 21: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 22: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 23: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 24: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 25: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 26: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 27: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 28: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 29: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 30: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 31: Qwen2_5OmniVisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 32</span><div class="cells"><i class="c0" title="layer 0: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 1: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 2: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 3: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 4: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 5: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 6: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 7: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 8: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 9: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 10: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 11: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 12: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 13: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 14: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 15: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 16: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 17: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 18: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 19: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 20: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 21: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 22: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 23: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 24: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 25: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 26: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 27: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 28: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 29: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 30: Qwen2_5OmniVisionBlock"></i><i class="c0" title="layer 31: Qwen2_5OmniVisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2_5OmniVisionBlock</td><td>32</td><td>32</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 36</span><div class="cells"><i class="c0" title="layer 0: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 1: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 2: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 3: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 4: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 5: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 6: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 7: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 8: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 9: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 10: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 11: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 12: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 13: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 14: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 15: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 16: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 17: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 18: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 19: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 20: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 21: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 22: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 23: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 24: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 25: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 26: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 27: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 28: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 29: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 30: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 31: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 32: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 33: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 34: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 35: Qwen2_5OmniDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2_5OmniDecoderLayer"></i><i class="c0" title="layer 1: Qwen2_5OmniDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2_5OmniDecoderLayer</td><td>36</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.589942" title="model.language_model.layers: 2.8B (59.0%)"></i><i class="c1" style="flex:0.134044" title="model.visual.blocks: 630.5M (13.4%)"></i><i class="c2" style="flex:0.133867" title="model.audio_tower.layers: 629.6M (13.4%)"></i><i class="c3" style="flex:0.066157" title="model.language_model: 311.2M (6.6%)"></i><i class="c4" style="flex:0.066157" title="lm_head: 311.2M (6.6%)"></i><i class="c5" style="flex:0.008125" title="model.visual: 38.2M (0.8%)"></i><i class="c6" style="flex:0.001709" title="model.audio_tower: 8.0M (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003515" title="model.language_model.layers: 74.2K (0.4%)"></i><i class="c1" style="flex:0.063124" title="model.visual.blocks: 1.3M (6.3%)"></i><i class="c2" style="flex:0.003163" title="model.audio_tower.layers: 66.8K (0.3%)"></i><i class="c3" style="flex:0.460390" title="model.language_model: 9.7M (46.0%)"></i><i class="c4" style="flex:0.460387" title="lm_head: 9.7M (46.0%)"></i><i class="c5" style="flex:0.007460" title="model.visual: 157.6K (0.7%)"></i><i class="c6" style="flex:0.001960" title="model.audio_tower: 41.4K (0.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>2.8B</td><td>59.0%</td><td>74.2K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.visual.blocks</td><td>630.5M</td><td>13.4%</td><td>1.3M</td><td>6.3%</td></tr><tr><td><i class="c2"></i>model.audio_tower.layers</td><td>629.6M</td><td>13.4%</td><td>66.8K</td><td>0.3%</td></tr><tr><td><i class="c3"></i>model.language_model</td><td>311.2M</td><td>6.6%</td><td>9.7M</td><td>46.0%</td></tr><tr><td><i class="c4"></i>lm_head</td><td>311.2M</td><td>6.6%</td><td>9.7M</td><td>46.0%</td></tr><tr><td><i class="c5"></i>model.visual</td><td>38.2M</td><td>0.8%</td><td>157.6K</td><td>0.7%</td></tr><tr><td><i class="c6"></i>model.audio_tower</td><td>8.0M</td><td>0.2%</td><td>41.4K</td><td>0.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen2_5_omni_thinker</td><td>qwen2_5_omni_thinker</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>qwen2_5_omni_audio_encoder</td><td>qwen2_5_omni_audio_encoder</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>128</td><td>128</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen2_5_omni_vision_encoder</td><td>qwen2_5_omni_vision_encoder</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2_5_omni_text</td><td>qwen2_5_omni_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen2_5OmniThinker omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>672</td><td>254</td></tr><tr class=""><td>Qwen2_5OmniRMSNorm</td><td>138</td><td>70</td></tr><tr class=""><td>SiLUActivation</td><td>68</td><td>34</td></tr><tr class=""><td>LayerNorm</td><td>65</td><td>5</td></tr><tr class=""><td>Qwen2MLP</td><td>36</td><td>2</td></tr><tr class=""><td>Qwen2_5OmniAttention</td><td>36</td><td>2</td></tr><tr class=""><td>Qwen2_5OmniDecoderLayer</td><td>36</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Qwen2_5OmniAudioAttention</td><td>32</td><td>2</td></tr><tr class=""><td>Qwen2_5OmniAudioEncoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>Qwen2_5OmniMLP</td><td>32</td><td>32</td></tr><tr class=""><td>Qwen2_5OmniVisionAttention</td><td>32</td><td>32</td></tr><tr class=""><td>Qwen2_5OmniVisionBlock</td><td>32</td><td>32</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Conv1d</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>AvgPool1d</td><td>1</td><td>1</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2_5OmniAudioEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2_5OmniPatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2_5OmniRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2_5OmniVisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2_5OmniVisionRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2_5_VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>Sequential</td><td>1</td><td>1</td></tr><tr class=""><td>SinusoidsPositionEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">16 fields changed, 102 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.d_model</td><td>1280</td><td>64</td></tr><tr><td>audio_config.encoder_attention_heads</td><td>20</td><td>4</td></tr><tr><td>audio_config.encoder_ffn_dim</td><td>5120</td><td>128</td></tr><tr><td>audio_config.encoder_layers</td><td>32</td><td>2</td></tr><tr><td>audio_config.output_dim</td><td>2048</td><td>64</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>11008</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>36 entries: 36× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>36</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_section&quot;: [16, 24, 24], &quot;rope_type&quot;: &quot;default&quot;, &quot;type&quot;: &quot;default&quot;, &quot;r…</td><td>{&quot;mrope_section&quot;: [2, 3, 3], &quot;rope_theta&quot;: 1000000.0, &quot;rope_type&quot;: &quot;default&quot;,…</td></tr><tr><td>vision_config.embed_dim</td><td>1280</td><td>64</td></tr><tr><td>vision_config.hidden_size</td><td>1280</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3420</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>2048</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/LCO-Embedding-Omni-3B-mic.json b/shrink/LCO-Embedding-Omni-3B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1b9d3203cd79f0a2bdd6f62411ab7e440ff021e7
--- /dev/null
+++ b/shrink/LCO-Embedding-Omni-3B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/LCO-Embedding-Omni-3B-mic@1e770f7015e573f2c4287893af7c6eb0c1d45c33", "findings": []}
\ No newline at end of file
diff --git a/shrink/LFM2-8B-A1B-mic.html b/shrink/LFM2-8B-A1B-mic.html
index 2374bbb352d035c84326b8a69b3cc008eea8f545..4a04e0821f979bd12e4a76d8c2c6032b08bc5bb6 100644
--- a/shrink/LFM2-8B-A1B-mic.html
+++ b/shrink/LFM2-8B-A1B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>LiquidAI/LFM2-8B-A1B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/LFM2-8B-A1B-mic@02493cd540778260b98790f375fb2070e8acd824 ← LiquidAI/LFM2-8B-A1B@c1c44ff9fc00db3ebf4516970563f5f383d23670 · Lfm2MoeForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>LiquidAI/LFM2-8B-A1B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/LFM2-8B-A1B-mic/tree/02493cd540778260b98790f375fb2070e8acd824" title="huggingface-mic/LFM2-8B-A1B-mic@02493cd540778260b98790f375fb2070e8acd824">huggingface-mic/LFM2-8B-A1B-mic@02493cd54077</a> ← <a href="https://huggingface.co/LiquidAI/LFM2-8B-A1B/tree/c1c44ff9fc00db3ebf4516970563f5f383d23670" title="LiquidAI/LFM2-8B-A1B@c1c44ff9fc00db3ebf4516970563f5f383d23670">LiquidAI/LFM2-8B-A1B@c1c44ff9fc00</a> · Lfm2MoeForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">8.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: conv, Conv1d, MLP …"></i><i class="c0" title="layer 1: conv, Conv1d, MLP …"></i><i class="c1" title="layer 2: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 3: conv, Conv1d, Experts …"></i><i class="c2" title="layer 4: conv, Conv1d, Experts …"></i><i class="c2" title="layer 5: conv, Conv1d, Experts …"></i><i class="c1" title="layer 6: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 7: conv, Conv1d, Experts …"></i><i class="c2" title="layer 8: conv, Conv1d, Experts …"></i><i class="c2" title="layer 9: conv, Conv1d, Experts …"></i><i class="c1" title="layer 10: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 11: conv, Conv1d, Experts …"></i><i class="c2" title="layer 12: conv, Conv1d, Experts …"></i><i class="c2" title="layer 13: conv, Conv1d, Experts …"></i><i class="c1" title="layer 14: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 15: conv, Conv1d, Experts …"></i><i class="c2" title="layer 16: conv, Conv1d, Experts …"></i><i class="c2" title="layer 17: conv, Conv1d, Experts …"></i><i class="c1" title="layer 18: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 19: conv, Conv1d, Experts …"></i><i class="c2" title="layer 20: conv, Conv1d, Experts …"></i><i class="c1" title="layer 21: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 22: conv, Conv1d, Experts …"></i><i class="c2" title="layer 23: conv, Conv1d, Experts …"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: conv, Conv1d, MLP …"></i><i class="c0" title="layer 1: conv, Conv1d, MLP …"></i><i class="c1" title="layer 2: is_causal=True, Attention, Experts …"></i><i class="c2" title="layer 3: conv, Conv1d, Experts …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>conv, Conv1d, MLP …</td><td>2</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention, Experts …</td><td>6</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>conv, Conv1d, Experts …</td><td>16</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.983906" title="model.layers: 8.2B (98.4%)"></i><i class="c1" style="flex:0.016093" title="model.embed_tokens: 134.2M (1.6%)"></i><i class="c2" style="flex:0.000000" title="model.embedding_norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.037049" title="model.layers: 161.4K (3.7%)"></i><i class="c1" style="flex:0.962936" title="model.embed_tokens: 4.2M (96.3%)"></i><i class="c2" style="flex:0.000015" title="model.embedding_norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>8.2B</td><td>98.4%</td><td>161.4K</td><td>3.7%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>134.2M</td><td>1.6%</td><td>4.2M</td><td>96.3%</td></tr><tr><td><i class="c2"></i>model.embedding_norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-4 of 32 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>lfm2_moe</td><td>lfm2_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>65536</td><td>65536</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>128000</td><td>128000</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Lfm2Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>67</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>61</td><td>11</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>4</td></tr><tr class=""><td>Experts</td><td>22</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>22</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>22</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>18</td><td>3</td></tr><tr class=""><td>ShortConv</td><td>18</td><td>3</td></tr><tr class=""><td>Attention</td><td>6</td><td>1</td></tr><tr class=""><td>MLP</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>7168</td><td>128</td></tr><tr><td>layer_types</td><td>24 entries: 18× conv, 6× full_attention</td><td>[&quot;conv&quot;, &quot;conv&quot;, &quot;full_attention&quot;, &quot;conv&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>1792</td><td>32</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_experts</td><td>32</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>4</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>4</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/LFM2-8B-A1B-mic.json b/shrink/LFM2-8B-A1B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1f5bf2b8434dcf0cc5fbf67f3e66bfe9634e0368
--- /dev/null
+++ b/shrink/LFM2-8B-A1B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/LFM2-8B-A1B-mic@02493cd540778260b98790f375fb2070e8acd824", "findings": []}
\ No newline at end of file
diff --git a/shrink/LFM2.5-230M-Base-mic.html b/shrink/LFM2.5-230M-Base-mic.html
index ff18db627fa25b7f9da28d9edaaf1c96f40c5bb4..de851ef6714f7de6db1da9b2f5ea6d0971277156 100644
--- a/shrink/LFM2.5-230M-Base-mic.html
+++ b/shrink/LFM2.5-230M-Base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>LiquidAI/LFM2.5-230M-Base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/LFM2.5-230M-Base-mic@d6697c673a52ed87ae4e8ed607e5b6a1ea6f7a8d ← LiquidAI/LFM2.5-230M-Base@9d2be5519834990d30996f878b6771cccbd24f2c · Lfm2ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>LiquidAI/LFM2.5-230M-Base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/LFM2.5-230M-Base-mic/tree/d6697c673a52ed87ae4e8ed607e5b6a1ea6f7a8d" title="huggingface-mic/LFM2.5-230M-Base-mic@d6697c673a52ed87ae4e8ed607e5b6a1ea6f7a8d">huggingface-mic/LFM2.5-230M-Base-mic@d6697c673a52</a> ← <a href="https://huggingface.co/LiquidAI/LFM2.5-230M-Base/tree/9d2be5519834990d30996f878b6771cccbd24f2c" title="LiquidAI/LFM2.5-230M-Base@9d2be5519834990d30996f878b6771cccbd24f2c">LiquidAI/LFM2.5-230M-Base@9d2be5519834</a> · Lfm2ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">229.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 14</span><div class="cells"><i class="c0" title="layer 0: conv, Conv1d, ShortConv"></i><i class="c0" title="layer 1: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 2: is_causal=True, Attention"></i><i class="c0" title="layer 3: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 4: is_causal=True, Attention"></i><i class="c0" title="layer 5: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 6: is_causal=True, Attention"></i><i class="c0" title="layer 7: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 8: is_causal=True, Attention"></i><i class="c0" title="layer 9: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 10: is_causal=True, Attention"></i><i class="c0" title="layer 11: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 12: is_causal=True, Attention"></i><i class="c0" title="layer 13: conv, Conv1d, ShortConv"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: conv, Conv1d, ShortConv"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>conv, Conv1d, ShortConv</td><td>8</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>6</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.707828" title="model.layers: 162.6M (70.8%)"></i><i class="c1" style="flex:0.292167" title="model.embed_tokens: 67.1M (29.2%)"></i><i class="c2" style="flex:0.000004" title="model.embedding_norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.018327" title="model.layers: 78.3K (1.8%)"></i><i class="c1" style="flex:0.981658" title="model.embed_tokens: 4.2M (98.2%)"></i><i class="c2" style="flex:0.000015" title="model.embedding_norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>162.6M</td><td>70.8%</td><td>78.3K</td><td>1.8%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>67.1M</td><td>29.2%</td><td>4.2M</td><td>98.2%</td></tr><tr><td><i class="c2"></i>model.embedding_norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>lfm2</td><td>lfm2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>65536</td><td>65536</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>128000</td><td>128000</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Lfm2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>83</td><td>13</td></tr><tr class=""><td>RMSNorm</td><td>41</td><td>7</td></tr><tr class=""><td>DecoderLayer</td><td>14</td><td>2</td></tr><tr class=""><td>MLP</td><td>14</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>8</td><td>1</td></tr><tr class=""><td>ShortConv</td><td>8</td><td>1</td></tr><tr class=""><td>Attention</td><td>6</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 41 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>2560</td><td>128</td></tr><tr><td>layer_types</td><td>14 entries: 8× conv, 6× full_attention</td><td>[&quot;conv&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>14</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/LFM2.5-230M-Base-mic.json b/shrink/LFM2.5-230M-Base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..f42c8fb2aed89d85eb9938cf4aa6dcbc2637f951
--- /dev/null
+++ b/shrink/LFM2.5-230M-Base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/LFM2.5-230M-Base-mic@d6697c673a52ed87ae4e8ed607e5b6a1ea6f7a8d", "findings": []}
\ No newline at end of file
diff --git a/shrink/LFM2.5-VL-450M-mic.html b/shrink/LFM2.5-VL-450M-mic.html
index 2391e81249a633171a7f2fb31f3fed94cece319a..9200d9b38c1afe90b9b5f31c246c7b216fde7ea5 100644
--- a/shrink/LFM2.5-VL-450M-mic.html
+++ b/shrink/LFM2.5-VL-450M-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>LiquidAI/LFM2.5-VL-450M: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/LFM2.5-VL-450M-mic@27b8f6b75e00ddd9758961d5abac08622f2b6d73 ← LiquidAI/LFM2.5-VL-450M@fc6221ca597f3315e4f82fc2df606783267b34ba · Lfm2VlForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>LiquidAI/LFM2.5-VL-450M</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/LFM2.5-VL-450M-mic/tree/27b8f6b75e00ddd9758961d5abac08622f2b6d73" title="huggingface-mic/LFM2.5-VL-450M-mic@27b8f6b75e00ddd9758961d5abac08622f2b6d73">huggingface-mic/LFM2.5-VL-450M-mic@27b8f6b75e00</a> ← <a href="https://huggingface.co/LiquidAI/LFM2.5-VL-450M/tree/fc6221ca597f3315e4f82fc2df606783267b34ba" title="LiquidAI/LFM2.5-VL-450M@fc6221ca597f3315e4f82fc2df606783267b34ba">LiquidAI/LFM2.5-VL-450M@fc6221ca597f</a> · Lfm2VlForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">448.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">5.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Siglip2EncoderLayer"></i><i class="c0" title="layer 1: Siglip2EncoderLayer"></i><i class="c0" title="layer 2: Siglip2EncoderLayer"></i><i class="c0" title="layer 3: Siglip2EncoderLayer"></i><i class="c0" title="layer 4: Siglip2EncoderLayer"></i><i class="c0" title="layer 5: Siglip2EncoderLayer"></i><i class="c0" title="layer 6: Siglip2EncoderLayer"></i><i class="c0" title="layer 7: Siglip2EncoderLayer"></i><i class="c0" title="layer 8: Siglip2EncoderLayer"></i><i class="c0" title="layer 9: Siglip2EncoderLayer"></i><i class="c0" title="layer 10: Siglip2EncoderLayer"></i><i class="c0" title="layer 11: Siglip2EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Siglip2EncoderLayer"></i><i class="c0" title="layer 1: Siglip2EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Siglip2EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: conv, Conv1d, Lfm2ShortConv"></i><i class="c0" title="layer 1: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 2: is_causal=True, Lfm2Attention"></i><i class="c0" title="layer 3: conv, Conv1d, Lfm2ShortConv"></i><i class="c0" title="layer 4: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 5: is_causal=True, Lfm2Attention"></i><i class="c0" title="layer 6: conv, Conv1d, Lfm2ShortConv"></i><i class="c0" title="layer 7: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 8: is_causal=True, Lfm2Attention"></i><i class="c0" title="layer 9: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 10: is_causal=True, Lfm2Attention"></i><i class="c0" title="layer 11: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 12: is_causal=True, Lfm2Attention"></i><i class="c0" title="layer 13: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 14: is_causal=True, Lfm2Attention"></i><i class="c0" title="layer 15: conv, Conv1d, Lfm2ShortConv"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: conv, Conv1d, Lfm2ShortConv"></i><i class="c1" title="layer 1: is_causal=True, Lfm2Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>conv, Conv1d, Lfm2ShortConv</td><td>10</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Lfm2Attention</td><td>6</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.640432" title="model.language_model.layers: 287.4M (64.0%)"></i><i class="c1" style="flex:0.189550" title="model.vision_tower.encoder.layers: 85.1M (19.0%)"></i><i class="c2" style="flex:0.149559" title="model.language_model: 67.1M (15.0%)"></i><i class="c3" style="flex:0.018701" title="model.multi_modal_projector: 8.4M (1.9%)"></i><i class="c4" style="flex:0.001758" title="model.vision_tower: 788.7K (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024933" title="model.language_model.layers: 127.5K (2.5%)"></i><i class="c1" style="flex:0.013096" title="model.vision_tower.encoder.layers: 66.9K (1.3%)"></i><i class="c2" style="flex:0.820500" title="model.language_model: 4.2M (82.0%)"></i><i class="c3" style="flex:0.128614" title="model.multi_modal_projector: 657.5K (12.9%)"></i><i class="c4" style="flex:0.012858" title="model.vision_tower: 65.7K (1.3%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>287.4M</td><td>64.0%</td><td>127.5K</td><td>2.5%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layers</td><td>85.1M</td><td>19.0%</td><td>66.9K</td><td>1.3%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>67.1M</td><td>15.0%</td><td>4.2M</td><td>82.0%</td></tr><tr><td><i class="c3"></i>model.multi_modal_projector</td><td>8.4M</td><td>1.9%</td><td>657.5K</td><td>12.9%</td></tr><tr><td><i class="c4"></i>model.vision_tower</td><td>788.7K</td><td>0.2%</td><td>65.7K</td><td>1.3%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>lfm2_vl</td><td>lfm2_vl</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>siglip2_vision_model</td><td>siglip2_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>lfm2</td><td>lfm2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>65536</td><td>65536</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>128000</td><td>128000</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Lfm2Vl omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>168</td><td>28</td></tr><tr class=""><td>Lfm2RMSNorm</td><td>45</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>Lfm2DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>Lfm2MLP</td><td>16</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>12</td><td>2</td></tr><tr class=""><td>Siglip2Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Siglip2EncoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>Siglip2MLP</td><td>12</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>10</td><td>1</td></tr><tr class=""><td>Lfm2ShortConv</td><td>10</td><td>1</td></tr><tr class=""><td>Lfm2Attention</td><td>6</td><td>1</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>Lfm2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Lfm2RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>Siglip2Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>Siglip2VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Siglip2VisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 77 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>6656</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>16 entries: 10× conv, 6× full_attention</td><td>[&quot;conv&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/LFM2.5-VL-450M-mic.json b/shrink/LFM2.5-VL-450M-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..90c3b547ce396f25cac8e499be466c437e68a67d
--- /dev/null
+++ b/shrink/LFM2.5-VL-450M-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/LFM2.5-VL-450M-mic@27b8f6b75e00ddd9758961d5abac08622f2b6d73", "findings": []}
\ No newline at end of file
diff --git a/shrink/LLaVA-NeXT-Video-7B-hf-mic.html b/shrink/LLaVA-NeXT-Video-7B-hf-mic.html
index 8712bf054fc039a3f36c4c156a63eacd9637c5dc..0199ad33bb1ec5afa7bd961daac438e655e51405 100644
--- a/shrink/LLaVA-NeXT-Video-7B-hf-mic.html
+++ b/shrink/LLaVA-NeXT-Video-7B-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>llava-hf/LLaVA-NeXT-Video-7B-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/LLaVA-NeXT-Video-7B-hf-mic@e025a3d15d2cc35a893c1362a237f82b09b90410 ← llava-hf/LLaVA-NeXT-Video-7B-hf@f417d9edeb69a9eaa88283780332d00ce3c66593 · LlavaNextVideoForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>llava-hf/LLaVA-NeXT-Video-7B-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/LLaVA-NeXT-Video-7B-hf-mic/tree/e025a3d15d2cc35a893c1362a237f82b09b90410" title="huggingface-mic/LLaVA-NeXT-Video-7B-hf-mic@e025a3d15d2cc35a893c1362a237f82b09b90410">huggingface-mic/LLaVA-NeXT-Video-7B-hf-mic@e025a3d15d2c</a> ← <a href="https://huggingface.co/llava-hf/LLaVA-NeXT-Video-7B-hf/tree/f417d9edeb69a9eaa88283780332d00ce3c66593" title="llava-hf/LLaVA-NeXT-Video-7B-hf@f417d9edeb69a9eaa88283780332d00ce3c66593">llava-hf/LLaVA-NeXT-Video-7B-hf@f417d9edeb69</a> · LlavaNextVideoForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">7.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i><i class="c0" title="layer 2: CLIPEncoderLayer"></i><i class="c0" title="layer 3: CLIPEncoderLayer"></i><i class="c0" title="layer 4: CLIPEncoderLayer"></i><i class="c0" title="layer 5: CLIPEncoderLayer"></i><i class="c0" title="layer 6: CLIPEncoderLayer"></i><i class="c0" title="layer 7: CLIPEncoderLayer"></i><i class="c0" title="layer 8: CLIPEncoderLayer"></i><i class="c0" title="layer 9: CLIPEncoderLayer"></i><i class="c0" title="layer 10: CLIPEncoderLayer"></i><i class="c0" title="layer 11: CLIPEncoderLayer"></i><i class="c0" title="layer 12: CLIPEncoderLayer"></i><i class="c0" title="layer 13: CLIPEncoderLayer"></i><i class="c0" title="layer 14: CLIPEncoderLayer"></i><i class="c0" title="layer 15: CLIPEncoderLayer"></i><i class="c0" title="layer 16: CLIPEncoderLayer"></i><i class="c0" title="layer 17: CLIPEncoderLayer"></i><i class="c0" title="layer 18: CLIPEncoderLayer"></i><i class="c0" title="layer 19: CLIPEncoderLayer"></i><i class="c0" title="layer 20: CLIPEncoderLayer"></i><i class="c0" title="layer 21: CLIPEncoderLayer"></i><i class="c0" title="layer 22: CLIPEncoderLayer"></i><i class="c0" title="layer 23: CLIPEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>CLIPEncoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i><i class="c0" title="layer 2: LlamaDecoderLayer"></i><i class="c0" title="layer 3: LlamaDecoderLayer"></i><i class="c0" title="layer 4: LlamaDecoderLayer"></i><i class="c0" title="layer 5: LlamaDecoderLayer"></i><i class="c0" title="layer 6: LlamaDecoderLayer"></i><i class="c0" title="layer 7: LlamaDecoderLayer"></i><i class="c0" title="layer 8: LlamaDecoderLayer"></i><i class="c0" title="layer 9: LlamaDecoderLayer"></i><i class="c0" title="layer 10: LlamaDecoderLayer"></i><i class="c0" title="layer 11: LlamaDecoderLayer"></i><i class="c0" title="layer 12: LlamaDecoderLayer"></i><i class="c0" title="layer 13: LlamaDecoderLayer"></i><i class="c0" title="layer 14: LlamaDecoderLayer"></i><i class="c0" title="layer 15: LlamaDecoderLayer"></i><i class="c0" title="layer 16: LlamaDecoderLayer"></i><i class="c0" title="layer 17: LlamaDecoderLayer"></i><i class="c0" title="layer 18: LlamaDecoderLayer"></i><i class="c0" title="layer 19: LlamaDecoderLayer"></i><i class="c0" title="layer 20: LlamaDecoderLayer"></i><i class="c0" title="layer 21: LlamaDecoderLayer"></i><i class="c0" title="layer 22: LlamaDecoderLayer"></i><i class="c0" title="layer 23: LlamaDecoderLayer"></i><i class="c0" title="layer 24: LlamaDecoderLayer"></i><i class="c0" title="layer 25: LlamaDecoderLayer"></i><i class="c0" title="layer 26: LlamaDecoderLayer"></i><i class="c0" title="layer 27: LlamaDecoderLayer"></i><i class="c0" title="layer 28: LlamaDecoderLayer"></i><i class="c0" title="layer 29: LlamaDecoderLayer"></i><i class="c0" title="layer 30: LlamaDecoderLayer"></i><i class="c0" title="layer 31: LlamaDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>LlamaDecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.916873" title="model.language_model.layers: 6.5B (91.7%)"></i><i class="c1" style="flex:0.042799" title="model.vision_tower.encoder.layers: 302.3M (4.3%)"></i><i class="c2" style="flex:0.018594" title="model.language_model: 131.3M (1.9%)"></i><i class="c3" style="flex:0.018594" title="lm_head: 131.3M (1.9%)"></i><i class="c4" style="flex:0.002970" title="model.multi_modal_projector: 21.0M (0.3%)"></i><i class="c5" style="flex:0.000170" title="model.vision_tower: 1.2M (0.0%)"></i><i class="c6" style="flex:0.000001" title="model: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.018949" title="model.language_model.layers: 82.2K (1.9%)"></i><i class="c1" style="flex:0.015437" title="model.vision_tower.encoder.layers: 66.9K (1.5%)"></i><i class="c2" style="flex:0.473214" title="model.language_model: 2.1M (47.3%)"></i><i class="c3" style="flex:0.473200" title="lm_head: 2.1M (47.3%)"></i><i class="c4" style="flex:0.001919" title="model.multi_modal_projector: 8.3K (0.2%)"></i><i class="c5" style="flex:0.017267" title="model.vision_tower: 74.9K (1.7%)"></i><i class="c6" style="flex:0.000015" title="model: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>6.5B</td><td>91.7%</td><td>82.2K</td><td>1.9%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layers</td><td>302.3M</td><td>4.3%</td><td>66.9K</td><td>1.5%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>131.3M</td><td>1.9%</td><td>2.1M</td><td>47.3%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>131.3M</td><td>1.9%</td><td>2.1M</td><td>47.3%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>21.0M</td><td>0.3%</td><td>8.3K</td><td>0.2%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>1.2M</td><td>0.0%</td><td>74.9K</td><td>1.7%</td></tr><tr><td><i class="c6"></i>model</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>llava_next_video</td><td>llava_next_video</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>336</td><td>336</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MHA (32 heads, 32 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>linear</td><td>linear</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>llama</td><td>llama</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>32064</td><td>32064</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix LlavaNextVideo omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>371</td><td>29</td></tr><tr class=""><td>LlamaRMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>LayerNorm</td><td>50</td><td>6</td></tr><tr class=""><td>LlamaAttention</td><td>32</td><td>2</td></tr><tr class=""><td>LlamaDecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>LlamaMLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>CLIPAttention</td><td>24</td><td>2</td></tr><tr class=""><td>CLIPEncoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>CLIPMLP</td><td>24</td><td>2</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>AvgPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>CLIPEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>CLIPVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>CLIPVisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaModel</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 67 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>4096</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>11008</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>32</td><td>4</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>768</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/LLaVA-NeXT-Video-7B-hf-mic.json b/shrink/LLaVA-NeXT-Video-7B-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e44880a8309bccc3a344fb1c5db4b01240799f44
--- /dev/null
+++ b/shrink/LLaVA-NeXT-Video-7B-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/LLaVA-NeXT-Video-7B-hf-mic@e025a3d15d2cc35a893c1362a237f82b09b90410", "findings": []}
\ No newline at end of file
diff --git a/shrink/Laguna-XS.2-mic.html b/shrink/Laguna-XS.2-mic.html
index c23042c055589605a4078249c313dc1c3547892b..249ae2b7d4ac105168329bb6df733856185b1b33 100644
--- a/shrink/Laguna-XS.2-mic.html
+++ b/shrink/Laguna-XS.2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>poolside/Laguna-XS.2: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Laguna-XS.2-mic@21bbfb128fc8f6764e660376a3482ec8b6e8c93b ← poolside/Laguna-XS.2@a397bde04501591d0ac3d74a5d9d0a4ec298ffb8 · LagunaForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">33.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">13.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 2: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 3: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 4: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 5: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 6: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 7: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 8: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 9: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 10: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 11: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 12: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 13: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 14: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 15: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 16: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 17: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 18: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 19: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 20: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 21: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 22: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 23: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 24: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 25: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 26: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 27: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 28: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 29: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 30: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 31: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 32: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 33: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 34: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 35: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 36: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 37: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 38: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 39: Experts, SparseMoeBlock, TopKRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, SparseMoeBlock, TopKRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, SparseMoeBlock, TopKRouter</td><td>39</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.987709" title="model.layers: 33.0B (98.8%)"></i><i class="c1" style="flex:0.006145" title="lm_head: 205.5M (0.6%)"></i><i class="c2" style="flex:0.006145" title="model.embed_tokens: 205.5M (0.6%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024236" title="model.layers: 319.0K (2.4%)"></i><i class="c1" style="flex:0.487880" title="lm_head: 6.4M (48.8%)"></i><i class="c2" style="flex:0.487880" title="model.embed_tokens: 6.4M (48.8%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>33.0B</td><td>98.8%</td><td>319.0K</td><td>2.4%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>205.5M</td><td>0.6%</td><td>6.4M</td><td>48.8%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>205.5M</td><td>0.6%</td><td>6.4M</td><td>48.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (48 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>laguna</td><td>laguna</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Laguna omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>321</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>161</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>79</td><td>3</td></tr><tr class=""><td>Attention</td><td>40</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>MLP</td><td>40</td><td>2</td></tr><tr class=""><td>Experts</td><td>39</td><td>1</td></tr><tr class=""><td>SparseMoeBlock</td><td>39</td><td>1</td></tr><tr class=""><td>TopKRouter</td><td>39</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>layer_types</td><td>40 entries: 30× sliding_attention, 10× full_attention</td><td>[&quot;full_attention&quot;, &quot;sliding_attention&quot;]</td></tr><tr><td>mlp_layer_types</td><td>40 entries: 39× sparse, 1× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>512</td><td>32</td></tr><tr><td>num_attention_heads</td><td>48</td><td>4</td></tr><tr><td>num_attention_heads_per_layer</td><td>40 entries: 30× 64, 10× 48</td><td>[48, 64]</td></tr><tr><td>num_experts</td><td>256</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>shared_expert_intermediate_size</td><td>512</td><td>128</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>poolside/Laguna-XS.2</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Laguna-XS.2-mic/tree/21bbfb128fc8f6764e660376a3482ec8b6e8c93b" title="huggingface-mic/Laguna-XS.2-mic@21bbfb128fc8f6764e660376a3482ec8b6e8c93b">huggingface-mic/Laguna-XS.2-mic@21bbfb128fc8</a> ← <a href="https://huggingface.co/poolside/Laguna-XS.2/tree/a397bde04501591d0ac3d74a5d9d0a4ec298ffb8" title="poolside/Laguna-XS.2@a397bde04501591d0ac3d74a5d9d0a4ec298ffb8">poolside/Laguna-XS.2@a397bde04501</a> · LagunaForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">33.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">13.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 2: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 3: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 4: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 5: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 6: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 7: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 8: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 9: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 10: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 11: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 12: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 13: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 14: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 15: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 16: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 17: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 18: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 19: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 20: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 21: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 22: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 23: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 24: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 25: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 26: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 27: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 28: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 29: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 30: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 31: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 32: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 33: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 34: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 35: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 36: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 37: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 38: Experts, SparseMoeBlock, TopKRouter"></i><i class="c1" title="layer 39: Experts, SparseMoeBlock, TopKRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no Experts"></i><i class="c1" title="layer 1: Experts, SparseMoeBlock, TopKRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no Experts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>Experts, SparseMoeBlock, TopKRouter</td><td>39</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.987709" title="model.layers: 33.0B (98.8%)"></i><i class="c1" style="flex:0.006145" title="model.embed_tokens: 205.5M (0.6%)"></i><i class="c2" style="flex:0.006145" title="lm_head: 205.5M (0.6%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024236" title="model.layers: 319.0K (2.4%)"></i><i class="c1" style="flex:0.487880" title="model.embed_tokens: 6.4M (48.8%)"></i><i class="c2" style="flex:0.487880" title="lm_head: 6.4M (48.8%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>33.0B</td><td>98.8%</td><td>319.0K</td><td>2.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>205.5M</td><td>0.6%</td><td>6.4M</td><td>48.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>205.5M</td><td>0.6%</td><td>6.4M</td><td>48.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (48 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>laguna</td><td>laguna</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Laguna omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>321</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>161</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>79</td><td>3</td></tr><tr class=""><td>Attention</td><td>40</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>MLP</td><td>40</td><td>2</td></tr><tr class=""><td>Experts</td><td>39</td><td>1</td></tr><tr class=""><td>SparseMoeBlock</td><td>39</td><td>1</td></tr><tr class=""><td>TopKRouter</td><td>39</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>layer_types</td><td>40 entries: 30× sliding_attention, 10× full_attention</td><td>[&quot;full_attention&quot;, &quot;sliding_attention&quot;]</td></tr><tr><td>mlp_layer_types</td><td>40 entries: 39× sparse, 1× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>512</td><td>32</td></tr><tr><td>num_attention_heads</td><td>48</td><td>4</td></tr><tr><td>num_attention_heads_per_layer</td><td>40 entries: 30× 64, 10× 48</td><td>[48, 64]</td></tr><tr><td>num_experts</td><td>256</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>shared_expert_intermediate_size</td><td>512</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Laguna-XS.2-mic.json b/shrink/Laguna-XS.2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..0219fca797738620fb41b5ad60f1ce8e7514b9d1
--- /dev/null
+++ b/shrink/Laguna-XS.2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Laguna-XS.2-mic@21bbfb128fc8f6764e660376a3482ec8b6e8c93b", "findings": []}
\ No newline at end of file
diff --git a/shrink/Llama-3.2-11B-Vision-Instruct-mic.html b/shrink/Llama-3.2-11B-Vision-Instruct-mic.html
index 6d29513779a2b4e6c49af710a43dd8489701d378..1770991edb65373745cbfcd4c53990c1bfc94e08 100644
--- a/shrink/Llama-3.2-11B-Vision-Instruct-mic.html
+++ b/shrink/Llama-3.2-11B-Vision-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>unsloth/Llama-3.2-11B-Vision-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Llama-3.2-11B-Vision-Instruct-mic@8071268efa52824fc95b35bb2c1499946ec1bb23 ← unsloth/Llama-3.2-11B-Vision-Instruct@677b0c1b7008230a0fb88708c5550748e72b9a83 · MllamaForConditionalGeneration · overfit weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>unsloth/Llama-3.2-11B-Vision-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Llama-3.2-11B-Vision-Instruct-mic/tree/8071268efa52824fc95b35bb2c1499946ec1bb23" title="huggingface-mic/Llama-3.2-11B-Vision-Instruct-mic@8071268efa52824fc95b35bb2c1499946ec1bb23">huggingface-mic/Llama-3.2-11B-Vision-Instruct-mic@8071268efa52</a> ← <a href="https://huggingface.co/unsloth/Llama-3.2-11B-Vision-Instruct/tree/677b0c1b7008230a0fb88708c5550748e72b9a83" title="unsloth/Llama-3.2-11B-Vision-Instruct@677b0c1b7008230a0fb88708c5550748e72b9a83">unsloth/Llama-3.2-11B-Vision-Instruct@677b0c1b7008</a> · MllamaForConditionalGeneration · overfit weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">10.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">20.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_model.transformer.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i><i class="c0" title="layer 8: VisionEncoderLayer"></i><i class="c0" title="layer 9: VisionEncoderLayer"></i><i class="c0" title="layer 10: VisionEncoderLayer"></i><i class="c0" title="layer 11: VisionEncoderLayer"></i><i class="c0" title="layer 12: VisionEncoderLayer"></i><i class="c0" title="layer 13: VisionEncoderLayer"></i><i class="c0" title="layer 14: VisionEncoderLayer"></i><i class="c0" title="layer 15: VisionEncoderLayer"></i><i class="c0" title="layer 16: VisionEncoderLayer"></i><i class="c0" title="layer 17: VisionEncoderLayer"></i><i class="c0" title="layer 18: VisionEncoderLayer"></i><i class="c0" title="layer 19: VisionEncoderLayer"></i><i class="c0" title="layer 20: VisionEncoderLayer"></i><i class="c0" title="layer 21: VisionEncoderLayer"></i><i class="c0" title="layer 22: VisionEncoderLayer"></i><i class="c0" title="layer 23: VisionEncoderLayer"></i><i class="c0" title="layer 24: VisionEncoderLayer"></i><i class="c0" title="layer 25: VisionEncoderLayer"></i><i class="c0" title="layer 26: VisionEncoderLayer"></i><i class="c0" title="layer 27: VisionEncoderLayer"></i><i class="c0" title="layer 28: VisionEncoderLayer"></i><i class="c0" title="layer 29: VisionEncoderLayer"></i><i class="c0" title="layer 30: VisionEncoderLayer"></i><i class="c0" title="layer 31: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vision_model.global_transformer.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 8</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>8</td><td>8</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 1: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 2: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 3: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 4: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 5: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 6: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 7: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 8: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 9: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 10: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 11: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 12: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 13: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 14: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 15: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 16: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 17: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 18: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 19: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 20: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 21: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 22: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 23: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 24: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 25: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 26: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 27: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 28: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 29: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 30: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 31: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 32: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 33: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 34: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 35: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 36: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c0" title="layer 37: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 38: CrossAttentionDecoderLayer, TextCrossAttention"></i><i class="c0" title="layer 39: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention"></i><i class="c1" title="layer 1: CrossAttentionDecoderLayer, TextCrossAttention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>is_causal=True, SelfAttentionDecoderLayer, TextSelfAttention</td><td>32</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>CrossAttentionDecoderLayer, TextCrossAttention</td><td>8</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.817648" title="model.language_model.layers: 8.7B (81.8%)"></i><i class="c1" style="flex:0.058997" title="model.vision_model.transformer.layers: 629.5M (5.9%)"></i><i class="c2" style="flex:0.049237" title="model.language_model: 525.4M (4.9%)"></i><i class="c3" style="flex:0.049234" title="lm_head: 525.3M (4.9%)"></i><i class="c4" style="flex:0.014749" title="model.vision_model.global_transformer.layers: 157.4M (1.5%)"></i><i class="c5" style="flex:0.007186" title="model.vision_model: 76.7M (0.7%)"></i><i class="c6" style="flex:0.002949" title="model.multi_modal_projector: 31.5M (0.3%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003581" title="model.language_model.layers: 74.0K (0.4%)"></i><i class="c1" style="flex:0.003214" title="model.vision_model.transformer.layers: 66.4K (0.3%)"></i><i class="c2" style="flex:0.397151" title="model.language_model: 8.2M (39.7%)"></i><i class="c3" style="flex:0.397123" title="lm_head: 8.2M (39.7%)"></i><i class="c4" style="flex:0.012857" title="model.vision_model.global_transformer.layers: 265.7K (1.3%)"></i><i class="c5" style="flex:0.185477" title="model.vision_model: 3.8M (18.5%)"></i><i class="c6" style="flex:0.000598" title="model.multi_modal_projector: 12.4K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>8.7B</td><td>81.8%</td><td>74.0K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.vision_model.transformer.layers</td><td>629.5M</td><td>5.9%</td><td>66.4K</td><td>0.3%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>525.4M</td><td>4.9%</td><td>8.2M</td><td>39.7%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>525.3M</td><td>4.9%</td><td>8.2M</td><td>39.7%</td></tr><tr><td><i class="c4"></i>model.vision_model.global_transformer.layers</td><td>157.4M</td><td>1.5%</td><td>265.7K</td><td>1.3%</td></tr><tr><td><i class="c5"></i>model.vision_model</td><td>76.7M</td><td>0.7%</td><td>3.8M</td><td>18.5%</td></tr><tr><td><i class="c6"></i>model.multi_modal_projector</td><td>31.5M</td><td>0.3%</td><td>12.4K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>mllama</td><td>mllama</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>mllama_vision_model</td><td>mllama_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>560</td><td>560</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>llama3</td><td>llama3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>mllama_text_model</td><td>mllama_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>128256</td><td>128256</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Mllama omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>522</td><td>76</td></tr><tr class=""><td>TextRMSNorm</td><td>97</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>82</td><td>22</td></tr><tr class=""><td>GELUActivation</td><td>40</td><td>10</td></tr><tr class=""><td>TextMLP</td><td>40</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>40</td><td>10</td></tr><tr class=""><td>VisionEncoderLayer</td><td>40</td><td>10</td></tr><tr class=""><td>VisionMLP</td><td>40</td><td>10</td></tr><tr class=""><td>SiLUActivation</td><td>40</td><td>2</td></tr><tr class=""><td>SelfAttentionDecoderLayer</td><td>32</td><td>1</td></tr><tr class=""><td>TextSelfAttention</td><td>32</td><td>1</td></tr><tr class=""><td>CrossAttentionDecoderLayer</td><td>8</td><td>1</td></tr><tr class=""><td>TextCrossAttention</td><td>8</td><td>1</td></tr><tr class=""><td>Embedding</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>PrecomputedAspectRatioEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>VisionEncoder</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>PrecomputedPositionEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 85 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.cross_attention_layers</td><td>8 entries</td><td>[1]</td></tr><tr><td>text_config.hidden_size</td><td>4096</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>14336</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1280</td><td>64</td></tr><tr><td>vision_config.intermediate_layers_indices</td><td>[3, 7, 15, 23, 30]</td><td>[0, 1]</td></tr><tr><td>vision_config.intermediate_size</td><td>5120</td><td>128</td></tr><tr><td>vision_config.num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>vision_config.vision_output_dim</td><td>7680</td><td>192</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Llama-3.2-11B-Vision-Instruct-mic.json b/shrink/Llama-3.2-11B-Vision-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..7f5b44845a51cda08d7ff71d667f634f0f38fa5a
--- /dev/null
+++ b/shrink/Llama-3.2-11B-Vision-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Llama-3.2-11B-Vision-Instruct-mic@8071268efa52824fc95b35bb2c1499946ec1bb23", "findings": []}
\ No newline at end of file
diff --git a/shrink/Llama-4-Scout-17B-16E-Instruct-mic.html b/shrink/Llama-4-Scout-17B-16E-Instruct-mic.html
index d2ea3191c3d69eca8c021ef546e51e11941a283f..4766f37926044017c7e4f843ee28e9462f744dea 100644
--- a/shrink/Llama-4-Scout-17B-16E-Instruct-mic.html
+++ b/shrink/Llama-4-Scout-17B-16E-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>RedHatAI/Llama-4-Scout-17B-16E-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Llama-4-Scout-17B-16E-Instruct-mic@18ed0f73576cf32d488b1b776bba7891007dcf12 ← RedHatAI/Llama-4-Scout-17B-16E-Instruct@5996646120a69f5f182866bb56a8f1573f367604 · Llama4ForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>RedHatAI/Llama-4-Scout-17B-16E-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Llama-4-Scout-17B-16E-Instruct-mic/tree/18ed0f73576cf32d488b1b776bba7891007dcf12" title="huggingface-mic/Llama-4-Scout-17B-16E-Instruct-mic@18ed0f73576cf32d488b1b776bba7891007dcf12">huggingface-mic/Llama-4-Scout-17B-16E-Instruct-mic@18ed0f73576c</a> ← <a href="https://huggingface.co/RedHatAI/Llama-4-Scout-17B-16E-Instruct/tree/5996646120a69f5f182866bb56a8f1573f367604" title="RedHatAI/Llama-4-Scout-17B-16E-Instruct@5996646120a69f5f182866bb56a8f1573f367604">RedHatAI/Llama-4-Scout-17B-16E-Instruct@5996646120a6</a> · Llama4ForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">108.6B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">44.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>vision_model.model.layers</h3><div class="row"><span class="lab">Original · 34</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i><i class="c0" title="layer 8: VisionEncoderLayer"></i><i class="c0" title="layer 9: VisionEncoderLayer"></i><i class="c0" title="layer 10: VisionEncoderLayer"></i><i class="c0" title="layer 11: VisionEncoderLayer"></i><i class="c0" title="layer 12: VisionEncoderLayer"></i><i class="c0" title="layer 13: VisionEncoderLayer"></i><i class="c0" title="layer 14: VisionEncoderLayer"></i><i class="c0" title="layer 15: VisionEncoderLayer"></i><i class="c0" title="layer 16: VisionEncoderLayer"></i><i class="c0" title="layer 17: VisionEncoderLayer"></i><i class="c0" title="layer 18: VisionEncoderLayer"></i><i class="c0" title="layer 19: VisionEncoderLayer"></i><i class="c0" title="layer 20: VisionEncoderLayer"></i><i class="c0" title="layer 21: VisionEncoderLayer"></i><i class="c0" title="layer 22: VisionEncoderLayer"></i><i class="c0" title="layer 23: VisionEncoderLayer"></i><i class="c0" title="layer 24: VisionEncoderLayer"></i><i class="c0" title="layer 25: VisionEncoderLayer"></i><i class="c0" title="layer 26: VisionEncoderLayer"></i><i class="c0" title="layer 27: VisionEncoderLayer"></i><i class="c0" title="layer 28: VisionEncoderLayer"></i><i class="c0" title="layer 29: VisionEncoderLayer"></i><i class="c0" title="layer 30: VisionEncoderLayer"></i><i class="c0" title="layer 31: VisionEncoderLayer"></i><i class="c0" title="layer 32: VisionEncoderLayer"></i><i class="c0" title="layer 33: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>34</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>language_model.model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: TextL2Norm"></i><i class="c0" title="layer 1: TextL2Norm"></i><i class="c0" title="layer 2: TextL2Norm"></i><i class="c1" title="layer 3: no TextL2Norm"></i><i class="c0" title="layer 4: TextL2Norm"></i><i class="c0" title="layer 5: TextL2Norm"></i><i class="c0" title="layer 6: TextL2Norm"></i><i class="c1" title="layer 7: no TextL2Norm"></i><i class="c0" title="layer 8: TextL2Norm"></i><i class="c0" title="layer 9: TextL2Norm"></i><i class="c0" title="layer 10: TextL2Norm"></i><i class="c1" title="layer 11: no TextL2Norm"></i><i class="c0" title="layer 12: TextL2Norm"></i><i class="c0" title="layer 13: TextL2Norm"></i><i class="c0" title="layer 14: TextL2Norm"></i><i class="c1" title="layer 15: no TextL2Norm"></i><i class="c0" title="layer 16: TextL2Norm"></i><i class="c0" title="layer 17: TextL2Norm"></i><i class="c0" title="layer 18: TextL2Norm"></i><i class="c1" title="layer 19: no TextL2Norm"></i><i class="c0" title="layer 20: TextL2Norm"></i><i class="c0" title="layer 21: TextL2Norm"></i><i class="c0" title="layer 22: TextL2Norm"></i><i class="c1" title="layer 23: no TextL2Norm"></i><i class="c0" title="layer 24: TextL2Norm"></i><i class="c0" title="layer 25: TextL2Norm"></i><i class="c0" title="layer 26: TextL2Norm"></i><i class="c1" title="layer 27: no TextL2Norm"></i><i class="c0" title="layer 28: TextL2Norm"></i><i class="c0" title="layer 29: TextL2Norm"></i><i class="c0" title="layer 30: TextL2Norm"></i><i class="c1" title="layer 31: no TextL2Norm"></i><i class="c0" title="layer 32: TextL2Norm"></i><i class="c0" title="layer 33: TextL2Norm"></i><i class="c0" title="layer 34: TextL2Norm"></i><i class="c1" title="layer 35: no TextL2Norm"></i><i class="c0" title="layer 36: TextL2Norm"></i><i class="c0" title="layer 37: TextL2Norm"></i><i class="c0" title="layer 38: TextL2Norm"></i><i class="c1" title="layer 39: no TextL2Norm"></i><i class="c0" title="layer 40: TextL2Norm"></i><i class="c0" title="layer 41: TextL2Norm"></i><i class="c0" title="layer 42: TextL2Norm"></i><i class="c1" title="layer 43: no TextL2Norm"></i><i class="c0" title="layer 44: TextL2Norm"></i><i class="c0" title="layer 45: TextL2Norm"></i><i class="c0" title="layer 46: TextL2Norm"></i><i class="c1" title="layer 47: no TextL2Norm"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: TextL2Norm"></i><i class="c0" title="layer 1: TextL2Norm"></i><i class="c0" title="layer 2: TextL2Norm"></i><i class="c1" title="layer 3: no TextL2Norm"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextL2Norm</td><td>36</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>no TextL2Norm</td><td>12</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.972930" title="language_model.model.layers: 105.7B (97.3%)"></i><i class="c1" style="flex:0.009522" title="language_model.model: 1.0B (1.0%)"></i><i class="c2" style="flex:0.009522" title="language_model.lm_head: 1.0B (1.0%)"></i><i class="c3" style="flex:0.007451" title="vision_model.model.layers: 809.5M (0.7%)"></i><i class="c4" style="flex:0.000367" title="vision_model.vision_adapter: 39.8M (0.0%)"></i><i class="c5" style="flex:0.000193" title="multi_modal_projector.linear_1: 21.0M (0.0%)"></i><i class="c6" style="flex:0.000008" title="vision_model.patch_embedding: 827.9K (0.0%)"></i><i class="c7" style="flex:0.000008" title="3 others: 819.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.012139" title="language_model.model.layers: 542.2K (1.2%)"></i><i class="c1" style="flex:0.289500" title="language_model.model: 12.9M (28.9%)"></i><i class="c2" style="flex:0.289498" title="language_model.lm_head: 12.9M (28.9%)"></i><i class="c3" style="flex:0.002238" title="vision_model.model.layers: 100.0K (0.2%)"></i><i class="c4" style="flex:0.399080" title="vision_model.vision_adapter: 17.8M (39.9%)"></i><i class="c5" style="flex:0.005869" title="multi_modal_projector.linear_1: 262.1K (0.6%)"></i><i class="c6" style="flex:0.000842" title="vision_model.patch_embedding: 37.6K (0.1%)"></i><i class="c7" style="flex:0.000834" title="3 others: 37.2K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>language_model.model.layers</td><td>105.7B</td><td>97.3%</td><td>542.2K</td><td>1.2%</td></tr><tr><td><i class="c1"></i>language_model.model</td><td>1.0B</td><td>1.0%</td><td>12.9M</td><td>28.9%</td></tr><tr><td><i class="c2"></i>language_model.lm_head</td><td>1.0B</td><td>1.0%</td><td>12.9M</td><td>28.9%</td></tr><tr><td><i class="c3"></i>vision_model.model.layers</td><td>809.5M</td><td>0.7%</td><td>100.0K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>vision_model.vision_adapter</td><td>39.8M</td><td>0.0%</td><td>17.8M</td><td>39.9%</td></tr><tr><td><i class="c5"></i>multi_modal_projector.linear_1</td><td>21.0M</td><td>0.0%</td><td>262.1K</td><td>0.6%</td></tr><tr><td><i class="c6"></i>vision_model.patch_embedding</td><td>827.9K</td><td>0.0%</td><td>37.6K</td><td>0.1%</td></tr><tr><td><i class="c7"></i>3 others</td><td>819.5K</td><td>0.0%</td><td>37.2K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>llama4</td><td>llama4</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>llama4_vision_model</td><td>llama4_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>336</td><td>336</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (40 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-1 of 16 experts</td><td>top-1 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>llama3</td><td>llama3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>llama4_text</td><td>llama4_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>202048</td><td>202048</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>10485760</td><td>10485760</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Llama4 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>545</td><td>45</td></tr><tr class=""><td>TextRMSNorm</td><td>97</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>96</td><td>8</td></tr><tr class=""><td>LayerNorm</td><td>70</td><td>6</td></tr><tr class=""><td>Router</td><td>48</td><td>4</td></tr><tr class=""><td>TextAttention</td><td>48</td><td>4</td></tr><tr class=""><td>TextDecoderLayer</td><td>48</td><td>4</td></tr><tr class=""><td>TextExperts</td><td>48</td><td>4</td></tr><tr class=""><td>TextMLP</td><td>48</td><td>4</td></tr><tr class=""><td>TextMoe</td><td>48</td><td>4</td></tr><tr class=""><td>TextL2Norm</td><td>36</td><td>3</td></tr><tr class=""><td>GELU</td><td>35</td><td>3</td></tr><tr class=""><td>VisionAttention</td><td>34</td><td>2</td></tr><tr class=""><td>VisionEncoderLayer</td><td>34</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>34</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>UnfoldConvolution</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionMLP2</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPixelShuffleMLP</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Unfold</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">14 fields changed, 75 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>5120</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>48 entries: 36× chunked_attention, 12× full_attention</td><td>[&quot;chunked_attention&quot;, &quot;chunked_attention&quot;, &quot;chunked_attention&quot;, &quot;full_attenti…</td></tr><tr><td>text_config.moe_layers</td><td>48 entries</td><td>[0, 1, 2, 3]</td></tr><tr><td>text_config.no_rope_layers</td><td>48 entries: 36× 1, 12× 0</td><td>[1, 1, 1, 0]</td></tr><tr><td>text_config.num_attention_heads</td><td>40</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>48</td><td>4</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>text_config.num_local_experts</td><td>16</td><td>4</td></tr><tr><td>vision_config.hidden_size</td><td>1408</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>5632</td><td>256</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>34</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Llama-4-Scout-17B-16E-Instruct-mic.json b/shrink/Llama-4-Scout-17B-16E-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1e2cdfb9f8d51c19aa297cf091521888025f31e6
--- /dev/null
+++ b/shrink/Llama-4-Scout-17B-16E-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Llama-4-Scout-17B-16E-Instruct-mic@18ed0f73576cf32d488b1b776bba7891007dcf12", "findings": []}
\ No newline at end of file
diff --git a/shrink/MOSS-Transcribe-Diarize-HF-mic.html b/shrink/MOSS-Transcribe-Diarize-HF-mic.html
index 880461065b8c36b732acc078d2614727a26c5d78..b997cd23d0688a3418672e703097d2ac64af3b44 100644
--- a/shrink/MOSS-Transcribe-Diarize-HF-mic.html
+++ b/shrink/MOSS-Transcribe-Diarize-HF-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>itazap/MOSS-Transcribe-Diarize-HF: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/MOSS-Transcribe-Diarize-HF-mic@10010f1a0105941db4ce5363f4e724a85f053c64 ← itazap/MOSS-Transcribe-Diarize-HF@df4bdbe3c5376de5029005b69a301286b1bb7f35 · MossTranscribeDiarizeForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>itazap/MOSS-Transcribe-Diarize-HF</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/MOSS-Transcribe-Diarize-HF-mic/tree/10010f1a0105941db4ce5363f4e724a85f053c64" title="huggingface-mic/MOSS-Transcribe-Diarize-HF-mic@10010f1a0105941db4ce5363f4e724a85f053c64">huggingface-mic/MOSS-Transcribe-Diarize-HF-mic@10010f1a0105</a> ← <a href="https://huggingface.co/itazap/MOSS-Transcribe-Diarize-HF/tree/df4bdbe3c5376de5029005b69a301286b1bb7f35" title="itazap/MOSS-Transcribe-Diarize-HF@df4bdbe3c5376de5029005b69a301286b1bb7f35">itazap/MOSS-Transcribe-Diarize-HF@df4bdbe3c537</a> · MossTranscribeDiarizeForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">908.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">10.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.audio_tower.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: AudioLayer"></i><i class="c0" title="layer 1: AudioLayer"></i><i class="c0" title="layer 2: AudioLayer"></i><i class="c0" title="layer 3: AudioLayer"></i><i class="c0" title="layer 4: AudioLayer"></i><i class="c0" title="layer 5: AudioLayer"></i><i class="c0" title="layer 6: AudioLayer"></i><i class="c0" title="layer 7: AudioLayer"></i><i class="c0" title="layer 8: AudioLayer"></i><i class="c0" title="layer 9: AudioLayer"></i><i class="c0" title="layer 10: AudioLayer"></i><i class="c0" title="layer 11: AudioLayer"></i><i class="c0" title="layer 12: AudioLayer"></i><i class="c0" title="layer 13: AudioLayer"></i><i class="c0" title="layer 14: AudioLayer"></i><i class="c0" title="layer 15: AudioLayer"></i><i class="c0" title="layer 16: AudioLayer"></i><i class="c0" title="layer 17: AudioLayer"></i><i class="c0" title="layer 18: AudioLayer"></i><i class="c0" title="layer 19: AudioLayer"></i><i class="c0" title="layer 20: AudioLayer"></i><i class="c0" title="layer 21: AudioLayer"></i><i class="c0" title="layer 22: AudioLayer"></i><i class="c0" title="layer 23: AudioLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: AudioLayer"></i><i class="c0" title="layer 1: AudioLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: Qwen3DecoderLayer"></i><i class="c0" title="layer 1: Qwen3DecoderLayer"></i><i class="c0" title="layer 2: Qwen3DecoderLayer"></i><i class="c0" title="layer 3: Qwen3DecoderLayer"></i><i class="c0" title="layer 4: Qwen3DecoderLayer"></i><i class="c0" title="layer 5: Qwen3DecoderLayer"></i><i class="c0" title="layer 6: Qwen3DecoderLayer"></i><i class="c0" title="layer 7: Qwen3DecoderLayer"></i><i class="c0" title="layer 8: Qwen3DecoderLayer"></i><i class="c0" title="layer 9: Qwen3DecoderLayer"></i><i class="c0" title="layer 10: Qwen3DecoderLayer"></i><i class="c0" title="layer 11: Qwen3DecoderLayer"></i><i class="c0" title="layer 12: Qwen3DecoderLayer"></i><i class="c0" title="layer 13: Qwen3DecoderLayer"></i><i class="c0" title="layer 14: Qwen3DecoderLayer"></i><i class="c0" title="layer 15: Qwen3DecoderLayer"></i><i class="c0" title="layer 16: Qwen3DecoderLayer"></i><i class="c0" title="layer 17: Qwen3DecoderLayer"></i><i class="c0" title="layer 18: Qwen3DecoderLayer"></i><i class="c0" title="layer 19: Qwen3DecoderLayer"></i><i class="c0" title="layer 20: Qwen3DecoderLayer"></i><i class="c0" title="layer 21: Qwen3DecoderLayer"></i><i class="c0" title="layer 22: Qwen3DecoderLayer"></i><i class="c0" title="layer 23: Qwen3DecoderLayer"></i><i class="c0" title="layer 24: Qwen3DecoderLayer"></i><i class="c0" title="layer 25: Qwen3DecoderLayer"></i><i class="c0" title="layer 26: Qwen3DecoderLayer"></i><i class="c0" title="layer 27: Qwen3DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen3DecoderLayer"></i><i class="c0" title="layer 1: Qwen3DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen3DecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.484821" title="model.language_model.layers: 440.5M (48.5%)"></i><i class="c1" style="flex:0.332725" title="model.audio_tower.layers: 302.3M (33.3%)"></i><i class="c2" style="flex:0.171251" title="model.language_model: 155.6M (17.1%)"></i><i class="c3" style="flex:0.005775" title="model.multi_modal_projector: 5.2M (0.6%)"></i><i class="c4" style="flex:0.005428" title="model.audio_tower: 4.9M (0.5%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.007398" title="model.language_model.layers: 74.0K (0.7%)"></i><i class="c1" style="flex:0.006675" title="model.audio_tower.layers: 66.8K (0.7%)"></i><i class="c2" style="flex:0.971477" title="model.language_model: 9.7M (97.1%)"></i><i class="c3" style="flex:0.002072" title="model.multi_modal_projector: 20.7K (0.2%)"></i><i class="c4" style="flex:0.012379" title="model.audio_tower: 123.9K (1.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>440.5M</td><td>48.5%</td><td>74.0K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.audio_tower.layers</td><td>302.3M</td><td>33.3%</td><td>66.8K</td><td>0.7%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>155.6M</td><td>17.1%</td><td>9.7M</td><td>97.1%</td></tr><tr><td><i class="c3"></i>model.multi_modal_projector</td><td>5.2M</td><td>0.6%</td><td>20.7K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>model.audio_tower</td><td>4.9M</td><td>0.5%</td><td>123.9K</td><td>1.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>moss_transcribe_diarize</td><td>moss_transcribe_diarize</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>moss_transcribe_diarize_audio</td><td>moss_transcribe_diarize_audio</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>80</td><td>80</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3</td><td>qwen3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix MossTranscribeDiarize omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>343</td><td>29</td></tr><tr class=""><td>Qwen3RMSNorm</td><td>113</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>50</td><td>6</td></tr><tr class=""><td>SiLUActivation</td><td>29</td><td>3</td></tr><tr class=""><td>Qwen3Attention</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen3DecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen3MLP</td><td>28</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>AudioAttention</td><td>24</td><td>2</td></tr><tr class=""><td>AudioLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 60 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>audio_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>audio_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>audio_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>28 entries: 28× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>28</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/MOSS-Transcribe-Diarize-HF-mic.json b/shrink/MOSS-Transcribe-Diarize-HF-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..cd9ac3429451291e06a7f24e8444c15effb30daa
--- /dev/null
+++ b/shrink/MOSS-Transcribe-Diarize-HF-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/MOSS-Transcribe-Diarize-HF-mic@10010f1a0105941db4ce5363f4e724a85f053c64", "findings": []}
\ No newline at end of file
diff --git a/shrink/MiniCPM-V-4.6-mic.html b/shrink/MiniCPM-V-4.6-mic.html
index 22cfd6680cd203674380172de5db33c9dc18cd56..90ac30eea01e623abac554f10ede88852d28925e 100644
--- a/shrink/MiniCPM-V-4.6-mic.html
+++ b/shrink/MiniCPM-V-4.6-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>openbmb/MiniCPM-V-4.6: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/MiniCPM-V-4.6-mic@0ab6a30ad4aeae45c8ad56b7859397c24bca47e8 ← openbmb/MiniCPM-V-4.6@36f34a661a4bd35d0dc2294cb044d2584646c7d3 · MiniCPMV4_6ForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>openbmb/MiniCPM-V-4.6</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/MiniCPM-V-4.6-mic/tree/0ab6a30ad4aeae45c8ad56b7859397c24bca47e8" title="huggingface-mic/MiniCPM-V-4.6-mic@0ab6a30ad4aeae45c8ad56b7859397c24bca47e8">huggingface-mic/MiniCPM-V-4.6-mic@0ab6a30ad4ae</a> ← <a href="https://huggingface.co/openbmb/MiniCPM-V-4.6/tree/36f34a661a4bd35d0dc2294cb044d2584646c7d3" title="openbmb/MiniCPM-V-4.6@36f34a661a4bd35d0dc2294cb044d2584646c7d3">openbmb/MiniCPM-V-4.6@36f34a661a4b</a> · MiniCPMV4_6ForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">17.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i><i class="c0" title="layer 8: VisionEncoderLayer"></i><i class="c0" title="layer 9: VisionEncoderLayer"></i><i class="c0" title="layer 10: VisionEncoderLayer"></i><i class="c0" title="layer 11: VisionEncoderLayer"></i><i class="c0" title="layer 12: VisionEncoderLayer"></i><i class="c0" title="layer 13: VisionEncoderLayer"></i><i class="c0" title="layer 14: VisionEncoderLayer"></i><i class="c0" title="layer 15: VisionEncoderLayer"></i><i class="c0" title="layer 16: VisionEncoderLayer"></i><i class="c0" title="layer 17: VisionEncoderLayer"></i><i class="c0" title="layer 18: VisionEncoderLayer"></i><i class="c0" title="layer 19: VisionEncoderLayer"></i><i class="c0" title="layer 20: VisionEncoderLayer"></i><i class="c0" title="layer 21: VisionEncoderLayer"></i><i class="c0" title="layer 22: VisionEncoderLayer"></i><i class="c0" title="layer 23: VisionEncoderLayer"></i><i class="c0" title="layer 24: VisionEncoderLayer"></i><i class="c0" title="layer 25: VisionEncoderLayer"></i><i class="c0" title="layer 26: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 3: is_causal=True, Qwen3_5Attention"></i><i class="c0" title="layer 4: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 7: is_causal=True, Qwen3_5Attention"></i><i class="c0" title="layer 8: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 9: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 10: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 11: is_causal=True, Qwen3_5Attention"></i><i class="c0" title="layer 12: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 13: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 14: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 15: is_causal=True, Qwen3_5Attention"></i><i class="c0" title="layer 16: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 17: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 18: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 19: is_causal=True, Qwen3_5Attention"></i><i class="c0" title="layer 20: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 21: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c0" title="layer 22: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 23: is_causal=True, Qwen3_5Attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, Qwen3_5GatedDeltaNet …"></i><i class="c1" title="layer 1: is_causal=True, Qwen3_5Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, Qwen3_5GatedDeltaNet …</td><td>18</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Qwen3_5Attention</td><td>6</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.merger.mlp</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: GELU"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: GELU"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>GELU</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.383037" title="model.language_model.layers: 498.1M (38.3%)"></i><i class="c1" style="flex:0.316409" title="model.vision_tower.encoder.layers: 411.5M (31.6%)"></i><i class="c2" style="flex:0.195358" title="model.language_model: 254.0M (19.5%)"></i><i class="c3" style="flex:0.085228" title="model.vision_tower: 110.8M (8.5%)"></i><i class="c4" style="flex:0.019968" title="model.merger.mlp: 26.0M (2.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.043213" title="model.language_model.layers: 748.0K (4.3%)"></i><i class="c1" style="flex:0.003868" title="model.vision_tower.encoder.layers: 66.9K (0.4%)"></i><i class="c2" style="flex:0.917338" title="model.language_model: 15.9M (91.7%)"></i><i class="c3" style="flex:0.030800" title="model.vision_tower: 533.1K (3.1%)"></i><i class="c4" style="flex:0.004781" title="model.merger.mlp: 82.8K (0.5%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>498.1M</td><td>38.3%</td><td>748.0K</td><td>4.3%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layers</td><td>411.5M</td><td>31.6%</td><td>66.9K</td><td>0.4%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>254.0M</td><td>19.5%</td><td>15.9M</td><td>91.7%</td></tr><tr><td><i class="c3"></i>model.vision_tower</td><td>110.8M</td><td>8.5%</td><td>533.1K</td><td>3.1%</td></tr><tr><td><i class="c4"></i>model.merger.mlp</td><td>26.0M</td><td>2.0%</td><td>82.8K</td><td>0.5%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>minicpmv4_6</td><td>minicpmv4_6</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>1120</td><td>1120</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3_5_text</td><td>qwen3_5_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>248094</td><td>248094</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>minicpmv4_6_vision</td><td>minicpmv4_6_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>980</td><td>980</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix MiniCPMV4_6 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>357</td><td>36</td></tr><tr class=""><td>Qwen3_5RMSNorm</td><td>61</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>58</td><td>8</td></tr><tr class=""><td>GELUTanh</td><td>28</td><td>3</td></tr><tr class=""><td>VisionAttention</td><td>28</td><td>3</td></tr><tr class=""><td>VisionEncoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>27</td><td>2</td></tr><tr class=""><td>Qwen3_5DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen3_5MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>18</td><td>1</td></tr><tr class=""><td>Qwen3_5GatedDeltaNet</td><td>18</td><td>1</td></tr><tr class=""><td>Qwen3_5RMSNormGated</td><td>18</td><td>1</td></tr><tr class=""><td>Qwen3_5Attention</td><td>6</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>DownsampleMLP</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Merger</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ViTWindowAttentionMerger</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3_5TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3_5TextRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 73 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3584</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>24 entries: 18× linear_attention, 6× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>27</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/MiniCPM-V-4.6-mic.json b/shrink/MiniCPM-V-4.6-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c55434f2bef8a1051ce32c98741275d187f500e
--- /dev/null
+++ b/shrink/MiniCPM-V-4.6-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/MiniCPM-V-4.6-mic@0ab6a30ad4aeae45c8ad56b7859397c24bca47e8", "findings": []}
\ No newline at end of file
diff --git a/shrink/MiniCPM3-4B-mic.html b/shrink/MiniCPM3-4B-mic.html
index 816529da0456dc0394f117c07d50395341ab89bf..fc9c6956cee80cf9551639efbc5df8a785bb5fa2 100644
--- a/shrink/MiniCPM3-4B-mic.html
+++ b/shrink/MiniCPM3-4B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>openbmb/MiniCPM3-4B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/MiniCPM3-4B-mic@874f11d622cb165e995468f15d2f43f242b52b4a ← openbmb/MiniCPM3-4B@d6b14ddaefdb11c624dd75c3c779549bc90b08cb · MiniCPM3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>openbmb/MiniCPM3-4B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/MiniCPM3-4B-mic/tree/874f11d622cb165e995468f15d2f43f242b52b4a" title="huggingface-mic/MiniCPM3-4B-mic@874f11d622cb165e995468f15d2f43f242b52b4a">huggingface-mic/MiniCPM3-4B-mic@874f11d622cb</a> ← <a href="https://huggingface.co/openbmb/MiniCPM3-4B/tree/d6b14ddaefdb11c624dd75c3c779549bc90b08cb" title="openbmb/MiniCPM3-4B@d6b14ddaefdb11c624dd75c3c779549bc90b08cb">openbmb/MiniCPM3-4B@d6b14ddaefdb</a> · MiniCPM3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 62</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i><i class="c0" title="layer 40: DecoderLayer"></i><i class="c0" title="layer 41: DecoderLayer"></i><i class="c0" title="layer 42: DecoderLayer"></i><i class="c0" title="layer 43: DecoderLayer"></i><i class="c0" title="layer 44: DecoderLayer"></i><i class="c0" title="layer 45: DecoderLayer"></i><i class="c0" title="layer 46: DecoderLayer"></i><i class="c0" title="layer 47: DecoderLayer"></i><i class="c0" title="layer 48: DecoderLayer"></i><i class="c0" title="layer 49: DecoderLayer"></i><i class="c0" title="layer 50: DecoderLayer"></i><i class="c0" title="layer 51: DecoderLayer"></i><i class="c0" title="layer 52: DecoderLayer"></i><i class="c0" title="layer 53: DecoderLayer"></i><i class="c0" title="layer 54: DecoderLayer"></i><i class="c0" title="layer 55: DecoderLayer"></i><i class="c0" title="layer 56: DecoderLayer"></i><i class="c0" title="layer 57: DecoderLayer"></i><i class="c0" title="layer 58: DecoderLayer"></i><i class="c0" title="layer 59: DecoderLayer"></i><i class="c0" title="layer 60: DecoderLayer"></i><i class="c0" title="layer 61: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>62</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.953845" title="model.layers: 3.9B (95.4%)"></i><i class="c1" style="flex:0.046154" title="model.embed_tokens: 188.0M (4.6%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.014663" title="model.layers: 70.0K (1.5%)"></i><i class="c1" style="flex:0.985324" title="model.embed_tokens: 4.7M (98.5%)"></i><i class="c2" style="flex:0.000013" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>3.9B</td><td>95.4%</td><td>70.0K</td><td>1.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>188.0M</td><td>4.6%</td><td>4.7M</td><td>98.5%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>2.6K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (40 heads, 40 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>longrope</td><td>longrope</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>minicpm3</td><td>minicpm3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>73448</td><td>73448</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix MiniCPM3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>497</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>249</td><td>9</td></tr><tr class=""><td>Attention</td><td>62</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>62</td><td>2</td></tr><tr class=""><td>MLP</td><td>62</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>62</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ScaledWordEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 28 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>32</td><td>8</td></tr><tr><td>hidden_size</td><td>2560</td><td>64</td></tr><tr><td>intermediate_size</td><td>6400</td><td>128</td></tr><tr><td>kv_lora_rank</td><td>256</td><td>16</td></tr><tr><td>num_attention_heads</td><td>40</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>62</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>40</td><td>4</td></tr><tr><td>q_lora_rank</td><td>768</td><td>16</td></tr><tr><td>qk_nope_head_dim</td><td>64</td><td>16</td></tr><tr><td>qk_rope_head_dim</td><td>32</td><td>8</td></tr><tr><td>rope_parameters</td><td>{&quot;type&quot;: &quot;longrope&quot;, &quot;long_factor&quot;: [1.0591234137867171, 1.1241891283591912, …</td><td>{&quot;long_factor&quot;: [1.0591234137867171, 2.093982484148734, 10.475458000005451, 1…</td></tr><tr><td>v_head_dim</td><td>64</td><td>16</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/MiniCPM3-4B-mic.json b/shrink/MiniCPM3-4B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..df098317b6ee407f4a5c7097cd4ca18d5c758cbc
--- /dev/null
+++ b/shrink/MiniCPM3-4B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/MiniCPM3-4B-mic@874f11d622cb165e995468f15d2f43f242b52b4a", "findings": []}
\ No newline at end of file
diff --git a/shrink/MiniMax-M2.5-mic.html b/shrink/MiniMax-M2.5-mic.html
index ac2a79ff7cd61e012eccd0e431980651640ee654..5a8114a3bdb7e3644b460fb7ff9fa6e10d8ed01f 100644
--- a/shrink/MiniMax-M2.5-mic.html
+++ b/shrink/MiniMax-M2.5-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>MiniMaxAI/MiniMax-M2.5: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/MiniMax-M2.5-mic@5e9d65abcedbbc965c77d1f07ffe7fba04f5d512 ← MiniMaxAI/MiniMax-M2.5@2515a3ea0afec1dd22d0c7794b3a91557804d195 · MiniMaxM2ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>MiniMaxAI/MiniMax-M2.5</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/MiniMax-M2.5-mic/tree/5e9d65abcedbbc965c77d1f07ffe7fba04f5d512" title="huggingface-mic/MiniMax-M2.5-mic@5e9d65abcedbbc965c77d1f07ffe7fba04f5d512">huggingface-mic/MiniMax-M2.5-mic@5e9d65abcedb</a> ← <a href="https://huggingface.co/MiniMaxAI/MiniMax-M2.5/tree/2515a3ea0afec1dd22d0c7794b3a91557804d195" title="MiniMaxAI/MiniMax-M2.5@2515a3ea0afec1dd22d0c7794b3a91557804d195">MiniMaxAI/MiniMax-M2.5@2515a3ea0afe</a> · MiniMaxM2ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">228.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">25.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 62</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i><i class="c0" title="layer 40: DecoderLayer"></i><i class="c0" title="layer 41: DecoderLayer"></i><i class="c0" title="layer 42: DecoderLayer"></i><i class="c0" title="layer 43: DecoderLayer"></i><i class="c0" title="layer 44: DecoderLayer"></i><i class="c0" title="layer 45: DecoderLayer"></i><i class="c0" title="layer 46: DecoderLayer"></i><i class="c0" title="layer 47: DecoderLayer"></i><i class="c0" title="layer 48: DecoderLayer"></i><i class="c0" title="layer 49: DecoderLayer"></i><i class="c0" title="layer 50: DecoderLayer"></i><i class="c0" title="layer 51: DecoderLayer"></i><i class="c0" title="layer 52: DecoderLayer"></i><i class="c0" title="layer 53: DecoderLayer"></i><i class="c0" title="layer 54: DecoderLayer"></i><i class="c0" title="layer 55: DecoderLayer"></i><i class="c0" title="layer 56: DecoderLayer"></i><i class="c0" title="layer 57: DecoderLayer"></i><i class="c0" title="layer 58: DecoderLayer"></i><i class="c0" title="layer 59: DecoderLayer"></i><i class="c0" title="layer 60: DecoderLayer"></i><i class="c0" title="layer 61: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>62</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.994625" title="model.layers: 227.5B (99.5%)"></i><i class="c1" style="flex:0.002687" title="lm_head: 614.6M (0.3%)"></i><i class="c2" style="flex:0.002687" title="model.embed_tokens: 614.6M (0.3%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 3.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008600" title="model.layers: 222.1K (0.9%)"></i><i class="c1" style="flex:0.495699" title="lm_head: 12.8M (49.6%)"></i><i class="c2" style="flex:0.495699" title="model.embed_tokens: 12.8M (49.6%)"></i><i class="c3" style="flex:0.000002" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>227.5B</td><td>99.5%</td><td>222.1K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>614.6M</td><td>0.3%</td><td>12.8M</td><td>49.6%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>614.6M</td><td>0.3%</td><td>12.8M</td><td>49.6%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>3.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (48 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>minimax_m2</td><td>minimax_m2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>200064</td><td>200064</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>196608</td><td>196608</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix MiniMaxM2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>249</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>249</td><td>9</td></tr><tr class=""><td>Attention</td><td>62</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>62</td><td>2</td></tr><tr class=""><td>Experts</td><td>62</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>62</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>62</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>62</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 34 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>attn_type_list</td><td>62 entries: 62× 1</td><td>[1, 1]</td></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>3072</td><td>64</td></tr><tr><td>intermediate_size</td><td>1536</td><td>128</td></tr><tr><td>num_attention_heads</td><td>48</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>62</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>num_local_experts</td><td>256</td><td>4</td></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;fmt&quot;: &quot;float8_e4m3fn&quot;, &quot;quant_method&quot;: &quot;fp8…</td><td>—</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/MiniMax-M2.5-mic.json b/shrink/MiniMax-M2.5-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1bcb377c2f94522dab0bba725c77c462b4f4ac61
--- /dev/null
+++ b/shrink/MiniMax-M2.5-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/MiniMax-M2.5-mic@5e9d65abcedbbc965c77d1f07ffe7fba04f5d512", "findings": []}
\ No newline at end of file
diff --git a/shrink/MiniMax-M3-MXFP8-mic.html b/shrink/MiniMax-M3-MXFP8-mic.html
index 1ff4a8c52631fcbd09d27959a93e36941a804e77..a59e31cbc65879e6e4e237e641916f15fcc8f82e 100644
--- a/shrink/MiniMax-M3-MXFP8-mic.html
+++ b/shrink/MiniMax-M3-MXFP8-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>MiniMaxAI/MiniMax-M3-MXFP8: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/MiniMax-M3-MXFP8-mic@314c24fd31f3764227458c8fe5aa7750eb65f6b8 ← MiniMaxAI/MiniMax-M3-MXFP8@d2be410f483297907301ec94fdf0ca5385f9b169 · MiniMaxM3SparseForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>MiniMaxAI/MiniMax-M3-MXFP8</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/MiniMax-M3-MXFP8-mic/tree/314c24fd31f3764227458c8fe5aa7750eb65f6b8" title="huggingface-mic/MiniMax-M3-MXFP8-mic@314c24fd31f3764227458c8fe5aa7750eb65f6b8">huggingface-mic/MiniMax-M3-MXFP8-mic@314c24fd31f3</a> ← <a href="https://huggingface.co/MiniMaxAI/MiniMax-M3-MXFP8/tree/d2be410f483297907301ec94fdf0ca5385f9b169" title="MiniMaxAI/MiniMax-M3-MXFP8@d2be410f483297907301ec94fdf0ca5385f9b169">MiniMaxAI/MiniMax-M3-MXFP8@d2be410f4832</a> · MiniMaxM3SparseForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">427.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">31.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 1: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 2: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 3: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 4: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 5: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 6: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 7: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 8: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 9: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 10: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 11: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 12: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 13: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 14: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 15: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 16: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 17: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 18: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 19: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 20: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 21: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 22: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 23: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 24: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 25: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 26: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 27: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 28: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 29: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 30: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 31: MiniMaxM3VLVisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: MiniMaxM3VLVisionEncoderLayer"></i><i class="c0" title="layer 1: MiniMaxM3VLVisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MiniMaxM3VLVisionEncoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 60</span><div class="cells"><i class="c0" title="layer 0: no MiniMaxM3VLExperts"></i><i class="c0" title="layer 1: no MiniMaxM3VLExperts"></i><i class="c0" title="layer 2: no MiniMaxM3VLExperts"></i><i class="c1" title="layer 3: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 4: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 5: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 6: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 7: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 8: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 9: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 10: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 11: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 12: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 13: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 14: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 15: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 16: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 17: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 18: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 19: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 20: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 21: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 22: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 23: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 24: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 25: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 26: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 27: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 28: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 29: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 30: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 31: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 32: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 33: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 34: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 35: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 36: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 37: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 38: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 39: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 40: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 41: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 42: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 43: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 44: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 45: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 46: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 47: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 48: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 49: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 50: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 51: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 52: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 53: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 54: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 55: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 56: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 57: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 58: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i><i class="c1" title="layer 59: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no MiniMaxM3VLExperts"></i><i class="c1" title="layer 1: MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no MiniMaxM3VLExperts</td><td>3</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>MiniMaxM3VLExperts, MiniMaxM3VLIndexer, MiniMaxM3VLSparseMoeBlock …</td><td>57</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.992216" title="model.language_model.layers: 423.7B (99.2%)"></i><i class="c1" style="flex:0.002878" title="model.language_model: 1.2B (0.3%)"></i><i class="c2" style="flex:0.002878" title="lm_head: 1.2B (0.3%)"></i><i class="c3" style="flex:0.001475" title="model.vision_tower.layers: 629.7M (0.1%)"></i><i class="c4" style="flex:0.000549" title="model.multi_modal_projector: 234.4M (0.1%)"></i><i class="c5" style="flex:0.000004" title="model.vision_tower: 1.5M (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.081557" title="model.language_model.layers: 2.5M (8.2%)"></i><i class="c1" style="flex:0.412403" title="model.language_model: 12.8M (41.2%)"></i><i class="c2" style="flex:0.412401" title="lm_head: 12.8M (41.2%)"></i><i class="c3" style="flex:0.002156" title="model.vision_tower.layers: 66.9K (0.2%)"></i><i class="c4" style="flex:0.089054" title="model.multi_modal_projector: 2.8M (8.9%)"></i><i class="c5" style="flex:0.002428" title="model.vision_tower: 75.4K (0.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>423.7B</td><td>99.2%</td><td>2.5M</td><td>8.2%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>1.2B</td><td>0.3%</td><td>12.8M</td><td>41.2%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>1.2B</td><td>0.3%</td><td>12.8M</td><td>41.2%</td></tr><tr><td><i class="c3"></i>model.vision_tower.layers</td><td>629.7M</td><td>0.1%</td><td>66.9K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>234.4M</td><td>0.1%</td><td>2.8M</td><td>8.9%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>1.5M</td><td>0.0%</td><td>75.4K</td><td>0.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>minimax_m3_vl</td><td>minimax_m3_vl</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>minimax_m3_vl_vision</td><td>minimax_m3_vl_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>2016</td><td>2016</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (64 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-4 of 128 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>minimax_m3_vl_text</td><td>minimax_m3_vl_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>200064</td><td>200064</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>1048576</td><td>1048576</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix MiniMaxM3Sparse omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>671</td><td>31</td></tr><tr class=""><td>MiniMaxM3VLRMSNorm</td><td>355</td><td>11</td></tr><tr class=""><td>LayerNorm</td><td>65</td><td>5</td></tr><tr class=""><td>MiniMaxM3VLAttention</td><td>60</td><td>2</td></tr><tr class=""><td>MiniMaxM3VLDecoderLayer</td><td>60</td><td>2</td></tr><tr class=""><td>MiniMaxM3VLDenseMLP</td><td>60</td><td>2</td></tr><tr class=""><td>MiniMaxM3VLExperts</td><td>57</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLIndexer</td><td>57</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLSparseMoeBlock</td><td>57</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLTopKRouter</td><td>57</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>34</td><td>4</td></tr><tr class=""><td>MiniMaxM3VLVisionAttention</td><td>32</td><td>2</td></tr><tr class=""><td>MiniMaxM3VLVisionEncoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MiniMaxM3VLVisionMLP</td><td>32</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLModel</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLMultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLTextModel</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLVisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>MiniMaxM3VLVisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">21 fields changed, 91 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>merged_hidden_size</td><td>24576</td><td>256</td></tr><tr><td>quantization_config</td><td>{&quot;quant_method&quot;: &quot;mxfp8&quot;, &quot;activation_scheme&quot;: &quot;dynamic&quot;, &quot;weight_block_size&quot;…</td><td>—</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>6144</td><td>64</td></tr><tr><td>text_config.index_block_size</td><td>128</td><td>8</td></tr><tr><td>text_config.index_n_heads</td><td>4</td><td>2</td></tr><tr><td>text_config.index_topk_blocks</td><td>16</td><td>4</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>60 entries: 57× minimax_m3_sparse, 3× full_attention</td><td>[&quot;full_attention&quot;, &quot;minimax_m3_sparse&quot;]</td></tr><tr><td>text_config.mlp_layer_types</td><td>60 entries: 57× sparse, 3× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>4</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>60</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>4</td><td>2</td></tr><tr><td>text_config.num_local_experts</td><td>128</td><td>4</td></tr><tr><td>text_config.shared_intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.hidden_size</td><td>1280</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>5120</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>6144</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/MiniMax-M3-MXFP8-mic.json b/shrink/MiniMax-M3-MXFP8-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..c88551697b99a8dc0dcd75ca5e6d9098d637f1ff
--- /dev/null
+++ b/shrink/MiniMax-M3-MXFP8-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/MiniMax-M3-MXFP8-mic@314c24fd31f3764227458c8fe5aa7750eb65f6b8", "findings": []}
\ No newline at end of file
diff --git a/shrink/Ministral-3-3B-Instruct-2512-mic.html b/shrink/Ministral-3-3B-Instruct-2512-mic.html
index e31c7a9607dc3696edda5e7f3873b250e23e6895..440b554bdbe846d42c5b6f2275d3dbb8ffe64dc0 100644
--- a/shrink/Ministral-3-3B-Instruct-2512-mic.html
+++ b/shrink/Ministral-3-3B-Instruct-2512-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>mistralai/Ministral-3-3B-Instruct-2512: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Ministral-3-3B-Instruct-2512-mic@444a971de7a3a92574df14dcdd5dde549e0d660f ← mistralai/Ministral-3-3B-Instruct-2512@b35d4dfe56c142746f54dbd64f579faab2744308 · Mistral3ForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>mistralai/Ministral-3-3B-Instruct-2512</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Ministral-3-3B-Instruct-2512-mic/tree/444a971de7a3a92574df14dcdd5dde549e0d660f" title="huggingface-mic/Ministral-3-3B-Instruct-2512-mic@444a971de7a3a92574df14dcdd5dde549e0d660f">huggingface-mic/Ministral-3-3B-Instruct-2512-mic@444a971de7a3</a> ← <a href="https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512/tree/b35d4dfe56c142746f54dbd64f579faab2744308" title="mistralai/Ministral-3-3B-Instruct-2512@b35d4dfe56c142746f54dbd64f579faab2744308">mistralai/Ministral-3-3B-Instruct-2512@b35d4dfe56c1</a> · Mistral3ForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">3.8B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.transformer.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: PixtralAttentionLayer"></i><i class="c0" title="layer 1: PixtralAttentionLayer"></i><i class="c0" title="layer 2: PixtralAttentionLayer"></i><i class="c0" title="layer 3: PixtralAttentionLayer"></i><i class="c0" title="layer 4: PixtralAttentionLayer"></i><i class="c0" title="layer 5: PixtralAttentionLayer"></i><i class="c0" title="layer 6: PixtralAttentionLayer"></i><i class="c0" title="layer 7: PixtralAttentionLayer"></i><i class="c0" title="layer 8: PixtralAttentionLayer"></i><i class="c0" title="layer 9: PixtralAttentionLayer"></i><i class="c0" title="layer 10: PixtralAttentionLayer"></i><i class="c0" title="layer 11: PixtralAttentionLayer"></i><i class="c0" title="layer 12: PixtralAttentionLayer"></i><i class="c0" title="layer 13: PixtralAttentionLayer"></i><i class="c0" title="layer 14: PixtralAttentionLayer"></i><i class="c0" title="layer 15: PixtralAttentionLayer"></i><i class="c0" title="layer 16: PixtralAttentionLayer"></i><i class="c0" title="layer 17: PixtralAttentionLayer"></i><i class="c0" title="layer 18: PixtralAttentionLayer"></i><i class="c0" title="layer 19: PixtralAttentionLayer"></i><i class="c0" title="layer 20: PixtralAttentionLayer"></i><i class="c0" title="layer 21: PixtralAttentionLayer"></i><i class="c0" title="layer 22: PixtralAttentionLayer"></i><i class="c0" title="layer 23: PixtralAttentionLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: PixtralAttentionLayer"></i><i class="c0" title="layer 1: PixtralAttentionLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>PixtralAttentionLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 26</span><div class="cells"><i class="c0" title="layer 0: Ministral3DecoderLayer"></i><i class="c0" title="layer 1: Ministral3DecoderLayer"></i><i class="c0" title="layer 2: Ministral3DecoderLayer"></i><i class="c0" title="layer 3: Ministral3DecoderLayer"></i><i class="c0" title="layer 4: Ministral3DecoderLayer"></i><i class="c0" title="layer 5: Ministral3DecoderLayer"></i><i class="c0" title="layer 6: Ministral3DecoderLayer"></i><i class="c0" title="layer 7: Ministral3DecoderLayer"></i><i class="c0" title="layer 8: Ministral3DecoderLayer"></i><i class="c0" title="layer 9: Ministral3DecoderLayer"></i><i class="c0" title="layer 10: Ministral3DecoderLayer"></i><i class="c0" title="layer 11: Ministral3DecoderLayer"></i><i class="c0" title="layer 12: Ministral3DecoderLayer"></i><i class="c0" title="layer 13: Ministral3DecoderLayer"></i><i class="c0" title="layer 14: Ministral3DecoderLayer"></i><i class="c0" title="layer 15: Ministral3DecoderLayer"></i><i class="c0" title="layer 16: Ministral3DecoderLayer"></i><i class="c0" title="layer 17: Ministral3DecoderLayer"></i><i class="c0" title="layer 18: Ministral3DecoderLayer"></i><i class="c0" title="layer 19: Ministral3DecoderLayer"></i><i class="c0" title="layer 20: Ministral3DecoderLayer"></i><i class="c0" title="layer 21: Ministral3DecoderLayer"></i><i class="c0" title="layer 22: Ministral3DecoderLayer"></i><i class="c0" title="layer 23: Ministral3DecoderLayer"></i><i class="c0" title="layer 24: Ministral3DecoderLayer"></i><i class="c0" title="layer 25: Ministral3DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Ministral3DecoderLayer"></i><i class="c0" title="layer 1: Ministral3DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Ministral3DecoderLayer</td><td>26</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.786251" title="model.language_model.layers: 3.0B (78.6%)"></i><i class="c1" style="flex:0.104623" title="model.vision_tower.transformer.layers: 402.7M (10.5%)"></i><i class="c2" style="flex:0.104611" title="model.language_model: 402.7M (10.5%)"></i><i class="c3" style="flex:0.004359" title="model.multi_modal_projector: 16.8M (0.4%)"></i><i class="c4" style="flex:0.000157" title="model.vision_tower: 603.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008596" title="model.language_model.layers: 74.0K (0.9%)"></i><i class="c1" style="flex:0.009547" title="model.vision_tower.transformer.layers: 82.2K (1.0%)"></i><i class="c2" style="flex:0.974615" title="model.language_model: 8.4M (97.5%)"></i><i class="c3" style="flex:0.002863" title="model.multi_modal_projector: 24.6K (0.3%)"></i><i class="c4" style="flex:0.004380" title="model.vision_tower: 37.7K (0.4%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>3.0B</td><td>78.6%</td><td>74.0K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>model.vision_tower.transformer.layers</td><td>402.7M</td><td>10.5%</td><td>82.2K</td><td>1.0%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>402.7M</td><td>10.5%</td><td>8.4M</td><td>97.5%</td></tr><tr><td><i class="c3"></i>model.multi_modal_projector</td><td>16.8M</td><td>0.4%</td><td>24.6K</td><td>0.3%</td></tr><tr><td><i class="c4"></i>model.vision_tower</td><td>603.1K</td><td>0.0%</td><td>37.7K</td><td>0.4%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>mistral3</td><td>mistral3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>pixtral</td><td>pixtral</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>1540</td><td>1540</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>ministral3</td><td>ministral3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Mistral3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>354</td><td>32</td></tr><tr class=""><td>Ministral3RMSNorm</td><td>53</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>50</td><td>4</td></tr><tr class=""><td>PixtralRMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Ministral3Attention</td><td>26</td><td>2</td></tr><tr class=""><td>Ministral3DecoderLayer</td><td>26</td><td>2</td></tr><tr class=""><td>Ministral3MLP</td><td>26</td><td>2</td></tr><tr class=""><td>PixtralAttention</td><td>24</td><td>2</td></tr><tr class=""><td>PixtralAttentionLayer</td><td>24</td><td>2</td></tr><tr class=""><td>PixtralMLP</td><td>24</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>Ministral3Model</td><td>1</td><td>1</td></tr><tr class=""><td>Ministral3RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>PatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>RMSNorm</td><td>1</td><td>1</td></tr><tr class=""><td>PixtralTransformer</td><td>1</td><td>1</td></tr><tr class=""><td>PixtralVisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>PixtralVisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 53 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>quantization_config</td><td>{&quot;activation_scheme&quot;: &quot;static&quot;, &quot;dequantize&quot;: false, &quot;modules_to_not_convert&quot;…</td><td>—</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>3072</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>9216</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>26</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>vision_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>24</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Ministral-3-3B-Instruct-2512-mic.json b/shrink/Ministral-3-3B-Instruct-2512-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..3838c66f87c2dd5171cca96ce09a8a7c3b308051
--- /dev/null
+++ b/shrink/Ministral-3-3B-Instruct-2512-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Ministral-3-3B-Instruct-2512-mic@444a971de7a3a92574df14dcdd5dde549e0d660f", "findings": []}
\ No newline at end of file
diff --git a/shrink/Mistral-7B-Instruct-v0.2-mic.html b/shrink/Mistral-7B-Instruct-v0.2-mic.html
index 1233543a12b96facaed9566e580bbeb4ef95d505..ad9f5457663c4d5adf1f93006777271a41de2267 100644
--- a/shrink/Mistral-7B-Instruct-v0.2-mic.html
+++ b/shrink/Mistral-7B-Instruct-v0.2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>mistralai/Mistral-7B-Instruct-v0.2: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Mistral-7B-Instruct-v0.2-mic@90938461b4cb2497a78ea7f817f89df453996848 ← mistralai/Mistral-7B-Instruct-v0.2@63a8b081895390a26e140280378bc85ec8bce07a · MistralForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>mistralai/Mistral-7B-Instruct-v0.2</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Mistral-7B-Instruct-v0.2-mic/tree/90938461b4cb2497a78ea7f817f89df453996848" title="huggingface-mic/Mistral-7B-Instruct-v0.2-mic@90938461b4cb2497a78ea7f817f89df453996848">huggingface-mic/Mistral-7B-Instruct-v0.2-mic@90938461b4cb</a> ← <a href="https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/tree/63a8b081895390a26e140280378bc85ec8bce07a" title="mistralai/Mistral-7B-Instruct-v0.2@63a8b081895390a26e140280378bc85ec8bce07a">mistralai/Mistral-7B-Instruct-v0.2@63a8b0818953</a> · MistralForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">7.2B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>model: max_position_embeddings 32768 became 512</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.963800" title="model.layers: 7.0B (96.4%)"></i><i class="c1" style="flex:0.018100" title="model.embed_tokens: 131.1M (1.8%)"></i><i class="c2" style="flex:0.018100" title="lm_head: 131.1M (1.8%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.029185" title="model.layers: 246.3K (2.9%)"></i><i class="c1" style="flex:0.485400" title="model.embed_tokens: 4.1M (48.5%)"></i><i class="c2" style="flex:0.485400" title="lm_head: 4.1M (48.5%)"></i><i class="c3" style="flex:0.000015" title="model.norm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>7.0B</td><td>96.4%</td><td>246.3K</td><td>2.9%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>131.1M</td><td>1.8%</td><td>4.1M</td><td>48.5%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>131.1M</td><td>1.8%</td><td>4.1M</td><td>48.5%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>mistral</td><td>mistral</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class="bad"><td>model</td><td>max_position_embeddings</td><td>32768</td><td>512</td><td>⚠ changed</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Mistral omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>225</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 21 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>4096</td><td>128</td></tr><tr><td>intermediate_size</td><td>14336</td><td>256</td></tr><tr><td>max_position_embeddings</td><td>32768</td><td>512</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Mistral-7B-Instruct-v0.2-mic.json b/shrink/Mistral-7B-Instruct-v0.2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a46786157b76cd34fd72b4cea34deadb8d008503
--- /dev/null
+++ b/shrink/Mistral-7B-Instruct-v0.2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Mistral-7B-Instruct-v0.2-mic@90938461b4cb2497a78ea7f817f89df453996848", "findings": ["model: max_position_embeddings 32768 became 512"]}
\ No newline at end of file
diff --git a/shrink/Mixtral-8x7B-Instruct-v0.1-mic.html b/shrink/Mixtral-8x7B-Instruct-v0.1-mic.html
index 7dd8050beb1d5fd1f0caeaebaf5d93df23c146dd..4849d41df7d0c16c8cf5cd28062629fd424e5a99 100644
--- a/shrink/Mixtral-8x7B-Instruct-v0.1-mic.html
+++ b/shrink/Mixtral-8x7B-Instruct-v0.1-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>mistralai/Mixtral-8x7B-Instruct-v0.1: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Mixtral-8x7B-Instruct-v0.1-mic@9c86dd7f51c866a4511616932943f95f6d163e66 ← mistralai/Mixtral-8x7B-Instruct-v0.1@eba92302a2861cdc0098cc54bc9f17cb2c47eb61 · MixtralForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">46.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.994387" title="model.layers: 46.4B (99.4%)"></i><i class="c1" style="flex:0.002807" title="lm_head: 131.1M (0.3%)"></i><i class="c2" style="flex:0.002807" title="model.embed_tokens: 131.1M (0.3%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.051401" title="model.layers: 222.0K (5.1%)"></i><i class="c1" style="flex:0.474292" title="lm_head: 2.0M (47.4%)"></i><i class="c2" style="flex:0.474292" title="model.embed_tokens: 2.0M (47.4%)"></i><i class="c3" style="flex:0.000015" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>46.4B</td><td>99.4%</td><td>222.0K</td><td>5.1%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>131.1M</td><td>0.3%</td><td>2.0M</td><td>47.4%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>131.1M</td><td>0.3%</td><td>2.0M</td><td>47.4%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 8 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>mixtral</td><td>mixtral</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Mixtral omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>129</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>Experts</td><td>32</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>32</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 27 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>14336</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>num_local_experts</td><td>8</td><td>4</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>mistralai/Mixtral-8x7B-Instruct-v0.1</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Mixtral-8x7B-Instruct-v0.1-mic/tree/9c86dd7f51c866a4511616932943f95f6d163e66" title="huggingface-mic/Mixtral-8x7B-Instruct-v0.1-mic@9c86dd7f51c866a4511616932943f95f6d163e66">huggingface-mic/Mixtral-8x7B-Instruct-v0.1-mic@9c86dd7f51c8</a> ← <a href="https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1/tree/eba92302a2861cdc0098cc54bc9f17cb2c47eb61" title="mistralai/Mixtral-8x7B-Instruct-v0.1@eba92302a2861cdc0098cc54bc9f17cb2c47eb61">mistralai/Mixtral-8x7B-Instruct-v0.1@eba92302a286</a> · MixtralForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">46.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.994387" title="model.layers: 46.4B (99.4%)"></i><i class="c1" style="flex:0.002807" title="model.embed_tokens: 131.1M (0.3%)"></i><i class="c2" style="flex:0.002807" title="lm_head: 131.1M (0.3%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.051401" title="model.layers: 222.0K (5.1%)"></i><i class="c1" style="flex:0.474292" title="model.embed_tokens: 2.0M (47.4%)"></i><i class="c2" style="flex:0.474292" title="lm_head: 2.0M (47.4%)"></i><i class="c3" style="flex:0.000015" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>46.4B</td><td>99.4%</td><td>222.0K</td><td>5.1%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>131.1M</td><td>0.3%</td><td>2.0M</td><td>47.4%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>131.1M</td><td>0.3%</td><td>2.0M</td><td>47.4%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 8 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>mixtral</td><td>mixtral</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Mixtral omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>129</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>Experts</td><td>32</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>32</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 27 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>14336</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>num_local_experts</td><td>8</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Mixtral-8x7B-Instruct-v0.1-mic.json b/shrink/Mixtral-8x7B-Instruct-v0.1-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a4e773cd8a6bf9ef94b7021022b1a215da7bfa87
--- /dev/null
+++ b/shrink/Mixtral-8x7B-Instruct-v0.1-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Mixtral-8x7B-Instruct-v0.1-mic@9c86dd7f51c866a4511616932943f95f6d163e66", "findings": []}
\ No newline at end of file
diff --git a/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.html b/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.html
index 60f8b259be959c1d26f30b277721e97c627c9000..6c263dff676e4088b3c59cc38381d26f3b078b23 100644
--- a/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.html
+++ b/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/NVIDIA-Nemotron-3-Nano-4B-BF16-mic@9e2cf8579314aef2d07fb651a0e47989c49ce98d ← nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16@dfaf35de3e30f1867dd8dbc38a7fc9fb52d3914f · NemotronHForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/NVIDIA-Nemotron-3-Nano-4B-BF16-mic/tree/9e2cf8579314aef2d07fb651a0e47989c49ce98d" title="huggingface-mic/NVIDIA-Nemotron-3-Nano-4B-BF16-mic@9e2cf8579314aef2d07fb651a0e47989c49ce98d">huggingface-mic/NVIDIA-Nemotron-3-Nano-4B-BF16-mic@9e2cf8579314</a> ← <a href="https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16/tree/dfaf35de3e30f1867dd8dbc38a7fc9fb52d3914f" title="nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16@dfaf35de3e30f1867dd8dbc38a7fc9fb52d3914f">nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16@dfaf35de3e30</a> · NemotronHForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">49.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 42</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 1: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 2: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 3: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 4: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 5: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 6: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 7: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 8: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 9: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 10: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 11: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 12: is_causal=True, Attention"></i><i class="c1" title="layer 13: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 14: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 15: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 16: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 17: is_causal=True, Attention"></i><i class="c1" title="layer 18: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 19: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 20: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 21: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 22: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 23: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 24: is_causal=True, Attention"></i><i class="c1" title="layer 25: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 26: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 27: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 28: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 29: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 30: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 31: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 32: is_causal=True, Attention"></i><i class="c1" title="layer 33: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 34: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 35: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 36: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 37: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 38: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 39: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 40: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 41: MLP, ReLUSquaredActivation"></i></div></div><div class="row"><span class="lab">Fixture · 42</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 1: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 2: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 3: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 4: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 5: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 6: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 7: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 8: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 9: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 10: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 11: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 12: is_causal=True, Attention"></i><i class="c1" title="layer 13: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 14: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 15: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 16: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 17: is_causal=True, Attention"></i><i class="c1" title="layer 18: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 19: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 20: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 21: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 22: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 23: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 24: is_causal=True, Attention"></i><i class="c1" title="layer 25: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 26: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 27: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 28: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 29: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 30: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 31: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c2" title="layer 32: is_causal=True, Attention"></i><i class="c1" title="layer 33: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 34: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 35: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c0" title="layer 36: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 37: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 38: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 39: MLP, ReLUSquaredActivation"></i><i class="c0" title="layer 40: linear_attention, Conv1d, Mamba2Mixer …"></i><i class="c1" title="layer 41: MLP, ReLUSquaredActivation"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, Mamba2Mixer …</td><td>21</td><td>21</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>MLP, ReLUSquaredActivation</td><td>17</td><td>17</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>is_causal=True, Attention</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.793111" title="model.layers: 3.2B (79.3%)"></i><i class="c1" style="flex:0.103444" title="lm_head: 411.0M (10.3%)"></i><i class="c2" style="flex:0.103444" title="model.embeddings: 411.0M (10.3%)"></i><i class="c3" style="flex:0.000001" title="model.norm_f: 3.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.661387" title="model.layers: 32.8M (66.1%)"></i><i class="c1" style="flex:0.169306" title="lm_head: 8.4M (16.9%)"></i><i class="c2" style="flex:0.169306" title="model.embeddings: 8.4M (16.9%)"></i><i class="c3" style="flex:0.000001" title="model.norm_f: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>3.2B</td><td>79.3%</td><td>32.8M</td><td>66.1%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>411.0M</td><td>10.3%</td><td>8.4M</td><td>16.9%</td></tr><tr><td><i class="c2"></i>model.embeddings</td><td>411.0M</td><td>10.3%</td><td>8.4M</td><td>16.9%</td></tr><tr><td><i class="c3"></i>model.norm_f</td><td>3.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (40 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 8 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>nemotron_h</td><td>nemotron_h</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix NemotronH omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>93</td><td>93</td></tr><tr class=""><td>RMSNorm</td><td>43</td><td>43</td></tr><tr class=""><td>Block</td><td>42</td><td>42</td></tr><tr class=""><td>Conv1d</td><td>21</td><td>21</td></tr><tr class=""><td>Mamba2Mixer</td><td>21</td><td>21</td></tr><tr class=""><td>RMSNormGated</td><td>21</td><td>21</td></tr><tr class=""><td>SiLUActivation</td><td>21</td><td>21</td></tr><tr class=""><td>MLP</td><td>17</td><td>17</td></tr><tr class=""><td>ReLUSquaredActivation</td><td>17</td><td>17</td></tr><tr class=""><td>Attention</td><td>4</td><td>4</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 58 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>chunk_size</td><td>256</td><td>8</td></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>3136</td><td>64</td></tr><tr><td>intermediate_size</td><td>12544</td><td>128</td></tr><tr><td>moe_intermediate_size</td><td>7688</td><td>32</td></tr><tr><td>n_routed_experts</td><td>8</td><td>4</td></tr><tr><td>num_attention_heads</td><td>40</td><td>4</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>ssm_state_size</td><td>128</td><td>16</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.json b/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..40faf88cb6bb3b566d36fc51a76437aa95c5b785
--- /dev/null
+++ b/shrink/NVIDIA-Nemotron-3-Nano-4B-BF16-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/NVIDIA-Nemotron-3-Nano-4B-BF16-mic@9e2cf8579314aef2d07fb651a0e47989c49ce98d", "findings": []}
\ No newline at end of file
diff --git a/shrink/Nemotron-3-Embed-1B-BF16-mic.html b/shrink/Nemotron-3-Embed-1B-BF16-mic.html
index 6f6adaadb9b384ef5d36477d5c7fbd666980c7d8..f05b1f9c89298ea488da1b015afe7f11dacd03d0 100644
--- a/shrink/Nemotron-3-Embed-1B-BF16-mic.html
+++ b/shrink/Nemotron-3-Embed-1B-BF16-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>nvidia/Nemotron-3-Embed-1B-BF16: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Nemotron-3-Embed-1B-BF16-mic@645864f9759d320c610786a2270ecb18e982534a ← nvidia/Nemotron-3-Embed-1B-BF16@c0c9fea93ea424587517f2c59e20db9f1d6bf615 · Ministral3Model · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>nvidia/Nemotron-3-Embed-1B-BF16</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Nemotron-3-Embed-1B-BF16-mic/tree/645864f9759d320c610786a2270ecb18e982534a" title="huggingface-mic/Nemotron-3-Embed-1B-BF16-mic@645864f9759d320c610786a2270ecb18e982534a">huggingface-mic/Nemotron-3-Embed-1B-BF16-mic@645864f9759d</a> ← <a href="https://huggingface.co/nvidia/Nemotron-3-Embed-1B-BF16/tree/c0c9fea93ea424587517f2c59e20db9f1d6bf615" title="nvidia/Nemotron-3-Embed-1B-BF16@c0c9fea93ea424587517f2c59e20db9f1d6bf615">nvidia/Nemotron-3-Embed-1B-BF16@c0c9fea93ea4</a> · Ministral3Model · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.764718" title="layers: 872.5M (76.5%)"></i><i class="c1" style="flex:0.235280" title="embed_tokens: 268.4M (23.5%)"></i><i class="c2" style="flex:0.000002" title="norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008742" title="layers: 74.0K (0.9%)"></i><i class="c1" style="flex:0.991250" title="embed_tokens: 8.4M (99.1%)"></i><i class="c2" style="flex:0.000008" title="norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>layers</td><td>872.5M</td><td>76.5%</td><td>74.0K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>embed_tokens</td><td>268.4M</td><td>23.5%</td><td>8.4M</td><td>99.1%</td></tr><tr><td><i class="c2"></i>norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (24 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>ministral3</td><td>ministral3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Ministral3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>112</td><td>14</td></tr><tr class=""><td>RMSNorm</td><td>33</td><td>5</td></tr><tr class=""><td>Attention</td><td>16</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>MLP</td><td>16</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>6144</td><td>128</td></tr><tr><td>num_attention_heads</td><td>24</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Nemotron-3-Embed-1B-BF16-mic.json b/shrink/Nemotron-3-Embed-1B-BF16-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..1a5f6984cb7096711be1d13bdb313cf5b672ad82
--- /dev/null
+++ b/shrink/Nemotron-3-Embed-1B-BF16-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Nemotron-3-Embed-1B-BF16-mic@645864f9759d320c610786a2270ecb18e982534a", "findings": []}
\ No newline at end of file
diff --git a/shrink/OLMo-1B-hf-mic.html b/shrink/OLMo-1B-hf-mic.html
index 629d5d1a54013f4a5fdd744195b2e53bbedc9f61..0027b39c8c511290139832d07d611a0e9fb70c1e 100644
--- a/shrink/OLMo-1B-hf-mic.html
+++ b/shrink/OLMo-1B-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>allenai/OLMo-1B-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/OLMo-1B-hf-mic@87c81bd7d9f93663d9906bbde0737125e4d85c54 ← allenai/OLMo-1B-hf@aee7752d9c08ee4775e9b0091426d8410e8f6a89 · OlmoForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>allenai/OLMo-1B-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/OLMo-1B-hf-mic/tree/87c81bd7d9f93663d9906bbde0737125e4d85c54" title="huggingface-mic/OLMo-1B-hf-mic@87c81bd7d9f93663d9906bbde0737125e4d85c54">huggingface-mic/OLMo-1B-hf-mic@87c81bd7d9f9</a> ← <a href="https://huggingface.co/allenai/OLMo-1B-hf/tree/aee7752d9c08ee4775e9b0091426d8410e8f6a89" title="allenai/OLMo-1B-hf@aee7752d9c08ee4775e9b0091426d8410e8f6a89">allenai/OLMo-1B-hf@aee7752d9c08</a> · OlmoForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.2B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.912453" title="model.layers: 1.1B (91.2%)"></i><i class="c1" style="flex:0.087547" title="model.embed_tokens: 103.0M (8.8%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024814" title="model.layers: 81.9K (2.5%)"></i><i class="c1" style="flex:0.975186" title="model.embed_tokens: 3.2M (97.5%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>1.1B</td><td>91.2%</td><td>81.9K</td><td>2.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>103.0M</td><td>8.8%</td><td>3.2M</td><td>97.5%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmo</td><td>olmo</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50304</td><td>50304</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmo omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>113</td><td>15</td></tr><tr class=""><td>LayerNorm</td><td>33</td><td>5</td></tr><tr class=""><td>Attention</td><td>16</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>MLP</td><td>16</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 23 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/OLMo-1B-hf-mic.json b/shrink/OLMo-1B-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..293956bdee8c3b8605bfeaf4b6f56a547ebc86a7
--- /dev/null
+++ b/shrink/OLMo-1B-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/OLMo-1B-hf-mic@87c81bd7d9f93663d9906bbde0737125e4d85c54", "findings": []}
\ No newline at end of file
diff --git a/shrink/OLMo-2-0425-1B-mic.html b/shrink/OLMo-2-0425-1B-mic.html
index 30f3e4376c6fb1219c1719d9d8feca661d30cc59..5c32ed86c004b31b33453133d852f0197cf1439d 100644
--- a/shrink/OLMo-2-0425-1B-mic.html
+++ b/shrink/OLMo-2-0425-1B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>allenai/OLMo-2-0425-1B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/OLMo-2-0425-1B-mic@7a6eea1e741066708957df80693b23000613b907 ← allenai/OLMo-2-0425-1B@a1847dff35000b4271fa70afc5db10fd29fedbdf · Olmo2ForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.5B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">12.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.723187" title="model.layers: 1.1B (72.3%)"></i><i class="c1" style="flex:0.138406" title="model.embed_tokens: 205.5M (13.8%)"></i><i class="c2" style="flex:0.138406" title="lm_head: 205.5M (13.8%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006376" title="model.layers: 82.4K (0.6%)"></i><i class="c1" style="flex:0.496809" title="model.embed_tokens: 6.4M (49.7%)"></i><i class="c2" style="flex:0.496809" title="lm_head: 6.4M (49.7%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>1.1B</td><td>72.3%</td><td>82.4K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>205.5M</td><td>13.8%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>205.5M</td><td>13.8%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmo2</td><td>olmo2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmo2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>113</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>9</td></tr><tr class=""><td>Attention</td><td>16</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>MLP</td><td>16</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 22 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>allenai/OLMo-2-0425-1B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/OLMo-2-0425-1B-mic/tree/7a6eea1e741066708957df80693b23000613b907" title="huggingface-mic/OLMo-2-0425-1B-mic@7a6eea1e741066708957df80693b23000613b907">huggingface-mic/OLMo-2-0425-1B-mic@7a6eea1e7410</a> ← <a href="https://huggingface.co/allenai/OLMo-2-0425-1B/tree/a1847dff35000b4271fa70afc5db10fd29fedbdf" title="allenai/OLMo-2-0425-1B@a1847dff35000b4271fa70afc5db10fd29fedbdf">allenai/OLMo-2-0425-1B@a1847dff3500</a> · Olmo2ForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.5B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">12.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.723187" title="model.layers: 1.1B (72.3%)"></i><i class="c1" style="flex:0.138406" title="lm_head: 205.5M (13.8%)"></i><i class="c2" style="flex:0.138406" title="model.embed_tokens: 205.5M (13.8%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006376" title="model.layers: 82.4K (0.6%)"></i><i class="c1" style="flex:0.496809" title="lm_head: 6.4M (49.7%)"></i><i class="c2" style="flex:0.496809" title="model.embed_tokens: 6.4M (49.7%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>1.1B</td><td>72.3%</td><td>82.4K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>205.5M</td><td>13.8%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>205.5M</td><td>13.8%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmo2</td><td>olmo2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmo2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>113</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>9</td></tr><tr class=""><td>Attention</td><td>16</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>MLP</td><td>16</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 22 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/OLMo-2-0425-1B-mic.json b/shrink/OLMo-2-0425-1B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..7e6c4bae06a36bbc1f9da5d933d8cfc7f2b52815
--- /dev/null
+++ b/shrink/OLMo-2-0425-1B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/OLMo-2-0425-1B-mic@7a6eea1e741066708957df80693b23000613b907", "findings": []}
\ No newline at end of file
diff --git a/shrink/OLMoE-1B-7B-0924-mic.html b/shrink/OLMoE-1B-7B-0924-mic.html
index b7697681bb69d4e281ce6de6e45207a3d6973324..93216b20e7300482907faa302cea26706a767095 100644
--- a/shrink/OLMoE-1B-7B-0924-mic.html
+++ b/shrink/OLMoE-1B-7B-0924-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>allenai/OLMoE-1B-7B-0924: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/OLMoE-1B-7B-0924-mic@abd7fb026a08bdb27dfa8ff4d6a1680721f1e85a ← allenai/OLMoE-1B-7B-0924@6d84c48581ece794365f2b8e9cfb043c68ade9c5 · OlmoeForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">6.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">6.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.970221" title="model.layers: 6.7B (97.0%)"></i><i class="c1" style="flex:0.014889" title="model.embed_tokens: 103.0M (1.5%)"></i><i class="c2" style="flex:0.014889" title="lm_head: 103.0M (1.5%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.034546" title="model.layers: 230.4K (3.5%)"></i><i class="c1" style="flex:0.482722" title="model.embed_tokens: 3.2M (48.3%)"></i><i class="c2" style="flex:0.482722" title="lm_head: 3.2M (48.3%)"></i><i class="c3" style="flex:0.000010" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>6.7B</td><td>97.0%</td><td>230.4K</td><td>3.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>103.0M</td><td>1.5%</td><td>3.2M</td><td>48.3%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>103.0M</td><td>1.5%</td><td>3.2M</td><td>48.3%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmoe</td><td>olmoe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50304</td><td>50304</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmoe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>65</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>9</td></tr><tr class=""><td>Attention</td><td>16</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>Experts</td><td>16</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>16</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>16</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>1024</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>allenai/OLMoE-1B-7B-0924</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/OLMoE-1B-7B-0924-mic/tree/abd7fb026a08bdb27dfa8ff4d6a1680721f1e85a" title="huggingface-mic/OLMoE-1B-7B-0924-mic@abd7fb026a08bdb27dfa8ff4d6a1680721f1e85a">huggingface-mic/OLMoE-1B-7B-0924-mic@abd7fb026a08</a> ← <a href="https://huggingface.co/allenai/OLMoE-1B-7B-0924/tree/6d84c48581ece794365f2b8e9cfb043c68ade9c5" title="allenai/OLMoE-1B-7B-0924@6d84c48581ece794365f2b8e9cfb043c68ade9c5">allenai/OLMoE-1B-7B-0924@6d84c48581ec</a> · OlmoeForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">6.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">6.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.970221" title="model.layers: 6.7B (97.0%)"></i><i class="c1" style="flex:0.014889" title="lm_head: 103.0M (1.5%)"></i><i class="c2" style="flex:0.014889" title="model.embed_tokens: 103.0M (1.5%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.034546" title="model.layers: 230.4K (3.5%)"></i><i class="c1" style="flex:0.482722" title="lm_head: 3.2M (48.3%)"></i><i class="c2" style="flex:0.482722" title="model.embed_tokens: 3.2M (48.3%)"></i><i class="c3" style="flex:0.000010" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>6.7B</td><td>97.0%</td><td>230.4K</td><td>3.5%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>103.0M</td><td>1.5%</td><td>3.2M</td><td>48.3%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>103.0M</td><td>1.5%</td><td>3.2M</td><td>48.3%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmoe</td><td>olmoe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50304</td><td>50304</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmoe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>65</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>9</td></tr><tr class=""><td>Attention</td><td>16</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>Experts</td><td>16</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>16</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>16</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>1024</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/OLMoE-1B-7B-0924-mic.json b/shrink/OLMoE-1B-7B-0924-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..02049285526f23e2050cfa358c1468e6fcb51f8a
--- /dev/null
+++ b/shrink/OLMoE-1B-7B-0924-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/OLMoE-1B-7B-0924-mic@abd7fb026a08bdb27dfa8ff4d6a1680721f1e85a", "findings": []}
\ No newline at end of file
diff --git a/shrink/Olmo-3-7B-Instruct-mic.html b/shrink/Olmo-3-7B-Instruct-mic.html
index 5867734cfec34930f34907058694e15ea6e0e110..8557b0ef96b1d88184029ad8e51df708e02abc00 100644
--- a/shrink/Olmo-3-7B-Instruct-mic.html
+++ b/shrink/Olmo-3-7B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>allenai/Olmo-3-7B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Olmo-3-7B-Instruct-mic@3d2269c8c387b9dec2f5a987d2739464ad990224 ← allenai/Olmo-3-7B-Instruct@6e5971d9eba42665f5bd5a0fcf047f299ce1dccc · Olmo3ForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">7.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">12.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c0" title="layer 1: sliding_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c0" title="layer 5: sliding_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c1" title="layer 7: full_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c0" title="layer 9: sliding_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c0" title="layer 13: sliding_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c1" title="layer 15: full_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c0" title="layer 17: sliding_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c0" title="layer 21: sliding_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i><i class="c0" title="layer 24: sliding_attention"></i><i class="c0" title="layer 25: sliding_attention"></i><i class="c0" title="layer 26: sliding_attention"></i><i class="c1" title="layer 27: full_attention"></i><i class="c0" title="layer 28: sliding_attention"></i><i class="c0" title="layer 29: sliding_attention"></i><i class="c0" title="layer 30: sliding_attention"></i><i class="c1" title="layer 31: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>24</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>8</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.887438" title="model.layers: 6.5B (88.7%)"></i><i class="c1" style="flex:0.056281" title="model.embed_tokens: 410.7M (5.6%)"></i><i class="c2" style="flex:0.056281" title="lm_head: 410.7M (5.6%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006381" title="model.layers: 82.4K (0.6%)"></i><i class="c1" style="flex:0.496807" title="model.embed_tokens: 6.4M (49.7%)"></i><i class="c2" style="flex:0.496807" title="lm_head: 6.4M (49.7%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>6.5B</td><td>88.7%</td><td>82.4K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>410.7M</td><td>5.6%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>410.7M</td><td>5.6%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (32 heads, 32 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmo3</td><td>olmo3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100278</td><td>100278</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>65536</td><td>65536</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmo3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>225</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>129</td><td>9</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 23 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>11008</td><td>128</td></tr><tr><td>layer_types</td><td>32 entries: 24× sliding_attention, 8× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>32</td><td>4</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>allenai/Olmo-3-7B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Olmo-3-7B-Instruct-mic/tree/3d2269c8c387b9dec2f5a987d2739464ad990224" title="huggingface-mic/Olmo-3-7B-Instruct-mic@3d2269c8c387b9dec2f5a987d2739464ad990224">huggingface-mic/Olmo-3-7B-Instruct-mic@3d2269c8c387</a> ← <a href="https://huggingface.co/allenai/Olmo-3-7B-Instruct/tree/6e5971d9eba42665f5bd5a0fcf047f299ce1dccc" title="allenai/Olmo-3-7B-Instruct@6e5971d9eba42665f5bd5a0fcf047f299ce1dccc">allenai/Olmo-3-7B-Instruct@6e5971d9eba4</a> · Olmo3ForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">7.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">12.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c0" title="layer 1: sliding_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c0" title="layer 5: sliding_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c1" title="layer 7: full_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c0" title="layer 9: sliding_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c0" title="layer 13: sliding_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c1" title="layer 15: full_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c0" title="layer 17: sliding_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c0" title="layer 21: sliding_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i><i class="c0" title="layer 24: sliding_attention"></i><i class="c0" title="layer 25: sliding_attention"></i><i class="c0" title="layer 26: sliding_attention"></i><i class="c1" title="layer 27: full_attention"></i><i class="c0" title="layer 28: sliding_attention"></i><i class="c0" title="layer 29: sliding_attention"></i><i class="c0" title="layer 30: sliding_attention"></i><i class="c1" title="layer 31: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>24</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>8</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.887438" title="model.layers: 6.5B (88.7%)"></i><i class="c1" style="flex:0.056281" title="lm_head: 410.7M (5.6%)"></i><i class="c2" style="flex:0.056281" title="model.embed_tokens: 410.7M (5.6%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006381" title="model.layers: 82.4K (0.6%)"></i><i class="c1" style="flex:0.496807" title="lm_head: 6.4M (49.7%)"></i><i class="c2" style="flex:0.496807" title="model.embed_tokens: 6.4M (49.7%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>6.5B</td><td>88.7%</td><td>82.4K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>410.7M</td><td>5.6%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>410.7M</td><td>5.6%</td><td>6.4M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (32 heads, 32 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>olmo3</td><td>olmo3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100278</td><td>100278</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>65536</td><td>65536</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Olmo3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>225</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>129</td><td>9</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 23 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>11008</td><td>128</td></tr><tr><td>layer_types</td><td>32 entries: 24× sliding_attention, 8× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>32</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Olmo-3-7B-Instruct-mic.json b/shrink/Olmo-3-7B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..01f18cd680088c0142bc59153750d272d57f496a
--- /dev/null
+++ b/shrink/Olmo-3-7B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Olmo-3-7B-Instruct-mic@3d2269c8c387b9dec2f5a987d2739464ad990224", "findings": []}
\ No newline at end of file
diff --git a/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.html b/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.html
index 14ec08c9b3ba91deac9a1d00dce9ff466cfa9780..ab772a1adf8999d809aba69587cab90e039cc077 100644
--- a/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.html
+++ b/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>inference-optimization/Phi-3.5-MoE-0.8B-A0.2B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Phi-3.5-MoE-0.8B-A0.2B-mic@38df14197e6a9ed50ccfe370498b9764d8dcadda ← inference-optimization/Phi-3.5-MoE-0.8B-A0.2B@f6258a2026894c90bffd3c0cfb11c42490935bb8 · PhimoeForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>inference-optimization/Phi-3.5-MoE-0.8B-A0.2B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Phi-3.5-MoE-0.8B-A0.2B-mic/tree/38df14197e6a9ed50ccfe370498b9764d8dcadda" title="huggingface-mic/Phi-3.5-MoE-0.8B-A0.2B-mic@38df14197e6a9ed50ccfe370498b9764d8dcadda">huggingface-mic/Phi-3.5-MoE-0.8B-A0.2B-mic@38df14197e6a</a> ← <a href="https://huggingface.co/inference-optimization/Phi-3.5-MoE-0.8B-A0.2B/tree/f6258a2026894c90bffd3c0cfb11c42490935bb8" title="inference-optimization/Phi-3.5-MoE-0.8B-A0.2B@f6258a2026894c90bffd3c0cfb11c42490935bb8">inference-optimization/Phi-3.5-MoE-0.8B-A0.2B@f6258a202689</a> · PhimoeForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">802.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>4</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.836315" title="model.layers: 671.2M (83.6%)"></i><i class="c1" style="flex:0.081860" title="lm_head: 65.7M (8.2%)"></i><i class="c2" style="flex:0.081820" title="model.embed_tokens: 65.7M (8.2%)"></i><i class="c3" style="flex:0.000005" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.051065" title="model.layers: 222.6K (5.1%)"></i><i class="c1" style="flex:0.478131" title="lm_head: 2.1M (47.8%)"></i><i class="c2" style="flex:0.470775" title="model.embed_tokens: 2.1M (47.1%)"></i><i class="c3" style="flex:0.000029" title="model.norm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>671.2M</td><td>83.6%</td><td>222.6K</td><td>5.1%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>65.7M</td><td>8.2%</td><td>2.1M</td><td>47.8%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>65.7M</td><td>8.2%</td><td>2.1M</td><td>47.1%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 8 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>longrope</td><td>longrope</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>phimoe</td><td>phimoe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32064</td><td>32064</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Phimoe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>17</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>9</td><td>5</td></tr><tr class=""><td>Attention</td><td>4</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>4</td><td>2</td></tr><tr class=""><td>Experts</td><td>4</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>4</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>4</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>4</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>3200</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>4</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr><tr><td>num_local_experts</td><td>8</td><td>4</td></tr><tr><td>rope_parameters</td><td>{&quot;long_factor&quot;: [1.0199999809265137, 1.0299999713897705, 1.0399999618530273, …</td><td>{&quot;long_factor&quot;: [1.0199999809265137, 1.059999942779541, 1.4899998903274536, 1…</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.json b/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..b8692da1cf7472334710d5930aa19a6f54c68fce
--- /dev/null
+++ b/shrink/Phi-3.5-MoE-0.8B-A0.2B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Phi-3.5-MoE-0.8B-A0.2B-mic@38df14197e6a9ed50ccfe370498b9764d8dcadda", "findings": []}
\ No newline at end of file
diff --git a/shrink/Phi-3.5-mini-instruct-mic.html b/shrink/Phi-3.5-mini-instruct-mic.html
index 097a62a59c1293d59b9008a2e4351a1bb461971d..2c8284252b60567571825d9030f6a78d5fa3a8b0 100644
--- a/shrink/Phi-3.5-mini-instruct-mic.html
+++ b/shrink/Phi-3.5-mini-instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>microsoft/Phi-3.5-mini-instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Phi-3.5-mini-instruct-mic@0c7964e6c77b965e4cccc04ccb4f4d4a20a9ff81 ← microsoft/Phi-3.5-mini-instruct@2fe192450127e6a83f7441aef6e3ca586c338b77 · Phi3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>microsoft/Phi-3.5-mini-instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Phi-3.5-mini-instruct-mic/tree/0c7964e6c77b965e4cccc04ccb4f4d4a20a9ff81" title="huggingface-mic/Phi-3.5-mini-instruct-mic@0c7964e6c77b965e4cccc04ccb4f4d4a20a9ff81">huggingface-mic/Phi-3.5-mini-instruct-mic@0c7964e6c77b</a> ← <a href="https://huggingface.co/microsoft/Phi-3.5-mini-instruct/tree/2fe192450127e6a83f7441aef6e3ca586c338b77" title="microsoft/Phi-3.5-mini-instruct@2fe192450127e6a83f7441aef6e3ca586c338b77">microsoft/Phi-3.5-mini-instruct@2fe192450127</a> · Phi3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">3.8B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.948443" title="model.layers: 3.6B (94.8%)"></i><i class="c1" style="flex:0.025778" title="lm_head: 98.5M (2.6%)"></i><i class="c2" style="flex:0.025778" title="model.embed_tokens: 98.5M (2.6%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 3.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.019629" title="model.layers: 82.2K (2.0%)"></i><i class="c1" style="flex:0.490178" title="lm_head: 2.1M (49.0%)"></i><i class="c2" style="flex:0.490178" title="model.embed_tokens: 2.1M (49.0%)"></i><i class="c3" style="flex:0.000015" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>3.6B</td><td>94.8%</td><td>82.2K</td><td>2.0%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>98.5M</td><td>2.6%</td><td>2.1M</td><td>49.0%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>98.5M</td><td>2.6%</td><td>2.1M</td><td>49.0%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>3.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (32 heads, 32 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>longrope</td><td>longrope</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>phi3</td><td>phi3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32064</td><td>32064</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Phi3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>129</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>Dropout</td><td>64</td><td>4</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>3072</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>32</td><td>4</td></tr><tr><td>rope_parameters</td><td>{&quot;long_factor&quot;: [1.0800000429153442, 1.1100000143051147, 1.1399999856948853, …</td><td>{&quot;long_factor&quot;: [1.0800000429153442, 1.6200000047683716, 7.700000286102295, 3…</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Phi-3.5-mini-instruct-mic.json b/shrink/Phi-3.5-mini-instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..ef0f6d48ed2692c2eb08d507dafaba149c3e830e
--- /dev/null
+++ b/shrink/Phi-3.5-mini-instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Phi-3.5-mini-instruct-mic@0c7964e6c77b965e4cccc04ccb4f4d4a20a9ff81", "findings": []}
\ No newline at end of file
diff --git a/shrink/Phi-4-multimodal-instruct-mic.html b/shrink/Phi-4-multimodal-instruct-mic.html
index a797629e9b7e17d67e6bff65cd85c596301b627e..322f85b14af96a727d7144cf9e1af3d55377f2dd 100644
--- a/shrink/Phi-4-multimodal-instruct-mic.html
+++ b/shrink/Phi-4-multimodal-instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>microsoft/Phi-4-multimodal-instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Phi-4-multimodal-instruct-mic@721c204bb9a6c7113710f11eb53f6798c3b87039 ← microsoft/Phi-4-multimodal-instruct@b5b5a5f60533515297d0e14ece86ed80d4b13b19 · Phi4MultimodalForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>microsoft/Phi-4-multimodal-instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Phi-4-multimodal-instruct-mic/tree/721c204bb9a6c7113710f11eb53f6798c3b87039" title="huggingface-mic/Phi-4-multimodal-instruct-mic@721c204bb9a6c7113710f11eb53f6798c3b87039">huggingface-mic/Phi-4-multimodal-instruct-mic@721c204bb9a6</a> ← <a href="https://huggingface.co/microsoft/Phi-4-multimodal-instruct/tree/b5b5a5f60533515297d0e14ece86ed80d4b13b19" title="microsoft/Phi-4-multimodal-instruct@b5b5a5f60533515297d0e14ece86ed80d4b13b19">microsoft/Phi-4-multimodal-instruct@b5b5a5f60533</a> · Phi4MultimodalForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.embed_tokens_extend.image_embed.img_processor.encoder.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i><i class="c0" title="layer 8: VisionEncoderLayer"></i><i class="c0" title="layer 9: VisionEncoderLayer"></i><i class="c0" title="layer 10: VisionEncoderLayer"></i><i class="c0" title="layer 11: VisionEncoderLayer"></i><i class="c0" title="layer 12: VisionEncoderLayer"></i><i class="c0" title="layer 13: VisionEncoderLayer"></i><i class="c0" title="layer 14: VisionEncoderLayer"></i><i class="c0" title="layer 15: VisionEncoderLayer"></i><i class="c0" title="layer 16: VisionEncoderLayer"></i><i class="c0" title="layer 17: VisionEncoderLayer"></i><i class="c0" title="layer 18: VisionEncoderLayer"></i><i class="c0" title="layer 19: VisionEncoderLayer"></i><i class="c0" title="layer 20: VisionEncoderLayer"></i><i class="c0" title="layer 21: VisionEncoderLayer"></i><i class="c0" title="layer 22: VisionEncoderLayer"></i><i class="c0" title="layer 23: VisionEncoderLayer"></i><i class="c0" title="layer 24: VisionEncoderLayer"></i><i class="c0" title="layer 25: VisionEncoderLayer"></i><i class="c0" title="layer 26: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.embed_tokens_extend.audio_embed.encoder.encoders</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: AudioConformerEncoderLayer"></i><i class="c0" title="layer 1: AudioConformerEncoderLayer"></i><i class="c0" title="layer 2: AudioConformerEncoderLayer"></i><i class="c0" title="layer 3: AudioConformerEncoderLayer"></i><i class="c0" title="layer 4: AudioConformerEncoderLayer"></i><i class="c0" title="layer 5: AudioConformerEncoderLayer"></i><i class="c0" title="layer 6: AudioConformerEncoderLayer"></i><i class="c0" title="layer 7: AudioConformerEncoderLayer"></i><i class="c0" title="layer 8: AudioConformerEncoderLayer"></i><i class="c0" title="layer 9: AudioConformerEncoderLayer"></i><i class="c0" title="layer 10: AudioConformerEncoderLayer"></i><i class="c0" title="layer 11: AudioConformerEncoderLayer"></i><i class="c0" title="layer 12: AudioConformerEncoderLayer"></i><i class="c0" title="layer 13: AudioConformerEncoderLayer"></i><i class="c0" title="layer 14: AudioConformerEncoderLayer"></i><i class="c0" title="layer 15: AudioConformerEncoderLayer"></i><i class="c0" title="layer 16: AudioConformerEncoderLayer"></i><i class="c0" title="layer 17: AudioConformerEncoderLayer"></i><i class="c0" title="layer 18: AudioConformerEncoderLayer"></i><i class="c0" title="layer 19: AudioConformerEncoderLayer"></i><i class="c0" title="layer 20: AudioConformerEncoderLayer"></i><i class="c0" title="layer 21: AudioConformerEncoderLayer"></i><i class="c0" title="layer 22: AudioConformerEncoderLayer"></i><i class="c0" title="layer 23: AudioConformerEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: AudioConformerEncoderLayer"></i><i class="c0" title="layer 1: AudioConformerEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioConformerEncoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.679054" title="model.layers: 3.2B (67.9%)"></i><i class="c1" style="flex:0.129553" title="model.embed_tokens: 614.6M (13.0%)"></i><i class="c2" style="flex:0.090347" title="model.embed_tokens_extend.audio_embed.encoder.encoders: 428.6M (9.0%)"></i><i class="c3" style="flex:0.086734" title="model.embed_tokens_extend.image_embed.img_processor.encoder.layers: 411.5M (8.7%)"></i><i class="c4" style="flex:0.014312" title="model.embed_tokens_extend: 67.9M (1.4%)"></i><i class="c5" style="flex:0.000001" title="model.norm: 3.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004466" title="model.layers: 74.0K (0.4%)"></i><i class="c1" style="flex:0.772843" title="model.embed_tokens: 12.8M (77.3%)"></i><i class="c2" style="flex:0.040515" title="model.embed_tokens_extend.audio_embed.encoder.encoders: 671.2K (4.1%)"></i><i class="c3" style="flex:0.004041" title="model.embed_tokens_extend.image_embed.img_processor.encoder.layers: 66.9K (0.4%)"></i><i class="c4" style="flex:0.178132" title="model.embed_tokens_extend: 3.0M (17.8%)"></i><i class="c5" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>3.2B</td><td>67.9%</td><td>74.0K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>614.6M</td><td>13.0%</td><td>12.8M</td><td>77.3%</td></tr><tr><td><i class="c2"></i>model.embed_tokens_extend.audio_embed.encoder.encoders</td><td>428.6M</td><td>9.0%</td><td>671.2K</td><td>4.1%</td></tr><tr><td><i class="c3"></i>model.embed_tokens_extend.image_embed.img_processor.encoder.layers</td><td>411.5M</td><td>8.7%</td><td>66.9K</td><td>0.4%</td></tr><tr><td><i class="c4"></i>model.embed_tokens_extend</td><td>67.9M</td><td>1.4%</td><td>3.0M</td><td>17.8%</td></tr><tr><td><i class="c5"></i>model.norm</td><td>3.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (24 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>longrope</td><td>longrope</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>phi4_multimodal</td><td>phi4_multimodal</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>200064</td><td>200064</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>phi4_multimodal_vision</td><td>phi4_multimodal_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>448</td><td>448</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>phi4_multimodal_audio</td><td>phi4_multimodal_audio</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Phi4Multimodal omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>492</td><td>46</td></tr><tr class=""><td>LayerNorm</td><td>176</td><td>16</td></tr><tr class=""><td>Dropout</td><td>139</td><td>13</td></tr><tr class=""><td>Conv1d</td><td>96</td><td>8</td></tr><tr class=""><td>SiLU</td><td>96</td><td>8</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>AudioMLP</td><td>48</td><td>4</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>28</td><td>3</td></tr><tr class=""><td>VisionMLP</td><td>28</td><td>3</td></tr><tr class=""><td>VisionAttention</td><td>27</td><td>2</td></tr><tr class=""><td>VisionEncoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>AudioAttention</td><td>24</td><td>2</td></tr><tr class=""><td>AudioConformerEncoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>AudioConvModule</td><td>24</td><td>2</td></tr><tr class=""><td>AudioDepthWiseSeparableConv1d</td><td>24</td><td>2</td></tr><tr class=""><td>AudioGluPointWiseConv</td><td>24</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>6</td><td>6</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>AvgPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>MultiheadAttention</td><td>1</td><td>1</td></tr><tr class=""><td>NonDynamicallyQuantizableLinear</td><td>1</td><td>1</td></tr><tr class=""><td>AudioEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>AudioMeanVarianceNormLayer</td><td>1</td><td>1</td></tr><tr class=""><td>AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>AudioNemoConvSubsampling</td><td>1</td><td>1</td></tr><tr class=""><td>AudioRelativeAttentionBias</td><td>1</td><td>1</td></tr><tr class=""><td>FeatureEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>ImageEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionMultiheadAttentionPoolingHead</td><td>1</td><td>1</td></tr><tr class=""><td>ReLU</td><td>1</td><td>1</td></tr><tr class=""><td>Sequential</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">15 fields changed, 78 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.chunk_size</td><td>-1</td><td>8</td></tr><tr><td>audio_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>audio_config.intermediate_size</td><td>1536</td><td>128</td></tr><tr><td>audio_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>audio_config.num_blocks</td><td>24</td><td>2</td></tr><tr><td>hidden_size</td><td>3072</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>24</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>rope_parameters</td><td>{&quot;long_factor&quot;: [1, 1.118320672, 1.250641126, 1.398617824, 1.564103225, 1.749…</td><td>{&quot;long_factor&quot;: [1, 2.446418898, 5.984965424, 14.64173252, 32.1, 32.71], &quot;ori…</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>27</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Phi-4-multimodal-instruct-mic.json b/shrink/Phi-4-multimodal-instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..089e2ee303465c2276e8c40e5f1c0b8a94f7b353
--- /dev/null
+++ b/shrink/Phi-4-multimodal-instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Phi-4-multimodal-instruct-mic@721c204bb9a6c7113710f11eb53f6798c3b87039", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen1.5-MoE-A2.7B-mic.html b/shrink/Qwen1.5-MoE-A2.7B-mic.html
index feed41be511ef8ba6219f9d10e3106efbec6abad..a64b80e1a8cada5c2644dfc4a847b4723efa2560 100644
--- a/shrink/Qwen1.5-MoE-A2.7B-mic.html
+++ b/shrink/Qwen1.5-MoE-A2.7B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen1.5-MoE-A2.7B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen1.5-MoE-A2.7B-mic@2fd4813071f2c23b77666d8c796ff80235d42b55 ← Qwen/Qwen1.5-MoE-A2.7B@1a758c50ecb6350748b9ce0a99d2352fd9fc11c9 · Qwen2MoeForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen1.5-MoE-A2.7B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen1.5-MoE-A2.7B-mic/tree/2fd4813071f2c23b77666d8c796ff80235d42b55" title="huggingface-mic/Qwen1.5-MoE-A2.7B-mic@2fd4813071f2c23b77666d8c796ff80235d42b55">huggingface-mic/Qwen1.5-MoE-A2.7B-mic@2fd4813071f2</a> ← <a href="https://huggingface.co/Qwen/Qwen1.5-MoE-A2.7B/tree/1a758c50ecb6350748b9ce0a99d2352fd9fc11c9" title="Qwen/Qwen1.5-MoE-A2.7B@1a758c50ecb6350748b9ce0a99d2352fd9fc11c9">Qwen/Qwen1.5-MoE-A2.7B@1a758c50ecb6</a> · Qwen2MoeForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">14.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.956528" title="model.layers: 13.7B (95.7%)"></i><i class="c1" style="flex:0.021736" title="model.embed_tokens: 311.2M (2.2%)"></i><i class="c2" style="flex:0.021736" title="lm_head: 311.2M (2.2%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006759" title="model.layers: 132.4K (0.7%)"></i><i class="c1" style="flex:0.496619" title="model.embed_tokens: 9.7M (49.7%)"></i><i class="c2" style="flex:0.496619" title="lm_head: 9.7M (49.7%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>13.7B</td><td>95.7%</td><td>132.4K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>311.2M</td><td>2.2%</td><td>9.7M</td><td>49.7%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>311.2M</td><td>2.2%</td><td>9.7M</td><td>49.7%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-4 of 60 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen2_moe</td><td>qwen2_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen2Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>193</td><td>17</td></tr><tr class=""><td>RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>48</td><td>4</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Experts</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>24</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 30 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>5632</td><td>128</td></tr><tr><td>layer_types</td><td>24 entries: 24× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>1408</td><td>32</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts</td><td>60</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>4</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>16</td><td>4</td></tr><tr><td>shared_expert_intermediate_size</td><td>5632</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen1.5-MoE-A2.7B-mic.json b/shrink/Qwen1.5-MoE-A2.7B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e0be52aadb17b4663f33172e318aac332b8d0d5c
--- /dev/null
+++ b/shrink/Qwen1.5-MoE-A2.7B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen1.5-MoE-A2.7B-mic@2fd4813071f2c23b77666d8c796ff80235d42b55", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen2-VL-2B-Instruct-mic.html b/shrink/Qwen2-VL-2B-Instruct-mic.html
index 690097fa88e9a6eb7c992fb9033f92ff306a77a6..c9c21466ac7941bbc2a1306affa80f6a9a029fc8 100644
--- a/shrink/Qwen2-VL-2B-Instruct-mic.html
+++ b/shrink/Qwen2-VL-2B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen2-VL-2B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen2-VL-2B-Instruct-mic@743a11f21777bbaacd3519d20014517268130159 ← Qwen/Qwen2-VL-2B-Instruct@895c3a49bc3fa70a340399125c650a463535e71c · Qwen2VLForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen2-VL-2B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen2-VL-2B-Instruct-mic/tree/743a11f21777bbaacd3519d20014517268130159" title="huggingface-mic/Qwen2-VL-2B-Instruct-mic@743a11f21777bbaacd3519d20014517268130159">huggingface-mic/Qwen2-VL-2B-Instruct-mic@743a11f21777</a> ← <a href="https://huggingface.co/Qwen/Qwen2-VL-2B-Instruct/tree/895c3a49bc3fa70a340399125c650a463535e71c" title="Qwen/Qwen2-VL-2B-Instruct@895c3a49bc3fa70a340399125c650a463535e71c">Qwen/Qwen2-VL-2B-Instruct@895c3a49bc3f</a> · Qwen2VLForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.2B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">10.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i><i class="c0" title="layer 27: VisionBlock"></i><i class="c0" title="layer 28: VisionBlock"></i><i class="c0" title="layer 29: VisionBlock"></i><i class="c0" title="layer 30: VisionBlock"></i><i class="c0" title="layer 31: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.593186" title="model.language_model.layers: 1.3B (59.3%)"></i><i class="c1" style="flex:0.285053" title="model.visual.blocks: 629.7M (28.5%)"></i><i class="c2" style="flex:0.105648" title="model.language_model: 233.4M (10.6%)"></i><i class="c3" style="flex:0.016113" title="model.visual: 35.6M (1.6%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.007383" title="model.language_model.layers: 74.2K (0.7%)"></i><i class="c1" style="flex:0.009941" title="model.visual.blocks: 100.0K (1.0%)"></i><i class="c2" style="flex:0.967000" title="model.language_model: 9.7M (96.7%)"></i><i class="c3" style="flex:0.015676" title="model.visual: 157.6K (1.6%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>1.3B</td><td>59.3%</td><td>74.2K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.visual.blocks</td><td>629.7M</td><td>28.5%</td><td>100.0K</td><td>1.0%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>233.4M</td><td>10.6%</td><td>9.7M</td><td>96.7%</td></tr><tr><td><i class="c3"></i>model.visual</td><td>35.6M</td><td>1.6%</td><td>157.6K</td><td>1.6%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen2_vl</td><td>qwen2_vl</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (12 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2_vl_text</td><td>qwen2_vl_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen2_vl_vision</td><td>qwen2_vl_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen2VL omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>327</td><td>25</td></tr><tr class=""><td>LayerNorm</td><td>65</td><td>5</td></tr><tr class=""><td>RMSNorm</td><td>57</td><td>5</td></tr><tr class=""><td>QuickGELUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>VisionBlock</td><td>32</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>32</td><td>2</td></tr><tr class=""><td>VisionMlp</td><td>32</td><td>2</td></tr><tr class=""><td>Qwen2MLP</td><td>28</td><td>2</td></tr><tr class=""><td>Attention</td><td>28</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>28</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>PatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>PatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2VisionTransformerPretrainedModel</td><td>1</td><td>1</td></tr><tr class=""><td>Sequential</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 57 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>8960</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>28 entries: 28× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>28</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;type&quot;: &quot;mrope&quot;, &quot;mrope_section&quot;: [16, 24, 24], &quot;rope_theta&quot;: 1000000.0, &quot;ro…</td><td>{&quot;mrope_section&quot;: [2, 3, 3], &quot;rope_theta&quot;: 1000000.0, &quot;rope_type&quot;: &quot;default&quot;,…</td></tr><tr><td>vision_config.depth</td><td>32</td><td>2</td></tr><tr><td>vision_config.embed_dim</td><td>1280</td><td>64</td></tr><tr><td>vision_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen2-VL-2B-Instruct-mic.json b/shrink/Qwen2-VL-2B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..cbc8b56ad795ee4023548354af4deb4dbca33652
--- /dev/null
+++ b/shrink/Qwen2-VL-2B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen2-VL-2B-Instruct-mic@743a11f21777bbaacd3519d20014517268130159", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3-0.6B-mic.html b/shrink/Qwen3-0.6B-mic.html
index c922ea44176f803a343dac6d8f0cbf105377c3c6..1d5d8335b8ffcfdec86e636fea437dbc3d73b825 100644
--- a/shrink/Qwen3-0.6B-mic.html
+++ b/shrink/Qwen3-0.6B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3-0.6B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3-0.6B-mic@f3c69d34835d4cf1d9aba9a08cd6af8282aae6a7 ← Qwen/Qwen3-0.6B@c1899de289a04d12100db370d81485cdf75e47ca · Qwen3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3-0.6B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3-0.6B-mic/tree/f3c69d34835d4cf1d9aba9a08cd6af8282aae6a7" title="huggingface-mic/Qwen3-0.6B-mic@f3c69d34835d4cf1d9aba9a08cd6af8282aae6a7">huggingface-mic/Qwen3-0.6B-mic@f3c69d34835d</a> ← <a href="https://huggingface.co/Qwen/Qwen3-0.6B/tree/c1899de289a04d12100db370d81485cdf75e47ca" title="Qwen/Qwen3-0.6B@c1899de289a04d12100db370d81485cdf75e47ca">Qwen/Qwen3-0.6B@c1899de289a0</a> · Qwen3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">596.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">9.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.738976" title="model.layers: 440.5M (73.9%)"></i><i class="c1" style="flex:0.261023" title="model.embed_tokens: 155.6M (26.1%)"></i><i class="c2" style="flex:0.000002" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.007557" title="model.layers: 74.0K (0.8%)"></i><i class="c1" style="flex:0.992436" title="model.embed_tokens: 9.7M (99.2%)"></i><i class="c2" style="flex:0.000007" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>440.5M</td><td>73.9%</td><td>74.0K</td><td>0.8%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>155.6M</td><td>26.1%</td><td>9.7M</td><td>99.2%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3</td><td>qwen3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>40960</td><td>40960</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>197</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>113</td><td>9</td></tr><tr class=""><td>Attention</td><td>28</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>MLP</td><td>28</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>28</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 25 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>layer_types</td><td>28 entries: 28× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>28</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3-0.6B-mic.json b/shrink/Qwen3-0.6B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..936c10f2fd9ffcf802d2086dcc43d2ed474990e5
--- /dev/null
+++ b/shrink/Qwen3-0.6B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3-0.6B-mic@f3c69d34835d4cf1d9aba9a08cd6af8282aae6a7", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3-30B-A3B-mic.html b/shrink/Qwen3-30B-A3B-mic.html
index aa65446848e3a1bac86aa50a17ac13094f70a271..4c87337ab0c0395e33731836768f4f4401442019 100644
--- a/shrink/Qwen3-30B-A3B-mic.html
+++ b/shrink/Qwen3-30B-A3B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3-30B-A3B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3-30B-A3B-mic@02a209fca46248ff0b3da72e3b190428138b2449 ← Qwen/Qwen3-30B-A3B@ad44e777bcd18fa416d9da3bd8f70d33ebb85d39 · Qwen3MoeForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3-30B-A3B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3-30B-A3B-mic/tree/02a209fca46248ff0b3da72e3b190428138b2449" title="huggingface-mic/Qwen3-30B-A3B-mic@02a209fca46248ff0b3da72e3b190428138b2449">huggingface-mic/Qwen3-30B-A3B-mic@02a209fca462</a> ← <a href="https://huggingface.co/Qwen/Qwen3-30B-A3B/tree/ad44e777bcd18fa416d9da3bd8f70d33ebb85d39" title="Qwen/Qwen3-30B-A3B@ad44e777bcd18fa416d9da3bd8f70d33ebb85d39">Qwen/Qwen3-30B-A3B@ad44e777bcd1</a> · Qwen3MoeForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">30.5B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i><i class="c0" title="layer 40: DecoderLayer"></i><i class="c0" title="layer 41: DecoderLayer"></i><i class="c0" title="layer 42: DecoderLayer"></i><i class="c0" title="layer 43: DecoderLayer"></i><i class="c0" title="layer 44: DecoderLayer"></i><i class="c0" title="layer 45: DecoderLayer"></i><i class="c0" title="layer 46: DecoderLayer"></i><i class="c0" title="layer 47: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>48</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.979617" title="model.layers: 29.9B (98.0%)"></i><i class="c1" style="flex:0.010191" title="model.embed_tokens: 311.2M (1.0%)"></i><i class="c2" style="flex:0.010191" title="lm_head: 311.2M (1.0%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003819" title="model.layers: 74.6K (0.4%)"></i><i class="c1" style="flex:0.498089" title="model.embed_tokens: 9.7M (49.8%)"></i><i class="c2" style="flex:0.498089" title="lm_head: 9.7M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>29.9B</td><td>98.0%</td><td>74.6K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>311.2M</td><td>1.0%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>311.2M</td><td>1.0%</td><td>9.7M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 128 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_moe</td><td>qwen3_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>40960</td><td>40960</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>193</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>193</td><td>9</td></tr><tr class=""><td>Attention</td><td>48</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>48</td><td>2</td></tr><tr class=""><td>Experts</td><td>48</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>48</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>48</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>48</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 30 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>6144</td><td>128</td></tr><tr><td>moe_intermediate_size</td><td>768</td><td>32</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>48</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr><tr><td>num_local_experts</td><td>128</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3-30B-A3B-mic.json b/shrink/Qwen3-30B-A3B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..9896b630424d382fbb1d8ae72263037c55c41514
--- /dev/null
+++ b/shrink/Qwen3-30B-A3B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3-30B-A3B-mic@02a209fca46248ff0b3da72e3b190428138b2449", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3-ASR-0.6B-hf-mic.html b/shrink/Qwen3-ASR-0.6B-hf-mic.html
index c86e23146a081607cbf2341597151941b7d23610..02c68a3c29735dd7c7edc92663e288392ec30bc2 100644
--- a/shrink/Qwen3-ASR-0.6B-hf-mic.html
+++ b/shrink/Qwen3-ASR-0.6B-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3-ASR-0.6B-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3-ASR-0.6B-hf-mic@6b969c14b8cf776fa7152fa143a3c0dcc6a30b40 ← Qwen/Qwen3-ASR-0.6B-hf@7f1569a48a89f3e3f4dc3a5c9d28bddd903bc76c · Qwen3ASRForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3-ASR-0.6B-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3-ASR-0.6B-hf-mic/tree/6b969c14b8cf776fa7152fa143a3c0dcc6a30b40" title="huggingface-mic/Qwen3-ASR-0.6B-hf-mic@6b969c14b8cf776fa7152fa143a3c0dcc6a30b40">huggingface-mic/Qwen3-ASR-0.6B-hf-mic@6b969c14b8cf</a> ← <a href="https://huggingface.co/Qwen/Qwen3-ASR-0.6B-hf/tree/7f1569a48a89f3e3f4dc3a5c9d28bddd903bc76c" title="Qwen/Qwen3-ASR-0.6B-hf@7f1569a48a89f3e3f4dc3a5c9d28bddd903bc76c">Qwen/Qwen3-ASR-0.6B-hf@7f1569a48a89</a> · Qwen3ASRForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">782.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">14.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.audio_tower.layers</h3><div class="row"><span class="lab">Original · 18</span><div class="cells"><i class="c0" title="layer 0: AudioEncoderLayer"></i><i class="c0" title="layer 1: AudioEncoderLayer"></i><i class="c0" title="layer 2: AudioEncoderLayer"></i><i class="c0" title="layer 3: AudioEncoderLayer"></i><i class="c0" title="layer 4: AudioEncoderLayer"></i><i class="c0" title="layer 5: AudioEncoderLayer"></i><i class="c0" title="layer 6: AudioEncoderLayer"></i><i class="c0" title="layer 7: AudioEncoderLayer"></i><i class="c0" title="layer 8: AudioEncoderLayer"></i><i class="c0" title="layer 9: AudioEncoderLayer"></i><i class="c0" title="layer 10: AudioEncoderLayer"></i><i class="c0" title="layer 11: AudioEncoderLayer"></i><i class="c0" title="layer 12: AudioEncoderLayer"></i><i class="c0" title="layer 13: AudioEncoderLayer"></i><i class="c0" title="layer 14: AudioEncoderLayer"></i><i class="c0" title="layer 15: AudioEncoderLayer"></i><i class="c0" title="layer 16: AudioEncoderLayer"></i><i class="c0" title="layer 17: AudioEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: AudioEncoderLayer"></i><i class="c0" title="layer 1: AudioEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioEncoderLayer</td><td>18</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: Qwen3DecoderLayer"></i><i class="c0" title="layer 1: Qwen3DecoderLayer"></i><i class="c0" title="layer 2: Qwen3DecoderLayer"></i><i class="c0" title="layer 3: Qwen3DecoderLayer"></i><i class="c0" title="layer 4: Qwen3DecoderLayer"></i><i class="c0" title="layer 5: Qwen3DecoderLayer"></i><i class="c0" title="layer 6: Qwen3DecoderLayer"></i><i class="c0" title="layer 7: Qwen3DecoderLayer"></i><i class="c0" title="layer 8: Qwen3DecoderLayer"></i><i class="c0" title="layer 9: Qwen3DecoderLayer"></i><i class="c0" title="layer 10: Qwen3DecoderLayer"></i><i class="c0" title="layer 11: Qwen3DecoderLayer"></i><i class="c0" title="layer 12: Qwen3DecoderLayer"></i><i class="c0" title="layer 13: Qwen3DecoderLayer"></i><i class="c0" title="layer 14: Qwen3DecoderLayer"></i><i class="c0" title="layer 15: Qwen3DecoderLayer"></i><i class="c0" title="layer 16: Qwen3DecoderLayer"></i><i class="c0" title="layer 17: Qwen3DecoderLayer"></i><i class="c0" title="layer 18: Qwen3DecoderLayer"></i><i class="c0" title="layer 19: Qwen3DecoderLayer"></i><i class="c0" title="layer 20: Qwen3DecoderLayer"></i><i class="c0" title="layer 21: Qwen3DecoderLayer"></i><i class="c0" title="layer 22: Qwen3DecoderLayer"></i><i class="c0" title="layer 23: Qwen3DecoderLayer"></i><i class="c0" title="layer 24: Qwen3DecoderLayer"></i><i class="c0" title="layer 25: Qwen3DecoderLayer"></i><i class="c0" title="layer 26: Qwen3DecoderLayer"></i><i class="c0" title="layer 27: Qwen3DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen3DecoderLayer"></i><i class="c0" title="layer 1: Qwen3DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen3DecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.562950" title="model.language_model.layers: 440.5M (56.3%)"></i><i class="c1" style="flex:0.221897" title="model.audio_tower.layers: 173.6M (22.2%)"></i><i class="c2" style="flex:0.198848" title="model.language_model: 155.6M (19.9%)"></i><i class="c3" style="flex:0.014105" title="model.audio_tower: 11.0M (1.4%)"></i><i class="c4" style="flex:0.002201" title="model.multi_modal_projector: 1.7M (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.005100" title="model.language_model.layers: 74.0K (0.5%)"></i><i class="c1" style="flex:0.004611" title="model.audio_tower.layers: 66.9K (0.5%)"></i><i class="c2" style="flex:0.669792" title="model.language_model: 9.7M (67.0%)"></i><i class="c3" style="flex:0.319923" title="model.audio_tower: 4.6M (32.0%)"></i><i class="c4" style="flex:0.000573" title="model.multi_modal_projector: 8.3K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>440.5M</td><td>56.3%</td><td>74.0K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>model.audio_tower.layers</td><td>173.6M</td><td>22.2%</td><td>66.9K</td><td>0.5%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>155.6M</td><td>19.9%</td><td>9.7M</td><td>67.0%</td></tr><tr><td><i class="c3"></i>model.audio_tower</td><td>11.0M</td><td>1.4%</td><td>4.6M</td><td>32.0%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>1.7M</td><td>0.2%</td><td>8.3K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_asr</td><td>qwen3_asr</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>qwen3_asr_encoder</td><td>qwen3_asr_encoder</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>max_position_embeddings</td><td>13</td><td>13</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>128</td><td>128</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3</td><td>qwen3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>65536</td><td>65536</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3ASR omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>308</td><td>30</td></tr><tr class=""><td>Qwen3RMSNorm</td><td>113</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>37</td><td>5</td></tr><tr class=""><td>Qwen3Attention</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen3DecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen3MLP</td><td>28</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>28</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>19</td><td>3</td></tr><tr class=""><td>AudioAttention</td><td>18</td><td>2</td></tr><tr class=""><td>AudioEncoderLayer</td><td>18</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>3</td><td>3</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen3RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>SinusoidsPositionEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 62 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.d_model</td><td>896</td><td>64</td></tr><tr><td>audio_config.encoder_attention_heads</td><td>14</td><td>4</td></tr><tr><td>audio_config.encoder_ffn_dim</td><td>3584</td><td>128</td></tr><tr><td>audio_config.encoder_layers</td><td>18</td><td>2</td></tr><tr><td>audio_config.num_key_value_heads</td><td>14</td><td>4</td></tr><tr><td>audio_config.output_dim</td><td>1024</td><td>64</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>28 entries: 28× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>28</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3-ASR-0.6B-hf-mic.json b/shrink/Qwen3-ASR-0.6B-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..db4dd0abba39c89c0d004ffe8a25a7f731f14830
--- /dev/null
+++ b/shrink/Qwen3-ASR-0.6B-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3-ASR-0.6B-hf-mic@6b969c14b8cf776fa7152fa143a3c0dcc6a30b40", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3-Next-80B-A3B-Instruct-mic.html b/shrink/Qwen3-Next-80B-A3B-Instruct-mic.html
index 9f73633c9e94019abd58e630c7b78afe76724c2c..b74640c7d2421b21be6ee294db963125af1b5b05 100644
--- a/shrink/Qwen3-Next-80B-A3B-Instruct-mic.html
+++ b/shrink/Qwen3-Next-80B-A3B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3-Next-80B-A3B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3-Next-80B-A3B-Instruct-mic@69466af2a651a5263712f1430d3baf98b0e16463 ← Qwen/Qwen3-Next-80B-A3B-Instruct@9c7f2fbe84465e40164a94cc16cd30b6999b0cc7 · Qwen3NextForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3-Next-80B-A3B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3-Next-80B-A3B-Instruct-mic/tree/69466af2a651a5263712f1430d3baf98b0e16463" title="huggingface-mic/Qwen3-Next-80B-A3B-Instruct-mic@69466af2a651a5263712f1430d3baf98b0e16463">huggingface-mic/Qwen3-Next-80B-A3B-Instruct-mic@69466af2a651</a> ← <a href="https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct/tree/9c7f2fbe84465e40164a94cc16cd30b6999b0cc7" title="Qwen/Qwen3-Next-80B-A3B-Instruct@9c7f2fbe84465e40164a94cc16cd30b6999b0cc7">Qwen/Qwen3-Next-80B-A3B-Instruct@9c7f2fbe8446</a> · Qwen3NextForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">79.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">20.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 3: is_causal=True, Attention"></i><i class="c0" title="layer 4: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 7: is_causal=True, Attention"></i><i class="c0" title="layer 8: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 9: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 10: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 11: is_causal=True, Attention"></i><i class="c0" title="layer 12: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 13: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 14: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 15: is_causal=True, Attention"></i><i class="c0" title="layer 16: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 17: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 18: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 19: is_causal=True, Attention"></i><i class="c0" title="layer 20: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 21: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 22: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 23: is_causal=True, Attention"></i><i class="c0" title="layer 24: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 25: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 26: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 27: is_causal=True, Attention"></i><i class="c0" title="layer 28: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 29: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 30: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 31: is_causal=True, Attention"></i><i class="c0" title="layer 32: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 33: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 34: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 35: is_causal=True, Attention"></i><i class="c0" title="layer 36: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 37: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 38: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 39: is_causal=True, Attention"></i><i class="c0" title="layer 40: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 41: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 42: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 43: is_causal=True, Attention"></i><i class="c0" title="layer 44: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 45: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 46: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 47: is_causal=True, Attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, GatedDeltaNet …</td><td>36</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>12</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.992189" title="model.layers: 79.1B (99.2%)"></i><i class="c1" style="flex:0.003905" title="lm_head: 311.2M (0.4%)"></i><i class="c2" style="flex:0.003905" title="model.embed_tokens: 311.2M (0.4%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.058174" title="model.layers: 1.2M (5.8%)"></i><i class="c1" style="flex:0.470911" title="lm_head: 9.7M (47.1%)"></i><i class="c2" style="flex:0.470911" title="model.embed_tokens: 9.7M (47.1%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>79.1B</td><td>99.2%</td><td>1.2M</td><td>5.8%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>311.2M</td><td>0.4%</td><td>9.7M</td><td>47.1%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>311.2M</td><td>0.4%</td><td>9.7M</td><td>47.1%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-10 of 512 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_next</td><td>qwen3_next</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3Next omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>349</td><td>16</td></tr><tr class=""><td>RMSNorm</td><td>121</td><td>7</td></tr><tr class=""><td>SiLUActivation</td><td>96</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>48</td><td>2</td></tr><tr class=""><td>Experts</td><td>48</td><td>2</td></tr><tr class=""><td>MLP</td><td>48</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>48</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>48</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>36</td><td>1</td></tr><tr class=""><td>GatedDeltaNet</td><td>36</td><td>1</td></tr><tr class=""><td>RMSNormGated</td><td>36</td><td>1</td></tr><tr class=""><td>Attention</td><td>12</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 34 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>256</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>5120</td><td>128</td></tr><tr><td>layer_types</td><td>48 entries: 36× linear_attention, 12× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>512</td><td>32</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts</td><td>512</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>10</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>48</td><td>2</td></tr><tr><td>shared_expert_intermediate_size</td><td>512</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3-Next-80B-A3B-Instruct-mic.json b/shrink/Qwen3-Next-80B-A3B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..447b4e5426f890a671c44958786f1260e9938898
--- /dev/null
+++ b/shrink/Qwen3-Next-80B-A3B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3-Next-80B-A3B-Instruct-mic@69466af2a651a5263712f1430d3baf98b0e16463", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3-VL-2B-Instruct-mic.html b/shrink/Qwen3-VL-2B-Instruct-mic.html
index 6b8c942f827dc941454bc337b615089f30c05ea0..edba00189fff820c505c532cfa9858a8a0e2966f 100644
--- a/shrink/Qwen3-VL-2B-Instruct-mic.html
+++ b/shrink/Qwen3-VL-2B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3-VL-2B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3-VL-2B-Instruct-mic@7e045dfc87cda2240c98f5ae78e8bb377a08dc38 ← Qwen/Qwen3-VL-2B-Instruct@89644892e4d85e24eaac8bacfd4f463576704203 · Qwen3VLForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3-VL-2B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3-VL-2B-Instruct-mic/tree/7e045dfc87cda2240c98f5ae78e8bb377a08dc38" title="huggingface-mic/Qwen3-VL-2B-Instruct-mic@7e045dfc87cda2240c98f5ae78e8bb377a08dc38">huggingface-mic/Qwen3-VL-2B-Instruct-mic@7e045dfc87cd</a> ← <a href="https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct/tree/89644892e4d85e24eaac8bacfd4f463576704203" title="Qwen/Qwen3-VL-2B-Instruct@89644892e4d85e24eaac8bacfd4f463576704203">Qwen/Qwen3-VL-2B-Instruct@89644892e4d8</a> · Qwen3VLForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">11.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 24</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>24</td><td>24</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.visual.deepstack_merger_list</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: GELU"></i><i class="c0" title="layer 1: GELU"></i><i class="c0" title="layer 2: GELU"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: GELU"></i><i class="c0" title="layer 1: GELU"></i><i class="c0" title="layer 2: GELU"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>GELU</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i><i class="c0" title="layer 2: TextDecoderLayer"></i><i class="c0" title="layer 3: TextDecoderLayer"></i><i class="c0" title="layer 4: TextDecoderLayer"></i><i class="c0" title="layer 5: TextDecoderLayer"></i><i class="c0" title="layer 6: TextDecoderLayer"></i><i class="c0" title="layer 7: TextDecoderLayer"></i><i class="c0" title="layer 8: TextDecoderLayer"></i><i class="c0" title="layer 9: TextDecoderLayer"></i><i class="c0" title="layer 10: TextDecoderLayer"></i><i class="c0" title="layer 11: TextDecoderLayer"></i><i class="c0" title="layer 12: TextDecoderLayer"></i><i class="c0" title="layer 13: TextDecoderLayer"></i><i class="c0" title="layer 14: TextDecoderLayer"></i><i class="c0" title="layer 15: TextDecoderLayer"></i><i class="c0" title="layer 16: TextDecoderLayer"></i><i class="c0" title="layer 17: TextDecoderLayer"></i><i class="c0" title="layer 18: TextDecoderLayer"></i><i class="c0" title="layer 19: TextDecoderLayer"></i><i class="c0" title="layer 20: TextDecoderLayer"></i><i class="c0" title="layer 21: TextDecoderLayer"></i><i class="c0" title="layer 22: TextDecoderLayer"></i><i class="c0" title="layer 23: TextDecoderLayer"></i><i class="c0" title="layer 24: TextDecoderLayer"></i><i class="c0" title="layer 25: TextDecoderLayer"></i><i class="c0" title="layer 26: TextDecoderLayer"></i><i class="c0" title="layer 27: TextDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextDecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.662461" title="model.language_model.layers: 1.4B (66.2%)"></i><i class="c1" style="flex:0.146257" title="model.language_model: 311.2M (14.6%)"></i><i class="c2" style="flex:0.142094" title="model.visual.blocks: 302.3M (14.2%)"></i><i class="c3" style="flex:0.035506" title="model.visual.deepstack_merger_list: 75.5M (3.6%)"></i><i class="c4" style="flex:0.013681" title="model.visual: 29.1M (1.4%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006625" title="model.language_model.layers: 74.0K (0.7%)"></i><i class="c1" style="flex:0.869936" title="model.language_model: 9.7M (87.0%)"></i><i class="c2" style="flex:0.071868" title="model.visual.blocks: 803.3K (7.2%)"></i><i class="c3" style="flex:0.022210" title="model.visual.deepstack_merger_list: 248.3K (2.2%)"></i><i class="c4" style="flex:0.029361" title="model.visual: 328.2K (2.9%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>1.4B</td><td>66.2%</td><td>74.0K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>311.2M</td><td>14.6%</td><td>9.7M</td><td>87.0%</td></tr><tr><td><i class="c2"></i>model.visual.blocks</td><td>302.3M</td><td>14.2%</td><td>803.3K</td><td>7.2%</td></tr><tr><td><i class="c3"></i>model.visual.deepstack_merger_list</td><td>75.5M</td><td>3.6%</td><td>248.3K</td><td>2.2%</td></tr><tr><td><i class="c4"></i>model.visual</td><td>29.1M</td><td>1.4%</td><td>328.2K</td><td>2.9%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_vl</td><td>qwen3_vl</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3_vl_text</td><td>qwen3_vl_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen3_vl_vision</td><td>qwen3_vl_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3VL omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>301</td><td>119</td></tr><tr class=""><td>TextRMSNorm</td><td>113</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>52</td><td>52</td></tr><tr class=""><td>TextAttention</td><td>28</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>TextMLP</td><td>28</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>28</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>24</td><td>24</td></tr><tr class=""><td>VisionAttention</td><td>24</td><td>24</td></tr><tr class=""><td>VisionBlock</td><td>24</td><td>24</td></tr><tr class=""><td>VisionMLP</td><td>24</td><td>24</td></tr><tr class=""><td>GELU</td><td>4</td><td>4</td></tr><tr class=""><td>VisionPatchMerger</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 54 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>6144</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>28</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [24, 20, 20], &quot;rope_type&quot;: &quot;defa…</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [3, 3, 2], &quot;rope_theta&quot;: 5000000…</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>2048</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3-VL-2B-Instruct-mic.json b/shrink/Qwen3-VL-2B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..556fd5f6324b52f1ee6e12d269a118afc671b393
--- /dev/null
+++ b/shrink/Qwen3-VL-2B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3-VL-2B-Instruct-mic@7e045dfc87cda2240c98f5ae78e8bb377a08dc38", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3-VL-30B-A3B-Instruct-mic.html b/shrink/Qwen3-VL-30B-A3B-Instruct-mic.html
index eac4522f115731506fdb57e3bd45f21423c4665b..8028a3b46aa22bec96c40d69222a9f1a595bb107 100644
--- a/shrink/Qwen3-VL-30B-A3B-Instruct-mic.html
+++ b/shrink/Qwen3-VL-30B-A3B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3-VL-30B-A3B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3-VL-30B-A3B-Instruct-mic@fce9290a8553fe0d943eb0e9c8e577bc5e7e3631 ← Qwen/Qwen3-VL-30B-A3B-Instruct@9c4b90e1e4ba969fd3b5378b57d966d725f1b86c · Qwen3VLMoeForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3-VL-30B-A3B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3-VL-30B-A3B-Instruct-mic/tree/fce9290a8553fe0d943eb0e9c8e577bc5e7e3631" title="huggingface-mic/Qwen3-VL-30B-A3B-Instruct-mic@fce9290a8553fe0d943eb0e9c8e577bc5e7e3631">huggingface-mic/Qwen3-VL-30B-A3B-Instruct-mic@fce9290a8553</a> ← <a href="https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Instruct/tree/9c4b90e1e4ba969fd3b5378b57d966d725f1b86c" title="Qwen/Qwen3-VL-30B-A3B-Instruct@9c4b90e1e4ba969fd3b5378b57d966d725f1b86c">Qwen/Qwen3-VL-30B-A3B-Instruct@9c4b90e1e4ba</a> · Qwen3VLMoeForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">31.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">21.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>27</td><td>27</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.visual.deepstack_merger_list</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: GELU"></i><i class="c0" title="layer 1: GELU"></i><i class="c0" title="layer 2: GELU"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: GELU"></i><i class="c0" title="layer 1: GELU"></i><i class="c0" title="layer 2: GELU"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>GELU</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i><i class="c0" title="layer 2: TextDecoderLayer"></i><i class="c0" title="layer 3: TextDecoderLayer"></i><i class="c0" title="layer 4: TextDecoderLayer"></i><i class="c0" title="layer 5: TextDecoderLayer"></i><i class="c0" title="layer 6: TextDecoderLayer"></i><i class="c0" title="layer 7: TextDecoderLayer"></i><i class="c0" title="layer 8: TextDecoderLayer"></i><i class="c0" title="layer 9: TextDecoderLayer"></i><i class="c0" title="layer 10: TextDecoderLayer"></i><i class="c0" title="layer 11: TextDecoderLayer"></i><i class="c0" title="layer 12: TextDecoderLayer"></i><i class="c0" title="layer 13: TextDecoderLayer"></i><i class="c0" title="layer 14: TextDecoderLayer"></i><i class="c0" title="layer 15: TextDecoderLayer"></i><i class="c0" title="layer 16: TextDecoderLayer"></i><i class="c0" title="layer 17: TextDecoderLayer"></i><i class="c0" title="layer 18: TextDecoderLayer"></i><i class="c0" title="layer 19: TextDecoderLayer"></i><i class="c0" title="layer 20: TextDecoderLayer"></i><i class="c0" title="layer 21: TextDecoderLayer"></i><i class="c0" title="layer 22: TextDecoderLayer"></i><i class="c0" title="layer 23: TextDecoderLayer"></i><i class="c0" title="layer 24: TextDecoderLayer"></i><i class="c0" title="layer 25: TextDecoderLayer"></i><i class="c0" title="layer 26: TextDecoderLayer"></i><i class="c0" title="layer 27: TextDecoderLayer"></i><i class="c0" title="layer 28: TextDecoderLayer"></i><i class="c0" title="layer 29: TextDecoderLayer"></i><i class="c0" title="layer 30: TextDecoderLayer"></i><i class="c0" title="layer 31: TextDecoderLayer"></i><i class="c0" title="layer 32: TextDecoderLayer"></i><i class="c0" title="layer 33: TextDecoderLayer"></i><i class="c0" title="layer 34: TextDecoderLayer"></i><i class="c0" title="layer 35: TextDecoderLayer"></i><i class="c0" title="layer 36: TextDecoderLayer"></i><i class="c0" title="layer 37: TextDecoderLayer"></i><i class="c0" title="layer 38: TextDecoderLayer"></i><i class="c0" title="layer 39: TextDecoderLayer"></i><i class="c0" title="layer 40: TextDecoderLayer"></i><i class="c0" title="layer 41: TextDecoderLayer"></i><i class="c0" title="layer 42: TextDecoderLayer"></i><i class="c0" title="layer 43: TextDecoderLayer"></i><i class="c0" title="layer 44: TextDecoderLayer"></i><i class="c0" title="layer 45: TextDecoderLayer"></i><i class="c0" title="layer 46: TextDecoderLayer"></i><i class="c0" title="layer 47: TextDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextDecoderLayer</td><td>48</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.962635" title="model.language_model.layers: 29.9B (96.3%)"></i><i class="c1" style="flex:0.013243" title="model.visual.blocks: 411.5M (1.3%)"></i><i class="c2" style="flex:0.010015" title="model.language_model: 311.2M (1.0%)"></i><i class="c3" style="flex:0.010015" title="lm_head: 311.2M (1.0%)"></i><i class="c4" style="flex:0.002963" title="model.visual.deepstack_merger_list: 92.1M (0.3%)"></i><i class="c5" style="flex:0.001130" title="model.visual: 35.1M (0.1%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003550" title="model.language_model.layers: 74.6K (0.4%)"></i><i class="c1" style="flex:0.043030" title="model.visual.blocks: 903.7K (4.3%)"></i><i class="c2" style="flex:0.462988" title="model.language_model: 9.7M (46.3%)"></i><i class="c3" style="flex:0.462985" title="lm_head: 9.7M (46.3%)"></i><i class="c4" style="flex:0.011820" title="model.visual.deepstack_merger_list: 248.3K (1.2%)"></i><i class="c5" style="flex:0.015626" title="model.visual: 328.2K (1.6%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>29.9B</td><td>96.3%</td><td>74.6K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.visual.blocks</td><td>411.5M</td><td>1.3%</td><td>903.7K</td><td>4.3%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>311.2M</td><td>1.0%</td><td>9.7M</td><td>46.3%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>311.2M</td><td>1.0%</td><td>9.7M</td><td>46.3%</td></tr><tr><td><i class="c4"></i>model.visual.deepstack_merger_list</td><td>92.1M</td><td>0.3%</td><td>248.3K</td><td>1.2%</td></tr><tr><td><i class="c5"></i>model.visual</td><td>35.1M</td><td>0.1%</td><td>328.2K</td><td>1.6%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_vl_moe</td><td>qwen3_vl_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (32 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-8 of 128 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3_vl_moe_text</td><td>qwen3_vl_moe_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen3_vl_moe_vision</td><td>qwen3_vl_moe_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3VLMoe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>309</td><td>125</td></tr><tr class=""><td>TextRMSNorm</td><td>193</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>58</td><td>58</td></tr><tr class=""><td>TextAttention</td><td>48</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>48</td><td>2</td></tr><tr class=""><td>TextExperts</td><td>48</td><td>2</td></tr><tr class=""><td>TextSparseMoeBlock</td><td>48</td><td>2</td></tr><tr class=""><td>TextTopKRouter</td><td>48</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>48</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>27</td><td>27</td></tr><tr class=""><td>VisionAttention</td><td>27</td><td>27</td></tr><tr class=""><td>VisionBlock</td><td>27</td><td>27</td></tr><tr class=""><td>VisionMLP</td><td>27</td><td>27</td></tr><tr class=""><td>GELU</td><td>4</td><td>4</td></tr><tr class=""><td>VisionPatchMerger</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">14 fields changed, 60 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>6144</td><td>128</td></tr><tr><td>text_config.moe_intermediate_size</td><td>768</td><td>32</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>48</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>4</td><td>2</td></tr><tr><td>text_config.num_local_experts</td><td>128</td><td>4</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [24, 20, 20], &quot;rope_type&quot;: &quot;defa…</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [3, 3, 2], &quot;rope_theta&quot;: 5000000…</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>2048</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3-VL-30B-A3B-Instruct-mic.json b/shrink/Qwen3-VL-30B-A3B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..28f67413c9517cd2879e2f7dcb2e26db994ff7ef
--- /dev/null
+++ b/shrink/Qwen3-VL-30B-A3B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3-VL-30B-A3B-Instruct-mic@fce9290a8553fe0d943eb0e9c8e577bc5e7e3631", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3.5-0.8B-text-only-mic.html b/shrink/Qwen3.5-0.8B-text-only-mic.html
index eab5fc6ce005615e3c5af4448ef06b58efa64036..1b162aaf2123b1290fe7fae607bec6cfce73b546 100644
--- a/shrink/Qwen3.5-0.8B-text-only-mic.html
+++ b/shrink/Qwen3.5-0.8B-text-only-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>principled-intelligence/Qwen3.5-0.8B-text-only: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3.5-0.8B-text-only-mic@6fbae6fc1ebdd102c3eb6a308aa9c8e9a73ac9e5 ← principled-intelligence/Qwen3.5-0.8B-text-only@d692a5787e0534f0fd3484d7a516bd3c2d23dbed · Qwen3_5ForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>principled-intelligence/Qwen3.5-0.8B-text-only</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3.5-0.8B-text-only-mic/tree/6fbae6fc1ebdd102c3eb6a308aa9c8e9a73ac9e5" title="huggingface-mic/Qwen3.5-0.8B-text-only-mic@6fbae6fc1ebdd102c3eb6a308aa9c8e9a73ac9e5">huggingface-mic/Qwen3.5-0.8B-text-only-mic@6fbae6fc1ebd</a> ← <a href="https://huggingface.co/principled-intelligence/Qwen3.5-0.8B-text-only/tree/d692a5787e0534f0fd3484d7a516bd3c2d23dbed" title="principled-intelligence/Qwen3.5-0.8B-text-only@d692a5787e0534f0fd3484d7a516bd3c2d23dbed">principled-intelligence/Qwen3.5-0.8B-text-only@d692a5787e05</a> · Qwen3_5ForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">752.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 3: is_causal=True, Attention"></i><i class="c0" title="layer 4: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 7: is_causal=True, Attention"></i><i class="c0" title="layer 8: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 9: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 10: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 11: is_causal=True, Attention"></i><i class="c0" title="layer 12: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 13: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 14: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 15: is_causal=True, Attention"></i><i class="c0" title="layer 16: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 17: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 18: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 19: is_causal=True, Attention"></i><i class="c0" title="layer 20: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 21: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 22: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 23: is_causal=True, Attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, GatedDeltaNet …</td><td>18</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>6</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.662037" title="model.layers: 498.1M (66.2%)"></i><i class="c1" style="flex:0.337961" title="model.embed_tokens: 254.3M (33.8%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.044949" title="model.layers: 748.0K (4.5%)"></i><i class="c1" style="flex:0.955048" title="model.embed_tokens: 15.9M (95.5%)"></i><i class="c2" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>498.1M</td><td>66.2%</td><td>748.0K</td><td>4.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>254.3M</td><td>33.8%</td><td>15.9M</td><td>95.5%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_5_text</td><td>qwen3_5_text</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>248320</td><td>248320</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3_5 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>187</td><td>16</td></tr><tr class=""><td>RMSNorm</td><td>61</td><td>7</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>18</td><td>1</td></tr><tr class=""><td>GatedDeltaNet</td><td>18</td><td>1</td></tr><tr class=""><td>RMSNormGated</td><td>18</td><td>1</td></tr><tr class=""><td>Attention</td><td>6</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 33 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>256</td><td>16</td></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>3584</td><td>128</td></tr><tr><td>layer_types</td><td>24 entries: 18× linear_attention, 6× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>rope_parameters</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [11, 11, 10], &quot;partial_rotary_fa…</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [3, 3, 2], &quot;partial_rotary_facto…</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3.5-0.8B-text-only-mic.json b/shrink/Qwen3.5-0.8B-text-only-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..62736c556389c167155965ce6c27f1625fe99fc1
--- /dev/null
+++ b/shrink/Qwen3.5-0.8B-text-only-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3.5-0.8B-text-only-mic@6fbae6fc1ebdd102c3eb6a308aa9c8e9a73ac9e5", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3.5-35B-A3B-mic.html b/shrink/Qwen3.5-35B-A3B-mic.html
index a8cc190e6437096bce5ba05f628a0120c12ec6a3..6f6be49324d6da7175c314f216c5f096b18390af 100644
--- a/shrink/Qwen3.5-35B-A3B-mic.html
+++ b/shrink/Qwen3.5-35B-A3B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3.5-35B-A3B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3.5-35B-A3B-mic@8197a27e48d360625a51c9ddf03148b0e186f558 ← Qwen/Qwen3.5-35B-A3B@59d61f3ce65a6d9863b86d2e96597125219dc754 · Qwen3_5MoeForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3.5-35B-A3B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3.5-35B-A3B-mic/tree/8197a27e48d360625a51c9ddf03148b0e186f558" title="huggingface-mic/Qwen3.5-35B-A3B-mic@8197a27e48d360625a51c9ddf03148b0e186f558">huggingface-mic/Qwen3.5-35B-A3B-mic@8197a27e48d3</a> ← <a href="https://huggingface.co/Qwen/Qwen3.5-35B-A3B/tree/59d61f3ce65a6d9863b86d2e96597125219dc754" title="Qwen/Qwen3.5-35B-A3B@59d61f3ce65a6d9863b86d2e96597125219dc754">Qwen/Qwen3.5-35B-A3B@59d61f3ce65a</a> · Qwen3_5MoeForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">35.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">34.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>27</td><td>27</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 3: is_causal=True, Attention"></i><i class="c0" title="layer 4: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 7: is_causal=True, Attention"></i><i class="c0" title="layer 8: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 9: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 10: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 11: is_causal=True, Attention"></i><i class="c0" title="layer 12: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 13: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 14: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 15: is_causal=True, Attention"></i><i class="c0" title="layer 16: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 17: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 18: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 19: is_causal=True, Attention"></i><i class="c0" title="layer 20: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 21: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 22: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 23: is_causal=True, Attention"></i><i class="c0" title="layer 24: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 25: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 26: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 27: is_causal=True, Attention"></i><i class="c0" title="layer 28: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 29: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 30: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 31: is_causal=True, Attention"></i><i class="c0" title="layer 32: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 33: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 34: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 35: is_causal=True, Attention"></i><i class="c0" title="layer 36: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 37: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 38: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 39: is_causal=True, Attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, GatedDeltaNet …</td><td>30</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>10</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.958308" title="model.language_model.layers: 33.6B (95.8%)"></i><i class="c1" style="flex:0.014486" title="model.language_model: 508.6M (1.4%)"></i><i class="c2" style="flex:0.014486" title="lm_head: 508.6M (1.4%)"></i><i class="c3" style="flex:0.011720" title="model.visual.blocks: 411.5M (1.2%)"></i><i class="c4" style="flex:0.001000" title="model.visual: 35.1M (0.1%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.035106" title="model.language_model.layers: 1.2M (3.5%)"></i><i class="c1" style="flex:0.464447" title="model.language_model: 15.9M (46.4%)"></i><i class="c2" style="flex:0.464445" title="lm_head: 15.9M (46.4%)"></i><i class="c3" style="flex:0.026411" title="model.visual.blocks: 903.7K (2.6%)"></i><i class="c4" style="flex:0.009591" title="model.visual: 328.2K (1.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>33.6B</td><td>95.8%</td><td>1.2M</td><td>3.5%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>508.6M</td><td>1.4%</td><td>15.9M</td><td>46.4%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>508.6M</td><td>1.4%</td><td>15.9M</td><td>46.4%</td></tr><tr><td><i class="c3"></i>model.visual.blocks</td><td>411.5M</td><td>1.2%</td><td>903.7K</td><td>2.6%</td></tr><tr><td><i class="c4"></i>model.visual</td><td>35.1M</td><td>0.1%</td><td>328.2K</td><td>1.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_5_moe</td><td>qwen3_5_moe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-8 of 256 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen3_5_moe_text</td><td>qwen3_5_moe_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>248320</td><td>248320</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen3_5_moe_vision</td><td>qwen3_5_moe_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3_5Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>461</td><td>128</td></tr><tr class=""><td>RMSNorm</td><td>101</td><td>7</td></tr><tr class=""><td>SiLUActivation</td><td>80</td><td>4</td></tr><tr class=""><td>LayerNorm</td><td>55</td><td>55</td></tr><tr class=""><td>DecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>Experts</td><td>40</td><td>2</td></tr><tr class=""><td>MLP</td><td>40</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>40</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>40</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>30</td><td>1</td></tr><tr class=""><td>GatedDeltaNet</td><td>30</td><td>1</td></tr><tr class=""><td>RMSNormGated</td><td>30</td><td>1</td></tr><tr class=""><td>GELUTanh</td><td>27</td><td>27</td></tr><tr class=""><td>VisionAttention</td><td>27</td><td>27</td></tr><tr class=""><td>VisionBlock</td><td>27</td><td>27</td></tr><tr class=""><td>VisionMLP</td><td>27</td><td>27</td></tr><tr class=""><td>Attention</td><td>10</td><td>1</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">14 fields changed, 68 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.layer_types</td><td>40 entries: 30× linear_attention, 10× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.moe_intermediate_size</td><td>512</td><td>32</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_experts</td><td>256</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [11, 11, 10], &quot;rope_type&quot;: &quot;defa…</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [3, 3, 2], &quot;partial_rotary_facto…</td></tr><tr><td>text_config.shared_expert_intermediate_size</td><td>512</td><td>128</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>2048</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3.5-35B-A3B-mic.json b/shrink/Qwen3.5-35B-A3B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..b45ae1cb663cdb35650df55af81dd84bc09badd7
--- /dev/null
+++ b/shrink/Qwen3.5-35B-A3B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3.5-35B-A3B-mic@8197a27e48d360625a51c9ddf03148b0e186f558", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3.8-1.0B-A0.6B-mic.html b/shrink/Qwen3.8-1.0B-A0.6B-mic.html
index 7072956badebe32c77bcd8112678eab91e9a0207..bf89157814e9cee5e9a1b8ad8ad9d06ec05ed0dd 100644
--- a/shrink/Qwen3.8-1.0B-A0.6B-mic.html
+++ b/shrink/Qwen3.8-1.0B-A0.6B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>inference-optimization/Qwen3.8-1.0B-A0.6B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3.8-1.0B-A0.6B-mic@9633d99d303ac2051fd49702bc4781e0f930cfcc ← inference-optimization/Qwen3.8-1.0B-A0.6B@36531b2b4eab12e9cf7b7b9cbfce0f0a0bb8b719 · Qwen3_5MoeForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">966.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">32.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 3: is_causal=True, Attention"></i><i class="c0" title="layer 4: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 7: is_causal=True, Attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, GatedDeltaNet …</td><td>6</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>2</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.473611" title="model.layers: 457.6M (47.4%)"></i><i class="c1" style="flex:0.263194" title="model.embed_tokens: 254.3M (26.3%)"></i><i class="c2" style="flex:0.263194" title="lm_head: 254.3M (26.3%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.010992" title="model.layers: 353.3K (1.1%)"></i><i class="c1" style="flex:0.494503" title="model.embed_tokens: 15.9M (49.5%)"></i><i class="c2" style="flex:0.494503" title="lm_head: 15.9M (49.5%)"></i><i class="c3" style="flex:0.000002" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>457.6M</td><td>47.4%</td><td>353.3K</td><td>1.1%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>254.3M</td><td>26.3%</td><td>15.9M</td><td>49.5%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>254.3M</td><td>26.3%</td><td>15.9M</td><td>49.5%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 16 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_5_moe_text</td><td>qwen3_5_moe_text</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>248320</td><td>248320</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3_5Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>71</td><td>18</td></tr><tr class=""><td>RMSNorm</td><td>21</td><td>7</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>8</td><td>2</td></tr><tr class=""><td>Experts</td><td>8</td><td>2</td></tr><tr class=""><td>MLP</td><td>8</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>8</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>8</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>6</td><td>1</td></tr><tr class=""><td>GatedDeltaNet</td><td>6</td><td>1</td></tr><tr class=""><td>RMSNormGated</td><td>6</td><td>1</td></tr><tr class=""><td>Attention</td><td>2</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">8 fields changed, 37 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>layer_types</td><td>8 entries: 6× linear_attention, 2× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>1024</td><td>32</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_experts</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>8</td><td>2</td></tr><tr><td>shared_expert_intermediate_size</td><td>1024</td><td>128</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>inference-optimization/Qwen3.8-1.0B-A0.6B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3.8-1.0B-A0.6B-mic/tree/9633d99d303ac2051fd49702bc4781e0f930cfcc" title="huggingface-mic/Qwen3.8-1.0B-A0.6B-mic@9633d99d303ac2051fd49702bc4781e0f930cfcc">huggingface-mic/Qwen3.8-1.0B-A0.6B-mic@9633d99d303a</a> ← <a href="https://huggingface.co/inference-optimization/Qwen3.8-1.0B-A0.6B/tree/36531b2b4eab12e9cf7b7b9cbfce0f0a0bb8b719" title="inference-optimization/Qwen3.8-1.0B-A0.6B@36531b2b4eab12e9cf7b7b9cbfce0f0a0bb8b719">inference-optimization/Qwen3.8-1.0B-A0.6B@36531b2b4eab</a> · Qwen3_5MoeForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">966.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">32.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 3: is_causal=True, Attention"></i><i class="c0" title="layer 4: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 7: is_causal=True, Attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, GatedDeltaNet …"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, GatedDeltaNet …</td><td>6</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>2</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.473611" title="model.layers: 457.6M (47.4%)"></i><i class="c1" style="flex:0.263194" title="lm_head: 254.3M (26.3%)"></i><i class="c2" style="flex:0.263194" title="model.embed_tokens: 254.3M (26.3%)"></i><i class="c3" style="flex:0.000001" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.010992" title="model.layers: 353.3K (1.1%)"></i><i class="c1" style="flex:0.494503" title="lm_head: 15.9M (49.5%)"></i><i class="c2" style="flex:0.494503" title="model.embed_tokens: 15.9M (49.5%)"></i><i class="c3" style="flex:0.000002" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>457.6M</td><td>47.4%</td><td>353.3K</td><td>1.1%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>254.3M</td><td>26.3%</td><td>15.9M</td><td>49.5%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>254.3M</td><td>26.3%</td><td>15.9M</td><td>49.5%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-2 of 16 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>qwen3_5_moe_text</td><td>qwen3_5_moe_text</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>248320</td><td>248320</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen3_5Moe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>71</td><td>18</td></tr><tr class=""><td>RMSNorm</td><td>21</td><td>7</td></tr><tr class=""><td>SiLUActivation</td><td>16</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>8</td><td>2</td></tr><tr class=""><td>Experts</td><td>8</td><td>2</td></tr><tr class=""><td>MLP</td><td>8</td><td>2</td></tr><tr class=""><td>SparseMoeBlock</td><td>8</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>8</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>6</td><td>1</td></tr><tr class=""><td>GatedDeltaNet</td><td>6</td><td>1</td></tr><tr class=""><td>RMSNormGated</td><td>6</td><td>1</td></tr><tr class=""><td>Attention</td><td>2</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">8 fields changed, 37 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>layer_types</td><td>8 entries: 6× linear_attention, 2× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>moe_intermediate_size</td><td>1024</td><td>32</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_experts</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>8</td><td>2</td></tr><tr><td>shared_expert_intermediate_size</td><td>1024</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3.8-1.0B-A0.6B-mic.json b/shrink/Qwen3.8-1.0B-A0.6B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..dbf8dd208401ef248226e6df22193717c93dbac7
--- /dev/null
+++ b/shrink/Qwen3.8-1.0B-A0.6B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3.8-1.0B-A0.6B-mic@9633d99d303ac2051fd49702bc4781e0f930cfcc", "findings": []}
\ No newline at end of file
diff --git a/shrink/Qwen3.8-Flash-Next-mic.html b/shrink/Qwen3.8-Flash-Next-mic.html
index b87488c936c4a924eb552c2ba6eb5a272f35613c..552ddde602765ae0d2e1687207680e8a6524ffae 100644
--- a/shrink/Qwen3.8-Flash-Next-mic.html
+++ b/shrink/Qwen3.8-Flash-Next-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Qwen/Qwen3.8-Flash-Next: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Qwen3.8-Flash-Next-mic@9b40023b0010df89774011881d27eb6165fd2ca9 ← Qwen/Qwen3.8-Flash-Next@de4b8e4d43b917e7706784d8bb445c9af86a3540 · Qwen4ExpForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Qwen/Qwen3.8-Flash-Next</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Qwen3.8-Flash-Next-mic/tree/9b40023b0010df89774011881d27eb6165fd2ca9" title="huggingface-mic/Qwen3.8-Flash-Next-mic@9b40023b0010df89774011881d27eb6165fd2ca9">huggingface-mic/Qwen3.8-Flash-Next-mic@9b40023b0010</a> ← <a href="https://huggingface.co/Qwen/Qwen3.8-Flash-Next/tree/de4b8e4d43b917e7706784d8bb445c9af86a3540" title="Qwen/Qwen3.8-Flash-Next@de4b8e4d43b917e7706784d8bb445c9af86a3540">Qwen/Qwen3.8-Flash-Next@de4b8e4d43b9</a> · Qwen4ExpForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">177.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">359.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.visual.blocks</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><div class="row"><span class="lab">Fixture · 27</span><div class="cells"><i class="c0" title="layer 0: VisionBlock"></i><i class="c0" title="layer 1: VisionBlock"></i><i class="c0" title="layer 2: VisionBlock"></i><i class="c0" title="layer 3: VisionBlock"></i><i class="c0" title="layer 4: VisionBlock"></i><i class="c0" title="layer 5: VisionBlock"></i><i class="c0" title="layer 6: VisionBlock"></i><i class="c0" title="layer 7: VisionBlock"></i><i class="c0" title="layer 8: VisionBlock"></i><i class="c0" title="layer 9: VisionBlock"></i><i class="c0" title="layer 10: VisionBlock"></i><i class="c0" title="layer 11: VisionBlock"></i><i class="c0" title="layer 12: VisionBlock"></i><i class="c0" title="layer 13: VisionBlock"></i><i class="c0" title="layer 14: VisionBlock"></i><i class="c0" title="layer 15: VisionBlock"></i><i class="c0" title="layer 16: VisionBlock"></i><i class="c0" title="layer 17: VisionBlock"></i><i class="c0" title="layer 18: VisionBlock"></i><i class="c0" title="layer 19: VisionBlock"></i><i class="c0" title="layer 20: VisionBlock"></i><i class="c0" title="layer 21: VisionBlock"></i><i class="c0" title="layer 22: VisionBlock"></i><i class="c0" title="layer 23: VisionBlock"></i><i class="c0" title="layer 24: VisionBlock"></i><i class="c0" title="layer 25: VisionBlock"></i><i class="c0" title="layer 26: VisionBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionBlock</td><td>27</td><td>27</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c1" title="layer 1: linear_attention, Conv1d, Embedding …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 3: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 4: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 7: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 8: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 9: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 10: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 11: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 12: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 13: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 14: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 15: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 16: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 17: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 18: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 19: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 20: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 21: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 22: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 23: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 24: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 25: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 26: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 27: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 28: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 29: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 30: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 31: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 32: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 33: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 34: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 35: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 36: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 37: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 38: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 39: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 40: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 41: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 42: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 43: indexed_attention, TextAttention, TextQSAIndexer"></i><i class="c0" title="layer 44: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 45: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c0" title="layer 46: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 47: indexed_attention, TextAttention, TextQSAIndexer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c1" title="layer 1: linear_attention, Conv1d, Embedding …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, TextGatedDeltaNet …"></i><i class="c2" title="layer 3: indexed_attention, TextAttention, TextQSAIndexer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, TextGatedDeltaNet …</td><td>35</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>linear_attention, Conv1d, Embedding …</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c2"></i>indexed_attention, TextAttention, TextQSAIndexer</td><td>12</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.990265" title="model.language_model.layers: 175.7B (99.0%)"></i><i class="c1" style="flex:0.003621" title="model.language_model: 642.3M (0.4%)"></i><i class="c2" style="flex:0.003584" title="lm_head: 635.7M (0.4%)"></i><i class="c3" style="flex:0.002320" title="model.visual.blocks: 411.5M (0.2%)"></i><i class="c4" style="flex:0.000211" title="model.visual: 37.5M (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.907634" title="model.language_model.layers: 326.1M (90.8%)"></i><i class="c1" style="flex:0.044697" title="model.language_model: 16.1M (4.5%)"></i><i class="c2" style="flex:0.044240" title="lm_head: 15.9M (4.4%)"></i><i class="c3" style="flex:0.002516" title="model.visual.blocks: 903.7K (0.3%)"></i><i class="c4" style="flex:0.000914" title="model.visual: 328.2K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>175.7B</td><td>99.0%</td><td>326.1M</td><td>90.8%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>642.3M</td><td>0.4%</td><td>16.1M</td><td>4.5%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>635.7M</td><td>0.4%</td><td>15.9M</td><td>4.4%</td></tr><tr><td><i class="c3"></i>model.visual.blocks</td><td>411.5M</td><td>0.2%</td><td>903.7K</td><td>0.3%</td></tr><tr><td><i class="c4"></i>model.visual</td><td>37.5M</td><td>0.0%</td><td>328.2K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>qwen4_exp</td><td>qwen4_exp</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (24 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-10 of 512 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen4_exp_text</td><td>qwen4_exp_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>248320</td><td>248320</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>qwen4_exp_vision</td><td>qwen4_exp_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Qwen4Exp omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>835</td><td>175</td></tr><tr class=""><td>TextRMSNorm</td><td>148</td><td>16</td></tr><tr class=""><td>TextGatedResidual</td><td>97</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>96</td><td>8</td></tr><tr class=""><td>LayerNorm</td><td>55</td><td>55</td></tr><tr class=""><td>TextDecoderLayer</td><td>48</td><td>4</td></tr><tr class=""><td>TextExperts</td><td>48</td><td>4</td></tr><tr class=""><td>TextMLP</td><td>48</td><td>4</td></tr><tr class=""><td>TextSparseMoeBlock</td><td>48</td><td>4</td></tr><tr class=""><td>TextTopKRouter</td><td>48</td><td>4</td></tr><tr class=""><td>Conv1d</td><td>37</td><td>4</td></tr><tr class=""><td>TextGatedDeltaNet</td><td>36</td><td>3</td></tr><tr class=""><td>TextRMSNormGated</td><td>36</td><td>3</td></tr><tr class=""><td>GELUTanh</td><td>27</td><td>27</td></tr><tr class=""><td>VisionAttention</td><td>27</td><td>27</td></tr><tr class=""><td>VisionBlock</td><td>27</td><td>27</td></tr><tr class=""><td>VisionMLP</td><td>27</td><td>27</td></tr><tr class=""><td>TextAttention</td><td>12</td><td>1</td></tr><tr class=""><td>TextQSAIndexer</td><td>12</td><td>1</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv3d</td><td>1</td><td>1</td></tr><tr class=""><td>GELU</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextNGramEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextPLELayer</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchMerger</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">15 fields changed, 86 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2560</td><td>64</td></tr><tr><td>text_config.layer_types</td><td>48 entries: 36× linear_attention, 12× indexed_attention</td><td>[&quot;linear_attention&quot;, &quot;linear_attention&quot;, &quot;linear_attention&quot;, &quot;indexed_attenti…</td></tr><tr><td>text_config.moe_intermediate_size</td><td>640</td><td>32</td></tr><tr><td>text_config.num_attention_heads</td><td>24</td><td>4</td></tr><tr><td>text_config.num_experts</td><td>512</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>10</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>48</td><td>4</td></tr><tr><td>text_config.ple_embed_dim</td><td>2560</td><td>16</td></tr><tr><td>text_config.rope_parameters</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [11, 11, 10], &quot;partial_rotary_fa…</td><td>{&quot;mrope_interleaved&quot;: true, &quot;mrope_section&quot;: [3, 3, 2], &quot;partial_rotary_facto…</td></tr><tr><td>text_config.shared_expert_intermediate_size</td><td>640</td><td>128</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.out_hidden_size</td><td>2560</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Qwen3.8-Flash-Next-mic.json b/shrink/Qwen3.8-Flash-Next-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5de55eadfe9db6fe3d55b544b98c37322ea77f5
--- /dev/null
+++ b/shrink/Qwen3.8-Flash-Next-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Qwen3.8-Flash-Next-mic@9b40023b0010df89774011881d27eb6165fd2ca9", "findings": []}
\ No newline at end of file
diff --git a/shrink/Seed-OSS-36B-Instruct-mic.html b/shrink/Seed-OSS-36B-Instruct-mic.html
index 91deea27e15834e150d01acba6953c054c6c2b6e..009380b3160bd5386f0d53a0641073f28cd79cf7 100644
--- a/shrink/Seed-OSS-36B-Instruct-mic.html
+++ b/shrink/Seed-OSS-36B-Instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ByteDance-Seed/Seed-OSS-36B-Instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Seed-OSS-36B-Instruct-mic@5a7c613134e50c88bce20805ea4b02458007b844 ← ByteDance-Seed/Seed-OSS-36B-Instruct@497f1dca95ebdec98e41d517b9f060ee753c902f · SeedOssForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">36.2B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 64</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i><i class="c0" title="layer 40: DecoderLayer"></i><i class="c0" title="layer 41: DecoderLayer"></i><i class="c0" title="layer 42: DecoderLayer"></i><i class="c0" title="layer 43: DecoderLayer"></i><i class="c0" title="layer 44: DecoderLayer"></i><i class="c0" title="layer 45: DecoderLayer"></i><i class="c0" title="layer 46: DecoderLayer"></i><i class="c0" title="layer 47: DecoderLayer"></i><i class="c0" title="layer 48: DecoderLayer"></i><i class="c0" title="layer 49: DecoderLayer"></i><i class="c0" title="layer 50: DecoderLayer"></i><i class="c0" title="layer 51: DecoderLayer"></i><i class="c0" title="layer 52: DecoderLayer"></i><i class="c0" title="layer 53: DecoderLayer"></i><i class="c0" title="layer 54: DecoderLayer"></i><i class="c0" title="layer 55: DecoderLayer"></i><i class="c0" title="layer 56: DecoderLayer"></i><i class="c0" title="layer 57: DecoderLayer"></i><i class="c0" title="layer 58: DecoderLayer"></i><i class="c0" title="layer 59: DecoderLayer"></i><i class="c0" title="layer 60: DecoderLayer"></i><i class="c0" title="layer 61: DecoderLayer"></i><i class="c0" title="layer 62: DecoderLayer"></i><i class="c0" title="layer 63: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>64</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.956057" title="model.layers: 34.6B (95.6%)"></i><i class="c1" style="flex:0.021972" title="lm_head: 794.3M (2.2%)"></i><i class="c2" style="flex:0.021972" title="model.embed_tokens: 794.3M (2.2%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 5.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003725" title="model.layers: 74.2K (0.4%)"></i><i class="c1" style="flex:0.498136" title="lm_head: 9.9M (49.8%)"></i><i class="c2" style="flex:0.498136" title="model.embed_tokens: 9.9M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>34.6B</td><td>95.6%</td><td>74.2K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>794.3M</td><td>2.2%</td><td>9.9M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>794.3M</td><td>2.2%</td><td>9.9M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>5.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (80 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>seed_oss</td><td>seed_oss</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>155136</td><td>155136</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>524288</td><td>524288</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix SeedOss omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>449</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>129</td><td>5</td></tr><tr class=""><td>Attention</td><td>64</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>64</td><td>2</td></tr><tr class=""><td>MLP</td><td>64</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>64</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>5120</td><td>64</td></tr><tr><td>intermediate_size</td><td>27648</td><td>128</td></tr><tr><td>num_attention_heads</td><td>80</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>64</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ByteDance-Seed/Seed-OSS-36B-Instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Seed-OSS-36B-Instruct-mic/tree/5a7c613134e50c88bce20805ea4b02458007b844" title="huggingface-mic/Seed-OSS-36B-Instruct-mic@5a7c613134e50c88bce20805ea4b02458007b844">huggingface-mic/Seed-OSS-36B-Instruct-mic@5a7c613134e5</a> ← <a href="https://huggingface.co/ByteDance-Seed/Seed-OSS-36B-Instruct/tree/497f1dca95ebdec98e41d517b9f060ee753c902f" title="ByteDance-Seed/Seed-OSS-36B-Instruct@497f1dca95ebdec98e41d517b9f060ee753c902f">ByteDance-Seed/Seed-OSS-36B-Instruct@497f1dca95eb</a> · SeedOssForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">36.2B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 64</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i><i class="c0" title="layer 40: DecoderLayer"></i><i class="c0" title="layer 41: DecoderLayer"></i><i class="c0" title="layer 42: DecoderLayer"></i><i class="c0" title="layer 43: DecoderLayer"></i><i class="c0" title="layer 44: DecoderLayer"></i><i class="c0" title="layer 45: DecoderLayer"></i><i class="c0" title="layer 46: DecoderLayer"></i><i class="c0" title="layer 47: DecoderLayer"></i><i class="c0" title="layer 48: DecoderLayer"></i><i class="c0" title="layer 49: DecoderLayer"></i><i class="c0" title="layer 50: DecoderLayer"></i><i class="c0" title="layer 51: DecoderLayer"></i><i class="c0" title="layer 52: DecoderLayer"></i><i class="c0" title="layer 53: DecoderLayer"></i><i class="c0" title="layer 54: DecoderLayer"></i><i class="c0" title="layer 55: DecoderLayer"></i><i class="c0" title="layer 56: DecoderLayer"></i><i class="c0" title="layer 57: DecoderLayer"></i><i class="c0" title="layer 58: DecoderLayer"></i><i class="c0" title="layer 59: DecoderLayer"></i><i class="c0" title="layer 60: DecoderLayer"></i><i class="c0" title="layer 61: DecoderLayer"></i><i class="c0" title="layer 62: DecoderLayer"></i><i class="c0" title="layer 63: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>64</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.956057" title="model.layers: 34.6B (95.6%)"></i><i class="c1" style="flex:0.021972" title="model.embed_tokens: 794.3M (2.2%)"></i><i class="c2" style="flex:0.021972" title="lm_head: 794.3M (2.2%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 5.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003725" title="model.layers: 74.2K (0.4%)"></i><i class="c1" style="flex:0.498136" title="model.embed_tokens: 9.9M (49.8%)"></i><i class="c2" style="flex:0.498136" title="lm_head: 9.9M (49.8%)"></i><i class="c3" style="flex:0.000003" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>34.6B</td><td>95.6%</td><td>74.2K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>794.3M</td><td>2.2%</td><td>9.9M</td><td>49.8%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>794.3M</td><td>2.2%</td><td>9.9M</td><td>49.8%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>5.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (80 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>seed_oss</td><td>seed_oss</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>155136</td><td>155136</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>524288</td><td>524288</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix SeedOss omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>449</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>129</td><td>5</td></tr><tr class=""><td>Attention</td><td>64</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>64</td><td>2</td></tr><tr class=""><td>MLP</td><td>64</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>64</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>128</td><td>16</td></tr><tr><td>hidden_size</td><td>5120</td><td>64</td></tr><tr><td>intermediate_size</td><td>27648</td><td>128</td></tr><tr><td>num_attention_heads</td><td>80</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>64</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Seed-OSS-36B-Instruct-mic.json b/shrink/Seed-OSS-36B-Instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..3f345a46b32e225699cce94310824cb53cf47c23
--- /dev/null
+++ b/shrink/Seed-OSS-36B-Instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Seed-OSS-36B-Instruct-mic@5a7c613134e50c88bce20805ea4b02458007b844", "findings": []}
\ No newline at end of file
diff --git a/shrink/SmolLM3-3B-mic.html b/shrink/SmolLM3-3B-mic.html
index c63f06f8f36abb976c3cef22ba5239aea3c44edf..891e020324a48d12c56f524e9407e47a50781307 100644
--- a/shrink/SmolLM3-3B-mic.html
+++ b/shrink/SmolLM3-3B-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>HuggingFaceTB/SmolLM3-3B: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/SmolLM3-3B-mic@717fbb55a2a1e454d1de0cd85e1c944183b0cbd7 ← HuggingFaceTB/SmolLM3-3B@a07cc9a04f16550a088caea529712d1d335b0ac1 · SmolLM3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>HuggingFaceTB/SmolLM3-3B</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/SmolLM3-3B-mic/tree/717fbb55a2a1e454d1de0cd85e1c944183b0cbd7" title="huggingface-mic/SmolLM3-3B-mic@717fbb55a2a1e454d1de0cd85e1c944183b0cbd7">huggingface-mic/SmolLM3-3B-mic@717fbb55a2a1</a> ← <a href="https://huggingface.co/HuggingFaceTB/SmolLM3-3B/tree/a07cc9a04f16550a088caea529712d1d335b0ac1" title="HuggingFaceTB/SmolLM3-3B@a07cc9a04f16550a088caea529712d1d335b0ac1">HuggingFaceTB/SmolLM3-3B@a07cc9a04f16</a> · SmolLM3ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">3.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">35.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>model: max_position_embeddings 65536 became 512</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 36</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>36</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.914581" title="model.layers: 2.8B (91.5%)"></i><i class="c1" style="flex:0.085418" title="model.embed_tokens: 262.7M (8.5%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.067093" title="model.layers: 2.4M (6.7%)"></i><i class="c1" style="flex:0.932900" title="model.embed_tokens: 32.8M (93.3%)"></i><i class="c2" style="flex:0.000007" title="model.norm: 256 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>2.8B</td><td>91.5%</td><td>2.4M</td><td>6.7%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>262.7M</td><td>8.5%</td><td>32.8M</td><td>93.3%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>256</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>smollm3</td><td>smollm3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>128256</td><td>128256</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class="bad"><td>model</td><td>max_position_embeddings</td><td>65536</td><td>512</td><td>⚠ changed</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix SmolLM3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>253</td><td>29</td></tr><tr class=""><td>RMSNorm</td><td>73</td><td>9</td></tr><tr class=""><td>SiLUActivation</td><td>36</td><td>4</td></tr><tr class=""><td>Attention</td><td>36</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>36</td><td>4</td></tr><tr class=""><td>MLP</td><td>36</td><td>4</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>256</td></tr><tr><td>intermediate_size</td><td>11008</td><td>512</td></tr><tr><td>layer_types</td><td>36 entries: 36× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;, &quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>max_position_embeddings</td><td>65536</td><td>512</td></tr><tr><td>max_window_layers</td><td>28</td><td>3</td></tr><tr><td>no_rope_layers</td><td>36 entries: 27× 1, 9× 0</td><td>[1, 1, 1, 0]</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>36</td><td>4</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/SmolLM3-3B-mic.json b/shrink/SmolLM3-3B-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..b2b28c3ab4923987a4dcb455d34b63301f2ed7fb
--- /dev/null
+++ b/shrink/SmolLM3-3B-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/SmolLM3-3B-mic@717fbb55a2a1e454d1de0cd85e1c944183b0cbd7", "findings": ["model: max_position_embeddings 65536 became 512"]}
\ No newline at end of file
diff --git a/shrink/Unlimited-OCR-mic.html b/shrink/Unlimited-OCR-mic.html
index e6ac2290e86344ae880f772365a6e98c7d2c9d35..9361bfd2480f1eb9d00909adf64ea78cc5d64d70 100644
--- a/shrink/Unlimited-OCR-mic.html
+++ b/shrink/Unlimited-OCR-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>baidu/Unlimited-OCR: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Unlimited-OCR-mic@7b86520360914f84047e85b7d0b11109baf6d934 ← baidu/Unlimited-OCR@8249d20b2822a7d8a4c22c8d8c9069b5ac36bda3 · UnlimitedOcrForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>baidu/Unlimited-OCR</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Unlimited-OCR-mic/tree/7b86520360914f84047e85b7d0b11109baf6d934" title="huggingface-mic/Unlimited-OCR-mic@7b86520360914f84047e85b7d0b11109baf6d934">huggingface-mic/Unlimited-OCR-mic@7b8652036091</a> ← <a href="https://huggingface.co/baidu/Unlimited-OCR/tree/8249d20b2822a7d8a4c22c8d8c9069b5ac36bda3" title="baidu/Unlimited-OCR@8249d20b2822a7d8a4c22c8d8c9069b5ac36bda3">baidu/Unlimited-OCR@8249d20b2822</a> · UnlimitedOcrForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">3.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">18.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.sam_encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: SamMLPBlock"></i><i class="c0" title="layer 1: SamMLPBlock"></i><i class="c0" title="layer 2: SamMLPBlock"></i><i class="c0" title="layer 3: SamMLPBlock"></i><i class="c0" title="layer 4: SamMLPBlock"></i><i class="c0" title="layer 5: SamMLPBlock"></i><i class="c0" title="layer 6: SamMLPBlock"></i><i class="c0" title="layer 7: SamMLPBlock"></i><i class="c0" title="layer 8: SamMLPBlock"></i><i class="c0" title="layer 9: SamMLPBlock"></i><i class="c0" title="layer 10: SamMLPBlock"></i><i class="c0" title="layer 11: SamMLPBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: SamMLPBlock"></i><i class="c0" title="layer 1: SamMLPBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>SamMLPBlock</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.vision_tower.vision_encoder.encoder.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i><i class="c0" title="layer 12: EncoderLayer"></i><i class="c0" title="layer 13: EncoderLayer"></i><i class="c0" title="layer 14: EncoderLayer"></i><i class="c0" title="layer 15: EncoderLayer"></i><i class="c0" title="layer 16: EncoderLayer"></i><i class="c0" title="layer 17: EncoderLayer"></i><i class="c0" title="layer 18: EncoderLayer"></i><i class="c0" title="layer 19: EncoderLayer"></i><i class="c0" title="layer 20: EncoderLayer"></i><i class="c0" title="layer 21: EncoderLayer"></i><i class="c0" title="layer 22: EncoderLayer"></i><i class="c0" title="layer 23: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: no TextExperts"></i><i class="c1" title="layer 1: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 2: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 3: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 4: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 5: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 6: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 7: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 8: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 9: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 10: TextExperts, TextMoe, TextTopkRouter"></i><i class="c1" title="layer 11: TextExperts, TextMoe, TextTopkRouter"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: no TextExperts"></i><i class="c1" title="layer 1: TextExperts, TextMoe, TextTopkRouter"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no TextExperts</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>TextExperts, TextMoe, TextTopkRouter</td><td>11</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.780624" title="model.language_model.layers: 2.6B (78.1%)"></i><i class="c1" style="flex:0.090634" title="model.vision_tower.vision_encoder.encoder.layers: 302.3M (9.1%)"></i><i class="c2" style="flex:0.049612" title="model.language_model: 165.5M (5.0%)"></i><i class="c3" style="flex:0.049611" title="lm_head: 165.5M (5.0%)"></i><i class="c4" style="flex:0.025528" title="model.vision_tower.sam_encoder.layers: 85.1M (2.6%)"></i><i class="c5" style="flex:0.003204" title="model.vision_tower: 10.7M (0.3%)"></i><i class="c6" style="flex:0.000786" title="model.multi_modal_projector: 2.6M (0.1%)"></i><i class="c7" style="flex:0.000001" title="1 others: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.005100" title="model.language_model.layers: 94.7K (0.5%)"></i><i class="c1" style="flex:0.003604" title="model.vision_tower.vision_encoder.encoder.layers: 66.9K (0.4%)"></i><i class="c2" style="flex:0.445471" title="model.language_model: 8.3M (44.5%)"></i><i class="c3" style="flex:0.445467" title="lm_head: 8.3M (44.5%)"></i><i class="c4" style="flex:0.044592" title="model.vision_tower.sam_encoder.layers: 828.2K (4.5%)"></i><i class="c5" style="flex:0.055315" title="model.vision_tower: 1.0M (5.5%)"></i><i class="c6" style="flex:0.000445" title="model.multi_modal_projector: 8.3K (0.0%)"></i><i class="c7" style="flex:0.000007" title="1 others: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>2.6B</td><td>78.1%</td><td>94.7K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>model.vision_tower.vision_encoder.encoder.layers</td><td>302.3M</td><td>9.1%</td><td>66.9K</td><td>0.4%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>165.5M</td><td>5.0%</td><td>8.3M</td><td>44.5%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>165.5M</td><td>5.0%</td><td>8.3M</td><td>44.5%</td></tr><tr><td><i class="c4"></i>model.vision_tower.sam_encoder.layers</td><td>85.1M</td><td>2.6%</td><td>828.2K</td><td>4.5%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>10.7M</td><td>0.3%</td><td>1.0M</td><td>5.5%</td></tr><tr><td><i class="c6"></i>model.multi_modal_projector</td><td>2.6M</td><td>0.1%</td><td>8.3K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>1 others</td><td>2.6K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MHA (10 heads, 10 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-6 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>unlimited_ocr</td><td>unlimited_ocr</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>129280</td><td>129280</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>unlimited_ocr_vision</td><td>unlimited_ocr_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>1024</td><td>1024</td><td>✓ kept</td></tr><tr class=""><td>vision_config.sam_config</td><td>model_type</td><td>unlimited_ocr_sam_vision_model</td><td>unlimited_ocr_sam_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config.sam_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config.sam_config</td><td>image_size</td><td>1024</td><td>1024</td><td>✓ kept</td></tr><tr class=""><td>vision_config.sam_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config.sam_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>vision_config.encoder_config</td><td>model_type</td><td>unlimited_ocr_vision_encoder</td><td>unlimited_ocr_vision_encoder</td><td>✓ kept</td></tr><tr class=""><td>vision_config.encoder_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config.encoder_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config.encoder_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config.encoder_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MHA (10 heads, 10 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>MoE routing</td><td>top-6 of 64 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>unlimited_ocr_text</td><td>unlimited_ocr_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>129280</td><td>129280</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr><tr class=""><td>projector_config</td><td>model_type</td><td>mlp_projector</td><td>mlp_projector</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix UnlimitedOcr omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>278</td><td>36</td></tr><tr class=""><td>LayerNorm</td><td>73</td><td>9</td></tr><tr class=""><td>TextRMSNorm</td><td>25</td><td>5</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>23</td><td>3</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>SamMLPBlock</td><td>12</td><td>2</td></tr><tr class=""><td>SamVisionLayer</td><td>12</td><td>2</td></tr><tr class=""><td>SamVisionSdpaAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>TextMLP</td><td>12</td><td>2</td></tr><tr class=""><td>TextExperts</td><td>11</td><td>1</td></tr><tr class=""><td>TextMoe</td><td>11</td><td>1</td></tr><tr class=""><td>TextTopkRouter</td><td>11</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>5</td><td>5</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>SamLayerNorm</td><td>2</td><td>2</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SamPatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>SamVisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>SamVisionNeck</td><td>1</td><td>1</td></tr><tr class=""><td>SamVisionProj</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">30 fields changed, 128 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1280</td><td>64</td></tr><tr><td>intermediate_size</td><td>6848</td><td>128</td></tr><tr><td>moe_intermediate_size</td><td>896</td><td>32</td></tr><tr><td>n_routed_experts</td><td>64</td><td>4</td></tr><tr><td>num_attention_heads</td><td>10</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>6</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>10</td><td>4</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1280</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>6848</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>12 entries: 12× reference_sliding_attention</td><td>[&quot;reference_sliding_attention&quot;, &quot;reference_sliding_attention&quot;]</td></tr><tr><td>text_config.mlp_layer_types</td><td>12 entries: 11× sparse, 1× dense</td><td>[&quot;dense&quot;, &quot;sparse&quot;]</td></tr><tr><td>text_config.moe_intermediate_size</td><td>896</td><td>32</td></tr><tr><td>text_config.n_routed_experts</td><td>64</td><td>4</td></tr><tr><td>text_config.num_attention_heads</td><td>10</td><td>4</td></tr><tr><td>text_config.num_experts_per_tok</td><td>6</td><td>2</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>10</td><td>4</td></tr><tr><td>text_config.v_head_dim</td><td>128</td><td>16</td></tr><tr><td>v_head_dim</td><td>128</td><td>16</td></tr><tr><td>vision_config.encoder_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.encoder_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.encoder_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.encoder_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.sam_config.downsample_channels</td><td>[512, 1024]</td><td>[32, 64]</td></tr><tr><td>vision_config.sam_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.sam_config.initializer_range</td><td>1e-10</td><td>0.02</td></tr><tr><td>vision_config.sam_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.sam_config.num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Unlimited-OCR-mic.json b/shrink/Unlimited-OCR-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..bde12c1dacd092792cd2acc0272838377a566705
--- /dev/null
+++ b/shrink/Unlimited-OCR-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Unlimited-OCR-mic@7b86520360914f84047e85b7d0b11109baf6d934", "findings": []}
\ No newline at end of file
diff --git a/shrink/VibeVoice-1.5B-hf-mic.html b/shrink/VibeVoice-1.5B-hf-mic.html
index fdf306855c6ed028a3044a9c867c32380a089a2d..4180ac42f1106b933ab274d0aef7a490dd32436c 100644
--- a/shrink/VibeVoice-1.5B-hf-mic.html
+++ b/shrink/VibeVoice-1.5B-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>vibevoice/VibeVoice-1.5B-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/VibeVoice-1.5B-hf-mic@a651225b62eeb6b3aa47c4f0600111b0055ab257 ← vibevoice/VibeVoice-1.5B-hf@d4cd82eb4371bf3e3d088e1ea7fa4eefb60449bf · VibeVoiceForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>vibevoice/VibeVoice-1.5B-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/VibeVoice-1.5B-hf-mic/tree/a651225b62eeb6b3aa47c4f0600111b0055ab257" title="huggingface-mic/VibeVoice-1.5B-hf-mic@a651225b62eeb6b3aa47c4f0600111b0055ab257">huggingface-mic/VibeVoice-1.5B-hf-mic@a651225b62ee</a> ← <a href="https://huggingface.co/vibevoice/VibeVoice-1.5B-hf/tree/d4cd82eb4371bf3e3d088e1ea7fa4eefb60449bf" title="vibevoice/VibeVoice-1.5B-hf@d4cd82eb4371bf3e3d088e1ea7fa4eefb60449bf">vibevoice/VibeVoice-1.5B-hf@d4cd82eb4371</a> · VibeVoiceForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">268.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">8</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.audio_tower.encoder.stem.stage</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 1: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 2: AcousticTokenizerConvNext1dLayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerConvNext1dLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AcousticTokenizerConvNext1dLayer</td><td>3</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.audio_tower.encoder.conv_layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 1: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 2: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 3: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 4: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 5: AcousticTokenizerEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 6</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 1: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 2: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 3: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 4: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 5: AcousticTokenizerEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AcousticTokenizerEncoderLayer</td><td>6</td><td>6</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.audio_tower.decoder.stem.stage</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 1: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 2: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 3: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 4: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 5: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 6: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 7: AcousticTokenizerConvNext1dLayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerConvNext1dLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AcousticTokenizerConvNext1dLayer</td><td>8</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.audio_tower.decoder.conv_layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 1: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 2: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 3: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 4: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 5: AcousticTokenizerDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 6</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 1: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 2: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 3: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 4: AcousticTokenizerDecoderLayer"></i><i class="c0" title="layer 5: AcousticTokenizerDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AcousticTokenizerDecoderLayer</td><td>6</td><td>6</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i><i class="c0" title="layer 2: Qwen2DecoderLayer"></i><i class="c0" title="layer 3: Qwen2DecoderLayer"></i><i class="c0" title="layer 4: Qwen2DecoderLayer"></i><i class="c0" title="layer 5: Qwen2DecoderLayer"></i><i class="c0" title="layer 6: Qwen2DecoderLayer"></i><i class="c0" title="layer 7: Qwen2DecoderLayer"></i><i class="c0" title="layer 8: Qwen2DecoderLayer"></i><i class="c0" title="layer 9: Qwen2DecoderLayer"></i><i class="c0" title="layer 10: Qwen2DecoderLayer"></i><i class="c0" title="layer 11: Qwen2DecoderLayer"></i><i class="c0" title="layer 12: Qwen2DecoderLayer"></i><i class="c0" title="layer 13: Qwen2DecoderLayer"></i><i class="c0" title="layer 14: Qwen2DecoderLayer"></i><i class="c0" title="layer 15: Qwen2DecoderLayer"></i><i class="c0" title="layer 16: Qwen2DecoderLayer"></i><i class="c0" title="layer 17: Qwen2DecoderLayer"></i><i class="c0" title="layer 18: Qwen2DecoderLayer"></i><i class="c0" title="layer 19: Qwen2DecoderLayer"></i><i class="c0" title="layer 20: Qwen2DecoderLayer"></i><i class="c0" title="layer 21: Qwen2DecoderLayer"></i><i class="c0" title="layer 22: Qwen2DecoderLayer"></i><i class="c0" title="layer 23: Qwen2DecoderLayer"></i><i class="c0" title="layer 24: Qwen2DecoderLayer"></i><i class="c0" title="layer 25: Qwen2DecoderLayer"></i><i class="c0" title="layer 26: Qwen2DecoderLayer"></i><i class="c0" title="layer 27: Qwen2DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2DecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.semantic_tokenizer_encoder.stem.stage</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 1: AcousticTokenizerConvNext1dLayer"></i><i class="c0" title="layer 2: AcousticTokenizerConvNext1dLayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerConvNext1dLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AcousticTokenizerConvNext1dLayer</td><td>3</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.semantic_tokenizer_encoder.conv_layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 1: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 2: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 3: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 4: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 5: AcousticTokenizerEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 6</span><div class="cells"><i class="c0" title="layer 0: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 1: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 2: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 3: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 4: AcousticTokenizerEncoderLayer"></i><i class="c0" title="layer 5: AcousticTokenizerEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AcousticTokenizerEncoderLayer</td><td>6</td><td>6</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.diffusion_head.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>4</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.484589" title="model.language_model.layers: 1.3B (48.5%)"></i><i class="c1" style="flex:0.126756" title="model.audio_tower.encoder.conv_layers: 342.8M (12.7%)"></i><i class="c2" style="flex:0.126756" title="model.semantic_tokenizer_encoder.conv_layers: 342.8M (12.7%)"></i><i class="c3" style="flex:0.099376" title="model.audio_tower.decoder.stem.stage: 268.7M (9.9%)"></i><i class="c4" style="flex:0.086307" title="model.language_model: 233.4M (8.6%)"></i><i class="c5" style="flex:0.041883" title="model.diffusion_head.layers: 113.3M (4.2%)"></i><i class="c6" style="flex:0.027390" title="model.audio_tower.decoder.conv_layers: 74.1M (2.7%)"></i><i class="c7" style="flex:0.006943" title="8 others: 18.8M (0.7%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.000277" title="model.language_model.layers: 74.2K (0.0%)"></i><i class="c1" style="flex:0.317546" title="model.audio_tower.encoder.conv_layers: 85.2M (31.8%)"></i><i class="c2" style="flex:0.317546" title="model.semantic_tokenizer_encoder.conv_layers: 85.2M (31.8%)"></i><i class="c3" style="flex:0.125172" title="model.audio_tower.decoder.stem.stage: 33.6M (12.5%)"></i><i class="c4" style="flex:0.036237" title="model.language_model: 9.7M (3.6%)"></i><i class="c5" style="flex:0.000275" title="model.diffusion_head.layers: 73.9K (0.0%)"></i><i class="c6" style="flex:0.192399" title="model.audio_tower.decoder.conv_layers: 51.6M (19.2%)"></i><i class="c7" style="flex:0.010548" title="8 others: 2.8M (1.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>1.3B</td><td>48.5%</td><td>74.2K</td><td>0.0%</td></tr><tr><td><i class="c1"></i>model.audio_tower.encoder.conv_layers</td><td>342.8M</td><td>12.7%</td><td>85.2M</td><td>31.8%</td></tr><tr><td><i class="c2"></i>model.semantic_tokenizer_encoder.conv_layers</td><td>342.8M</td><td>12.7%</td><td>85.2M</td><td>31.8%</td></tr><tr><td><i class="c3"></i>model.audio_tower.decoder.stem.stage</td><td>268.7M</td><td>9.9%</td><td>33.6M</td><td>12.5%</td></tr><tr><td><i class="c4"></i>model.language_model</td><td>233.4M</td><td>8.6%</td><td>9.7M</td><td>3.6%</td></tr><tr><td><i class="c5"></i>model.diffusion_head.layers</td><td>113.3M</td><td>4.2%</td><td>73.9K</td><td>0.0%</td></tr><tr><td><i class="c6"></i>model.audio_tower.decoder.conv_layers</td><td>74.1M</td><td>2.7%</td><td>51.6M</td><td>19.2%</td></tr><tr><td><i class="c7"></i>8 others</td><td>18.8M</td><td>0.7%</td><td>2.8M</td><td>1.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>vibevoice</td><td>vibevoice</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>vibevoice_acoustic_tokenizer</td><td>vibevoice_acoustic_tokenizer</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>semantic_model_config</td><td>model_type</td><td>vibevoice_acoustic_tokenizer_encoder</td><td>vibevoice_acoustic_tokenizer_encoder</td><td>✓ kept</td></tr><tr class=""><td>semantic_model_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (12 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2</td><td>qwen2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151936</td><td>151936</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>65536</td><td>65536</td><td>✓ kept</td></tr><tr class=""><td>diffusion_head_config</td><td>model_type</td><td>vibevoice_diffusion</td><td>vibevoice_diffusion</td><td>✓ kept</td></tr><tr class=""><td>diffusion_head_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix VibeVoice omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>379</td><td>75</td></tr><tr class=""><td>AcousticTokenizerRMSNorm</td><td>156</td><td>42</td></tr><tr class=""><td>Conv1d</td><td>96</td><td>39</td></tr><tr class=""><td>AcousticTokenizerCausalConv1d</td><td>96</td><td>39</td></tr><tr class=""><td>GELUActivation</td><td>78</td><td>21</td></tr><tr class=""><td>AcousticTokenizerConvNext1dLayer</td><td>78</td><td>21</td></tr><tr class=""><td>AcousticTokenizerFeedForward</td><td>78</td><td>21</td></tr><tr class=""><td>Qwen2RMSNorm</td><td>57</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>38</td><td>8</td></tr><tr class=""><td>Qwen2Attention</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen2DecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen2MLP</td><td>28</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>26</td><td>26</td></tr><tr class=""><td>AcousticTokenizerEncoderLayer</td><td>12</td><td>12</td></tr><tr class=""><td>ConvTranspose1d</td><td>6</td><td>6</td></tr><tr class=""><td>AcousticTokenizerCausalConvTranspose1d</td><td>6</td><td>6</td></tr><tr class=""><td>AcousticTokenizerDecoderLayer</td><td>6</td><td>6</td></tr><tr class=""><td>RMSNorm</td><td>6</td><td>4</td></tr><tr class=""><td>DiffusionHeadAdaLayerNorm</td><td>4</td><td>2</td></tr><tr class=""><td>MLP</td><td>4</td><td>2</td></tr><tr class=""><td>AcousticTokenizerEncoderModel</td><td>2</td><td>2</td></tr><tr class=""><td>AcousticTokenizerEncoderStem</td><td>2</td><td>2</td></tr><tr class=""><td>MultiModalProjector</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>AcousticTokenizerDecoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>AcousticTokenizerDecoderStem</td><td>1</td><td>1</td></tr><tr class=""><td>AcousticTokenizerModel</td><td>1</td><td>1</td></tr><tr class=""><td>DiffusionHead</td><td>1</td><td>1</td></tr><tr class=""><td>DiffusionHeadFinalLayer</td><td>1</td><td>1</td></tr><tr class=""><td>DiffusionHeadMLP</td><td>1</td><td>1</td></tr><tr class=""><td>DiffusionHeadSinusoidalEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 98 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.depths</td><td>7 entries: 6× 3, 1× 8</td><td>7 entries: 7× 1</td></tr><tr><td>diffusion_head_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>diffusion_head_config.intermediate_size</td><td>4608</td><td>128</td></tr><tr><td>diffusion_head_config.num_hidden_layers</td><td>4</td><td>2</td></tr><tr><td>semantic_model_config.depths</td><td>7 entries: 6× 3, 1× 8</td><td>7 entries: 7× 1</td></tr><tr><td>semantic_model_config.hidden_size</td><td>128</td><td>64</td></tr><tr><td>text_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>8960</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>28 entries: 28× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>28</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/VibeVoice-1.5B-hf-mic.json b/shrink/VibeVoice-1.5B-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..07ccbbac9dfd01b4beacf81827ffb05a0cb61d1f
--- /dev/null
+++ b/shrink/VibeVoice-1.5B-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/VibeVoice-1.5B-hf-mic@a651225b62eeb6b3aa47c4f0600111b0055ab257", "findings": []}
\ No newline at end of file
diff --git a/shrink/Voxtral-Mini-3B-2507-mic.html b/shrink/Voxtral-Mini-3B-2507-mic.html
index 6bedf6367706cb32ddc4ac87c066572cfaf2c034..8aa2fe3d126822d21fac8982673bbdcd19dacc12 100644
--- a/shrink/Voxtral-Mini-3B-2507-mic.html
+++ b/shrink/Voxtral-Mini-3B-2507-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>mistralai/Voxtral-Mini-3B-2507: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/Voxtral-Mini-3B-2507-mic@88954b6754a54ae7b39c94a53b59d3d54313fae9 ← mistralai/Voxtral-Mini-3B-2507@3060fe34b35ba5d44202ce9ff3c097642914f8f3 · VoxtralForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>mistralai/Voxtral-Mini-3B-2507</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/Voxtral-Mini-3B-2507-mic/tree/88954b6754a54ae7b39c94a53b59d3d54313fae9" title="huggingface-mic/Voxtral-Mini-3B-2507-mic@88954b6754a54ae7b39c94a53b59d3d54313fae9">huggingface-mic/Voxtral-Mini-3B-2507-mic@88954b6754a5</a> ← <a href="https://huggingface.co/mistralai/Voxtral-Mini-3B-2507/tree/3060fe34b35ba5d44202ce9ff3c097642914f8f3" title="mistralai/Voxtral-Mini-3B-2507@3060fe34b35ba5d44202ce9ff3c097642914f8f3">mistralai/Voxtral-Mini-3B-2507@3060fe34b35b</a> · VoxtralForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.7B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">17.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.audio_tower.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i><i class="c0" title="layer 12: EncoderLayer"></i><i class="c0" title="layer 13: EncoderLayer"></i><i class="c0" title="layer 14: EncoderLayer"></i><i class="c0" title="layer 15: EncoderLayer"></i><i class="c0" title="layer 16: EncoderLayer"></i><i class="c0" title="layer 17: EncoderLayer"></i><i class="c0" title="layer 18: EncoderLayer"></i><i class="c0" title="layer 19: EncoderLayer"></i><i class="c0" title="layer 20: EncoderLayer"></i><i class="c0" title="layer 21: EncoderLayer"></i><i class="c0" title="layer 22: EncoderLayer"></i><i class="c0" title="layer 23: EncoderLayer"></i><i class="c0" title="layer 24: EncoderLayer"></i><i class="c0" title="layer 25: EncoderLayer"></i><i class="c0" title="layer 26: EncoderLayer"></i><i class="c0" title="layer 27: EncoderLayer"></i><i class="c0" title="layer 28: EncoderLayer"></i><i class="c0" title="layer 29: EncoderLayer"></i><i class="c0" title="layer 30: EncoderLayer"></i><i class="c0" title="layer 31: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 30</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i><i class="c0" title="layer 2: LlamaDecoderLayer"></i><i class="c0" title="layer 3: LlamaDecoderLayer"></i><i class="c0" title="layer 4: LlamaDecoderLayer"></i><i class="c0" title="layer 5: LlamaDecoderLayer"></i><i class="c0" title="layer 6: LlamaDecoderLayer"></i><i class="c0" title="layer 7: LlamaDecoderLayer"></i><i class="c0" title="layer 8: LlamaDecoderLayer"></i><i class="c0" title="layer 9: LlamaDecoderLayer"></i><i class="c0" title="layer 10: LlamaDecoderLayer"></i><i class="c0" title="layer 11: LlamaDecoderLayer"></i><i class="c0" title="layer 12: LlamaDecoderLayer"></i><i class="c0" title="layer 13: LlamaDecoderLayer"></i><i class="c0" title="layer 14: LlamaDecoderLayer"></i><i class="c0" title="layer 15: LlamaDecoderLayer"></i><i class="c0" title="layer 16: LlamaDecoderLayer"></i><i class="c0" title="layer 17: LlamaDecoderLayer"></i><i class="c0" title="layer 18: LlamaDecoderLayer"></i><i class="c0" title="layer 19: LlamaDecoderLayer"></i><i class="c0" title="layer 20: LlamaDecoderLayer"></i><i class="c0" title="layer 21: LlamaDecoderLayer"></i><i class="c0" title="layer 22: LlamaDecoderLayer"></i><i class="c0" title="layer 23: LlamaDecoderLayer"></i><i class="c0" title="layer 24: LlamaDecoderLayer"></i><i class="c0" title="layer 25: LlamaDecoderLayer"></i><i class="c0" title="layer 26: LlamaDecoderLayer"></i><i class="c0" title="layer 27: LlamaDecoderLayer"></i><i class="c0" title="layer 28: LlamaDecoderLayer"></i><i class="c0" title="layer 29: LlamaDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>LlamaDecoderLayer</td><td>30</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.686194" title="model.language_model.layers: 3.2B (68.6%)"></i><i class="c1" style="flex:0.134645" title="model.audio_tower.layers: 629.6M (13.5%)"></i><i class="c2" style="flex:0.086106" title="model.language_model: 402.7M (8.6%)"></i><i class="c3" style="flex:0.086106" title="lm_head: 402.7M (8.6%)"></i><i class="c4" style="flex:0.005382" title="model.multi_modal_projector: 25.2M (0.5%)"></i><i class="c5" style="flex:0.001568" title="model.audio_tower: 7.3M (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004325" title="model.language_model.layers: 74.0K (0.4%)"></i><i class="c1" style="flex:0.005837" title="model.audio_tower.layers: 99.8K (0.6%)"></i><i class="c2" style="flex:0.490431" title="model.language_model: 8.4M (49.0%)"></i><i class="c3" style="flex:0.490427" title="lm_head: 8.4M (49.0%)"></i><i class="c4" style="flex:0.001197" title="model.multi_modal_projector: 20.5K (0.1%)"></i><i class="c5" style="flex:0.007783" title="model.audio_tower: 133.1K (0.8%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>3.2B</td><td>68.6%</td><td>74.0K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.audio_tower.layers</td><td>629.6M</td><td>13.5%</td><td>99.8K</td><td>0.6%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>402.7M</td><td>8.6%</td><td>8.4M</td><td>49.0%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>402.7M</td><td>8.6%</td><td>8.4M</td><td>49.0%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>25.2M</td><td>0.5%</td><td>20.5K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>model.audio_tower</td><td>7.3M</td><td>0.2%</td><td>133.1K</td><td>0.8%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>voxtral</td><td>voxtral</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>attention style</td><td>MHA (20 heads, 20 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>voxtral_encoder</td><td>voxtral_encoder</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>vocab_size</td><td>51866</td><td>51866</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>128</td><td>128</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>llama</td><td>llama</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Voxtral omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>405</td><td>29</td></tr><tr class=""><td>LayerNorm</td><td>65</td><td>5</td></tr><tr class=""><td>LlamaRMSNorm</td><td>61</td><td>5</td></tr><tr class=""><td>GELUActivation</td><td>33</td><td>3</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>LlamaAttention</td><td>30</td><td>2</td></tr><tr class=""><td>LlamaDecoderLayer</td><td>30</td><td>2</td></tr><tr class=""><td>LlamaMLP</td><td>30</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>30</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>AvgPool1d</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaModel</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 56 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>audio_config.hidden_size</td><td>1280</td><td>64</td></tr><tr><td>audio_config.intermediate_size</td><td>5120</td><td>256</td></tr><tr><td>audio_config.num_attention_heads</td><td>20</td><td>4</td></tr><tr><td>audio_config.num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>audio_config.num_key_value_heads</td><td>20</td><td>4</td></tr><tr><td>hidden_size</td><td>3072</td><td>64</td></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>3072</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>30</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/Voxtral-Mini-3B-2507-mic.json b/shrink/Voxtral-Mini-3B-2507-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..8f786dab42a38f0ffd6794641cebaa906c443bf1
--- /dev/null
+++ b/shrink/Voxtral-Mini-3B-2507-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/Voxtral-Mini-3B-2507-mic@88954b6754a54ae7b39c94a53b59d3d54313fae9", "findings": []}
\ No newline at end of file
diff --git a/shrink/albert-base-v2-mic.html b/shrink/albert-base-v2-mic.html
index d49bb5e130b7d36a9f9b9f6d9dcef8a3351d8832..ab7ce6ac34aacb4b3d942dd25b72d3c1265a2d8b 100644
--- a/shrink/albert-base-v2-mic.html
+++ b/shrink/albert-base-v2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>albert/albert-base-v2: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/albert-base-v2-mic@258e9f08fe272ff4b05d6823ad031e0ba86d9be6 ← albert/albert-base-v2@8e2f239c5f8a2c0f253781ca60135db913e5c80c · AlbertForMaskedLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>albert/albert-base-v2</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/albert-base-v2-mic/tree/258e9f08fe272ff4b05d6823ad031e0ba86d9be6" title="huggingface-mic/albert-base-v2-mic@258e9f08fe272ff4b05d6823ad031e0ba86d9be6">huggingface-mic/albert-base-v2-mic@258e9f08fe27</a> ← <a href="https://huggingface.co/albert/albert-base-v2/tree/8e2f239c5f8a2c0f253781ca60135db913e5c80c" title="albert/albert-base-v2@8e2f239c5f8a2c0f253781ca60135db913e5c80c">albert/albert-base-v2@8e2f239c5f8a</a> · AlbertForMaskedLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">11.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>albert.encoder.albert_layer_groups</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.631623" title="albert.encoder.albert_layer_groups: 7.1M (63.2%)"></i><i class="c1" style="flex:0.348081" title="albert.embeddings: 3.9M (34.8%)"></i><i class="c2" style="flex:0.008829" title="albert.encoder: 99.1K (0.9%)"></i><i class="c3" style="flex:0.008772" title="predictions.dense: 98.4K (0.9%)"></i><i class="c4" style="flex:0.002673" title="predictions: 30.0K (0.3%)"></i><i class="c5" style="flex:0.000023" title="predictions.LayerNorm: 256 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008397" title="albert.encoder.albert_layer_groups: 33.5K (0.8%)"></i><i class="c1" style="flex:0.979855" title="albert.embeddings: 3.9M (98.0%)"></i><i class="c2" style="flex:0.002071" title="albert.encoder: 8.3K (0.2%)"></i><i class="c3" style="flex:0.002087" title="predictions.dense: 8.3K (0.2%)"></i><i class="c4" style="flex:0.007526" title="predictions: 30.0K (0.8%)"></i><i class="c5" style="flex:0.000064" title="predictions.LayerNorm: 256 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>albert.encoder.albert_layer_groups</td><td>7.1M</td><td>63.2%</td><td>33.5K</td><td>0.8%</td></tr><tr><td><i class="c1"></i>albert.embeddings</td><td>3.9M</td><td>34.8%</td><td>3.9M</td><td>98.0%</td></tr><tr><td><i class="c2"></i>albert.encoder</td><td>99.1K</td><td>0.9%</td><td>8.3K</td><td>0.2%</td></tr><tr><td><i class="c3"></i>predictions.dense</td><td>98.4K</td><td>0.9%</td><td>8.3K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>predictions</td><td>30.0K</td><td>0.3%</td><td>30.0K</td><td>0.8%</td></tr><tr><td><i class="c5"></i>predictions.LayerNorm</td><td>256</td><td>0.0%</td><td>256</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>albert</td><td>albert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30000</td><td>30000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu_new</td><td>gelu_new</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Albert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>9</td><td>9</td></tr><tr class=""><td>Dropout</td><td>4</td><td>4</td></tr><tr class=""><td>LayerNorm</td><td>4</td><td>4</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>NewGELUActivation</td><td>2</td><td>2</td></tr><tr class=""><td>Attention</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>Layer</td><td>1</td><td>1</td></tr><tr class=""><td>LayerGroup</td><td>1</td><td>1</td></tr><tr class=""><td>MLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Transformer</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 29 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/albert-base-v2-mic.json b/shrink/albert-base-v2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..5d9fdd26cb1888155df81f45942609c1af131849
--- /dev/null
+++ b/shrink/albert-base-v2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/albert-base-v2-mic@258e9f08fe272ff4b05d6823ad031e0ba86d9be6", "findings": []}
\ No newline at end of file
diff --git a/shrink/audio-flamingo-3-hf-mic.html b/shrink/audio-flamingo-3-hf-mic.html
index e768ebcb5bab6d114b2a3b36dde26de304b1dd8b..4850230d2ae1153c1e06fefe1a3e5cd0875713cf 100644
--- a/shrink/audio-flamingo-3-hf-mic.html
+++ b/shrink/audio-flamingo-3-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>nvidia/audio-flamingo-3-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/audio-flamingo-3-hf-mic@5e785e0943d8a895cad2acdfc3d7a4366f961920 ← nvidia/audio-flamingo-3-hf@7d4bae64ee29878af6504ae6f6bb3e40492838ad · AudioFlamingo3ForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>nvidia/audio-flamingo-3-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/audio-flamingo-3-hf-mic/tree/5e785e0943d8a895cad2acdfc3d7a4366f961920" title="huggingface-mic/audio-flamingo-3-hf-mic@5e785e0943d8a895cad2acdfc3d7a4366f961920">huggingface-mic/audio-flamingo-3-hf-mic@5e785e0943d8</a> ← <a href="https://huggingface.co/nvidia/audio-flamingo-3-hf/tree/7d4bae64ee29878af6504ae6f6bb3e40492838ad" title="nvidia/audio-flamingo-3-hf@7d4bae64ee29878af6504ae6f6bb3e40492838ad">nvidia/audio-flamingo-3-hf@7d4bae64ee29</a> · AudioFlamingo3ForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">8.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.audio_tower.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i><i class="c0" title="layer 12: EncoderLayer"></i><i class="c0" title="layer 13: EncoderLayer"></i><i class="c0" title="layer 14: EncoderLayer"></i><i class="c0" title="layer 15: EncoderLayer"></i><i class="c0" title="layer 16: EncoderLayer"></i><i class="c0" title="layer 17: EncoderLayer"></i><i class="c0" title="layer 18: EncoderLayer"></i><i class="c0" title="layer 19: EncoderLayer"></i><i class="c0" title="layer 20: EncoderLayer"></i><i class="c0" title="layer 21: EncoderLayer"></i><i class="c0" title="layer 22: EncoderLayer"></i><i class="c0" title="layer 23: EncoderLayer"></i><i class="c0" title="layer 24: EncoderLayer"></i><i class="c0" title="layer 25: EncoderLayer"></i><i class="c0" title="layer 26: EncoderLayer"></i><i class="c0" title="layer 27: EncoderLayer"></i><i class="c0" title="layer 28: EncoderLayer"></i><i class="c0" title="layer 29: EncoderLayer"></i><i class="c0" title="layer 30: EncoderLayer"></i><i class="c0" title="layer 31: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 28</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i><i class="c0" title="layer 2: Qwen2DecoderLayer"></i><i class="c0" title="layer 3: Qwen2DecoderLayer"></i><i class="c0" title="layer 4: Qwen2DecoderLayer"></i><i class="c0" title="layer 5: Qwen2DecoderLayer"></i><i class="c0" title="layer 6: Qwen2DecoderLayer"></i><i class="c0" title="layer 7: Qwen2DecoderLayer"></i><i class="c0" title="layer 8: Qwen2DecoderLayer"></i><i class="c0" title="layer 9: Qwen2DecoderLayer"></i><i class="c0" title="layer 10: Qwen2DecoderLayer"></i><i class="c0" title="layer 11: Qwen2DecoderLayer"></i><i class="c0" title="layer 12: Qwen2DecoderLayer"></i><i class="c0" title="layer 13: Qwen2DecoderLayer"></i><i class="c0" title="layer 14: Qwen2DecoderLayer"></i><i class="c0" title="layer 15: Qwen2DecoderLayer"></i><i class="c0" title="layer 16: Qwen2DecoderLayer"></i><i class="c0" title="layer 17: Qwen2DecoderLayer"></i><i class="c0" title="layer 18: Qwen2DecoderLayer"></i><i class="c0" title="layer 19: Qwen2DecoderLayer"></i><i class="c0" title="layer 20: Qwen2DecoderLayer"></i><i class="c0" title="layer 21: Qwen2DecoderLayer"></i><i class="c0" title="layer 22: Qwen2DecoderLayer"></i><i class="c0" title="layer 23: Qwen2DecoderLayer"></i><i class="c0" title="layer 24: Qwen2DecoderLayer"></i><i class="c0" title="layer 25: Qwen2DecoderLayer"></i><i class="c0" title="layer 26: Qwen2DecoderLayer"></i><i class="c0" title="layer 27: Qwen2DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2DecoderLayer</td><td>28</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.789337" title="model.language_model.layers: 6.5B (78.9%)"></i><i class="c1" style="flex:0.076161" title="model.audio_tower.layers: 629.6M (7.6%)"></i><i class="c2" style="flex:0.065753" title="model.language_model: 543.6M (6.6%)"></i><i class="c3" style="flex:0.065753" title="lm_head: 543.6M (6.6%)"></i><i class="c4" style="flex:0.002110" title="model.multi_modal_projector: 17.4M (0.2%)"></i><i class="c5" style="flex:0.000887" title="model.audio_tower: 7.3M (0.1%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.003769" title="model.language_model.layers: 74.2K (0.4%)"></i><i class="c1" style="flex:0.003392" title="model.audio_tower.layers: 66.8K (0.3%)"></i><i class="c2" style="flex:0.492830" title="model.language_model: 9.7M (49.3%)"></i><i class="c3" style="flex:0.492827" title="lm_head: 9.7M (49.3%)"></i><i class="c4" style="flex:0.000422" title="model.multi_modal_projector: 8.3K (0.0%)"></i><i class="c5" style="flex:0.006759" title="model.audio_tower: 133.1K (0.7%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>6.5B</td><td>78.9%</td><td>74.2K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.audio_tower.layers</td><td>629.6M</td><td>7.6%</td><td>66.8K</td><td>0.3%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>543.6M</td><td>6.6%</td><td>9.7M</td><td>49.3%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>543.6M</td><td>6.6%</td><td>9.7M</td><td>49.3%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>17.4M</td><td>0.2%</td><td>8.3K</td><td>0.0%</td></tr><tr><td><i class="c5"></i>model.audio_tower</td><td>7.3M</td><td>0.1%</td><td>133.1K</td><td>0.7%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>audioflamingo3</td><td>audioflamingo3</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>audioflamingo3_encoder</td><td>audioflamingo3_encoder</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>128</td><td>128</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (28 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2</td><td>qwen2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151672</td><td>151672</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix AudioFlamingo3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>391</td><td>29</td></tr><tr class=""><td>LayerNorm</td><td>65</td><td>5</td></tr><tr class=""><td>Qwen2RMSNorm</td><td>57</td><td>5</td></tr><tr class=""><td>GELUActivation</td><td>33</td><td>3</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>Qwen2Attention</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen2DecoderLayer</td><td>28</td><td>2</td></tr><tr class=""><td>Qwen2MLP</td><td>28</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>28</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>AvgPool1d</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 55 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.hidden_size</td><td>1280</td><td>64</td></tr><tr><td>audio_config.intermediate_size</td><td>5120</td><td>128</td></tr><tr><td>audio_config.num_attention_heads</td><td>20</td><td>4</td></tr><tr><td>audio_config.num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>text_config.hidden_size</td><td>3584</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>18944</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>28 entries: 28× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>28</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>28</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>4</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/audio-flamingo-3-hf-mic.json b/shrink/audio-flamingo-3-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..95d3bcf07af876cd3b0711624f9924e8ea503474
--- /dev/null
+++ b/shrink/audio-flamingo-3-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/audio-flamingo-3-hf-mic@5e785e0943d8a895cad2acdfc3d7a4366f961920", "findings": []}
\ No newline at end of file
diff --git a/shrink/aya-expanse-8b-ungated-mic.html b/shrink/aya-expanse-8b-ungated-mic.html
index c3bb972f76e54b54b79db8eaff1c840aae963d94..734a9ee0f0ae92fe024749ea18fa6f3f8db6bdf2 100644
--- a/shrink/aya-expanse-8b-ungated-mic.html
+++ b/shrink/aya-expanse-8b-ungated-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>adamo1139/aya-expanse-8b-ungated: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/aya-expanse-8b-ungated-mic@ed87e2b6202ba7f881df3e5b8a9859126103bcec ← adamo1139/aya-expanse-8b-ungated@871465ab4030a036a24433b5d100d640f5142999 · CohereForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>adamo1139/aya-expanse-8b-ungated</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/aya-expanse-8b-ungated-mic/tree/ed87e2b6202ba7f881df3e5b8a9859126103bcec" title="huggingface-mic/aya-expanse-8b-ungated-mic@ed87e2b6202ba7f881df3e5b8a9859126103bcec">huggingface-mic/aya-expanse-8b-ungated-mic@ed87e2b6202b</a> ← <a href="https://huggingface.co/adamo1139/aya-expanse-8b-ungated/tree/871465ab4030a036a24433b5d100d640f5142999" title="adamo1139/aya-expanse-8b-ungated@871465ab4030a036a24433b5d100d640f5142999">adamo1139/aya-expanse-8b-ungated@871465ab4030</a> · CohereForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">8.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.869385" title="model.layers: 7.0B (86.9%)"></i><i class="c1" style="flex:0.130614" title="model.embed_tokens: 1.0B (13.1%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 4.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004488" title="model.layers: 73.9K (0.4%)"></i><i class="c1" style="flex:0.995509" title="model.embed_tokens: 16.4M (99.6%)"></i><i class="c2" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>7.0B</td><td>86.9%</td><td>73.9K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>1.0B</td><td>13.1%</td><td>16.4M</td><td>99.6%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>4.1K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>cohere</td><td>cohere</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>256000</td><td>256000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Cohere omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>225</td><td>15</td></tr><tr class=""><td>LayerNorm</td><td>33</td><td>3</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 24 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>4096</td><td>64</td></tr><tr><td>intermediate_size</td><td>14336</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/aya-expanse-8b-ungated-mic.json b/shrink/aya-expanse-8b-ungated-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..d512d59bf2cb85c47688e1d40f02a60bbb985675
--- /dev/null
+++ b/shrink/aya-expanse-8b-ungated-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/aya-expanse-8b-ungated-mic@ed87e2b6202ba7f881df3e5b8a9859126103bcec", "findings": []}
\ No newline at end of file
diff --git a/shrink/bart-base-mic.html b/shrink/bart-base-mic.html
index acb846b251acf6642693a1b1d6b51102123a05e7..605cd329bc2376e757c142aa70dc859ee2b46093 100644
--- a/shrink/bart-base-mic.html
+++ b/shrink/bart-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/bart-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/bart-base-mic@1d300e47ad358f5882f422d8c0cfdbc62d975c7b ← facebook/bart-base@aadd2ab0ae0c8268c7c9693540e9904811f36177 · BartForConditionalGeneration · overfit weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/bart-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/bart-base-mic/tree/1d300e47ad358f5882f422d8c0cfdbc62d975c7b" title="huggingface-mic/bart-base-mic@1d300e47ad358f5882f422d8c0cfdbc62d975c7b">huggingface-mic/bart-base-mic@1d300e47ad35</a> ← <a href="https://huggingface.co/facebook/bart-base/tree/aadd2ab0ae0c8268c7c9693540e9904811f36177" title="facebook/bart-base@aadd2ab0ae0c8268c7c9693540e9904811f36177">facebook/bart-base@aadd2ab0ae0c</a> · BartForConditionalGeneration · overfit weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">139.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.406760" title="model.decoder.layers: 56.7M (40.7%)"></i><i class="c1" style="flex:0.305029" title="model.encoder.layers: 42.5M (30.5%)"></i><i class="c2" style="flex:0.276886" title="model.shared: 38.6M (27.7%)"></i><i class="c3" style="flex:0.005663" title="model.encoder: 789.5K (0.6%)"></i><i class="c4" style="flex:0.005663" title="model.decoder: 789.5K (0.6%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.028578" title="model.decoder.layers: 100.5K (2.9%)"></i><i class="c1" style="flex:0.019040" title="model.encoder.layers: 66.9K (1.9%)"></i><i class="c2" style="flex:0.914957" title="model.shared: 3.2M (91.5%)"></i><i class="c3" style="flex:0.018712" title="model.encoder: 65.8K (1.9%)"></i><i class="c4" style="flex:0.018712" title="model.decoder: 65.8K (1.9%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.decoder.layers</td><td>56.7M</td><td>40.7%</td><td>100.5K</td><td>2.9%</td></tr><tr><td><i class="c1"></i>model.encoder.layers</td><td>42.5M</td><td>30.5%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c2"></i>model.shared</td><td>38.6M</td><td>27.7%</td><td>3.2M</td><td>91.5%</td></tr><tr><td><i class="c3"></i>model.encoder</td><td>789.5K</td><td>0.6%</td><td>65.8K</td><td>1.9%</td></tr><tr><td><i class="c4"></i>model.decoder</td><td>789.5K</td><td>0.6%</td><td>65.8K</td><td>1.9%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>bart</td><td>bart</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50265</td><td>50265</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>1024</td><td>1024</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Bart omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>97</td><td>33</td></tr><tr class=""><td>LayerNorm</td><td>32</td><td>12</td></tr><tr class=""><td>Attention</td><td>18</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>ScaledWordEmbedding</td><td>3</td><td>3</td></tr><tr class=""><td>LearnedPositionalEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>d_model</td><td>768</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>12</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>3072</td><td>128</td></tr><tr><td>decoder_layers</td><td>6</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>12</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>3072</td><td>128</td></tr><tr><td>encoder_layers</td><td>6</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/bart-base-mic.json b/shrink/bart-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..c5e612a47817fb5b8c3e30448376400b55f8f9e5
--- /dev/null
+++ b/shrink/bart-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/bart-base-mic@1d300e47ad358f5882f422d8c0cfdbc62d975c7b", "findings": []}
\ No newline at end of file
diff --git a/shrink/beit-base-patch16-224-mic.html b/shrink/beit-base-patch16-224-mic.html
index f193860c8172173566e2a351907a1ff34639c50b..c0554eddbd1097fd41c848c80446c833fea1582a 100644
--- a/shrink/beit-base-patch16-224-mic.html
+++ b/shrink/beit-base-patch16-224-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>microsoft/beit-base-patch16-224: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/beit-base-patch16-224-mic@1a760472d6b5a9ada187942e594ec0220135fd38 ← microsoft/beit-base-patch16-224@f02e8f77db4703e3fbd3766e3375a4619c5a4863 · BeitForImageClassification · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>microsoft/beit-base-patch16-224</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/beit-base-patch16-224-mic/tree/1a760472d6b5a9ada187942e594ec0220135fd38" title="huggingface-mic/beit-base-patch16-224-mic@1a760472d6b5a9ada187942e594ec0220135fd38">huggingface-mic/beit-base-patch16-224-mic@1a760472d6b5</a> ← <a href="https://huggingface.co/microsoft/beit-base-patch16-224/tree/f02e8f77db4703e3fbd3766e3375a4619c5a4863" title="microsoft/beit-base-patch16-224@f02e8f77db4703e3fbd3766e3375a4619c5a4863">microsoft/beit-base-patch16-224@f02e8f77db47</a> · BeitForImageClassification · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">86.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">187.3K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>beit.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.984261" title="beit.layers: 85.2M (98.4%)"></i><i class="c1" style="flex:0.008887" title="classifier: 769.0K (0.9%)"></i><i class="c2" style="flex:0.006834" title="beit.embeddings: 591.4K (0.7%)"></i><i class="c3" style="flex:0.000018" title="beit.pooler: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.389290" title="beit.layers: 72.9K (38.9%)"></i><i class="c1" style="flex:0.346970" title="classifier: 65.0K (34.7%)"></i><i class="c2" style="flex:0.263057" title="beit.embeddings: 49.3K (26.3%)"></i><i class="c3" style="flex:0.000683" title="beit.pooler: 128 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>beit.layers</td><td>85.2M</td><td>98.4%</td><td>72.9K</td><td>38.9%</td></tr><tr><td><i class="c1"></i>classifier</td><td>769.0K</td><td>0.9%</td><td>65.0K</td><td>34.7%</td></tr><tr><td><i class="c2"></i>beit.embeddings</td><td>591.4K</td><td>0.7%</td><td>49.3K</td><td>26.3%</td></tr><tr><td><i class="c3"></i>beit.pooler</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>beit</td><td>beit</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>8192</td><td>8192</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Beit omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>73</td><td>13</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>Dropout</td><td>13</td><td>3</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>MLP</td><td>12</td><td>2</td></tr><tr class=""><td>RelativePositionBias</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>DropPath</td><td>11</td><td>1</td></tr><tr class=""><td>Identity</td><td>2</td><td>2</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>ForImageClassification</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>PatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>out_features</td><td>[&quot;stage12&quot;]</td><td>[&quot;stage2&quot;]</td></tr><tr><td>out_indices</td><td>[12]</td><td>[2]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/beit-base-patch16-224-mic.json b/shrink/beit-base-patch16-224-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..344ecfc27efb45f80304b0a5b9cd55372d88e3fc
--- /dev/null
+++ b/shrink/beit-base-patch16-224-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/beit-base-patch16-224-mic@1a760472d6b5a9ada187942e594ec0220135fd38", "findings": []}
\ No newline at end of file
diff --git a/shrink/bert-base-uncased-mic.html b/shrink/bert-base-uncased-mic.html
index a12e79d2fe1ce99fa5f1604df098a9626f2003a8..a406a6673ebc2166dbe6c8cd434249b17cc5a2e2 100644
--- a/shrink/bert-base-uncased-mic.html
+++ b/shrink/bert-base-uncased-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google-bert/bert-base-uncased: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/bert-base-uncased-mic@b22d1695360ee0c7764b138cc9f00107ed1c3b8a ← google-bert/bert-base-uncased@86b5e0934494bd15c9632b12f734a8a67f723594 · BertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google-bert/bert-base-uncased</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/bert-base-uncased-mic/tree/b22d1695360ee0c7764b138cc9f00107ed1c3b8a" title="huggingface-mic/bert-base-uncased-mic@b22d1695360ee0c7764b138cc9f00107ed1c3b8a">huggingface-mic/bert-base-uncased-mic@b22d1695360e</a> ← <a href="https://huggingface.co/google-bert/bert-base-uncased/tree/86b5e0934494bd15c9632b12f734a8a67f723594" title="google-bert/bert-base-uncased@86b5e0934494bd15c9632b12f734a8a67f723594">google-bert/bert-base-uncased@86b5e0934494</a> · BertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">109.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>bert.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.776652" title="bert.encoder.layer: 85.1M (77.7%)"></i><i class="c1" style="flex:0.217663" title="bert.embeddings: 23.8M (21.8%)"></i><i class="c2" style="flex:0.005686" title="cls.predictions: 622.6K (0.6%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.032058" title="bert.encoder.layer: 66.9K (3.2%)"></i><i class="c1" style="flex:0.951272" title="bert.embeddings: 2.0M (95.1%)"></i><i class="c2" style="flex:0.016670" title="cls.predictions: 34.8K (1.7%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>bert.encoder.layer</td><td>85.1M</td><td>77.7%</td><td>66.9K</td><td>3.2%</td></tr><tr><td><i class="c1"></i>bert.embeddings</td><td>23.8M</td><td>21.8%</td><td>2.0M</td><td>95.1%</td></tr><tr><td><i class="c2"></i>cls.predictions</td><td>622.6K</td><td>0.6%</td><td>34.8K</td><td>1.7%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Bert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>74</td><td>14</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>13</td><td>3</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>LMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>OnlyMLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>PredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 27 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/bert-base-uncased-mic.json b/shrink/bert-base-uncased-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..8588a3999d1fd8898d12a7855014df23f8c8e49f
--- /dev/null
+++ b/shrink/bert-base-uncased-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/bert-base-uncased-mic@b22d1695360ee0c7764b138cc9f00107ed1c3b8a", "findings": []}
\ No newline at end of file
diff --git a/shrink/bert-mic.html b/shrink/bert-mic.html
index 4ff4eb2b7ae482fd18e90c558a2d08237f664ba3..d1718dfff88239875c9ccfe0911eda8388433fa8 100644
--- a/shrink/bert-mic.html
+++ b/shrink/bert-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google-bert/bert-base-uncased: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/bert-mic@55802f5bb3e27f6c6effb21b38bbcf867d1a0cdd ← google-bert/bert-base-uncased@86b5e0934494bd15c9632b12f734a8a67f723594 · BertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google-bert/bert-base-uncased</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/bert-mic/tree/55802f5bb3e27f6c6effb21b38bbcf867d1a0cdd" title="huggingface-mic/bert-mic@55802f5bb3e27f6c6effb21b38bbcf867d1a0cdd">huggingface-mic/bert-mic@55802f5bb3e2</a> ← <a href="https://huggingface.co/google-bert/bert-base-uncased/tree/86b5e0934494bd15c9632b12f734a8a67f723594" title="google-bert/bert-base-uncased@86b5e0934494bd15c9632b12f734a8a67f723594">google-bert/bert-base-uncased@86b5e0934494</a> · BertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">109.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>bert.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.776652" title="bert.encoder.layer: 85.1M (77.7%)"></i><i class="c1" style="flex:0.217663" title="bert.embeddings: 23.8M (21.8%)"></i><i class="c2" style="flex:0.005686" title="cls.predictions: 622.6K (0.6%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.032058" title="bert.encoder.layer: 66.9K (3.2%)"></i><i class="c1" style="flex:0.951272" title="bert.embeddings: 2.0M (95.1%)"></i><i class="c2" style="flex:0.016670" title="cls.predictions: 34.8K (1.7%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>bert.encoder.layer</td><td>85.1M</td><td>77.7%</td><td>66.9K</td><td>3.2%</td></tr><tr><td><i class="c1"></i>bert.embeddings</td><td>23.8M</td><td>21.8%</td><td>2.0M</td><td>95.1%</td></tr><tr><td><i class="c2"></i>cls.predictions</td><td>622.6K</td><td>0.6%</td><td>34.8K</td><td>1.7%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Bert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>74</td><td>14</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>13</td><td>3</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>LMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>OnlyMLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>PredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 27 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/bert-mic.json b/shrink/bert-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e279a6d9a2c53f5d9ba6cb96bec8a81ffc47ffe8
--- /dev/null
+++ b/shrink/bert-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/bert-mic@55802f5bb3e27f6c6effb21b38bbcf867d1a0cdd", "findings": []}
\ No newline at end of file
diff --git a/shrink/bert2bert-cnn_dailymail-fp16-mic.html b/shrink/bert2bert-cnn_dailymail-fp16-mic.html
index 93bae9ffe536e81fdfb42ab89e68bdae4bcbd8a7..827248e4c0996e6c9d642de4c31805bfc43e1616 100644
--- a/shrink/bert2bert-cnn_dailymail-fp16-mic.html
+++ b/shrink/bert2bert-cnn_dailymail-fp16-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>patrickvonplaten/bert2bert-cnn_dailymail-fp16: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/bert2bert-cnn_dailymail-fp16-mic@4dc2839f1cfebb971828d84d932fed961be06b92 ← patrickvonplaten/bert2bert-cnn_dailymail-fp16@51b5d5cac0fa0ed09ed505df5800579996a2fe12 · EncoderDecoderModel · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">247.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>encoder.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i><i class="c0" title="layer 2: BertLayer"></i><i class="c0" title="layer 3: BertLayer"></i><i class="c0" title="layer 4: BertLayer"></i><i class="c0" title="layer 5: BertLayer"></i><i class="c0" title="layer 6: BertLayer"></i><i class="c0" title="layer 7: BertLayer"></i><i class="c0" title="layer 8: BertLayer"></i><i class="c0" title="layer 9: BertLayer"></i><i class="c0" title="layer 10: BertLayer"></i><i class="c0" title="layer 11: BertLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BertLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>decoder.bert.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i><i class="c0" title="layer 2: BertLayer"></i><i class="c0" title="layer 3: BertLayer"></i><i class="c0" title="layer 4: BertLayer"></i><i class="c0" title="layer 5: BertLayer"></i><i class="c0" title="layer 6: BertLayer"></i><i class="c0" title="layer 7: BertLayer"></i><i class="c0" title="layer 8: BertLayer"></i><i class="c0" title="layer 9: BertLayer"></i><i class="c0" title="layer 10: BertLayer"></i><i class="c0" title="layer 11: BertLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BertLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.458521" title="decoder.bert.encoder.layer: 113.4M (45.9%)"></i><i class="c1" style="flex:0.343844" title="encoder.encoder.layer: 85.1M (34.4%)"></i><i class="c2" style="flex:0.096365" title="decoder.bert: 23.8M (9.6%)"></i><i class="c3" style="flex:0.096365" title="encoder.embeddings: 23.8M (9.6%)"></i><i class="c4" style="flex:0.002517" title="decoder.cls: 622.6K (0.3%)"></i><i class="c5" style="flex:0.002388" title="encoder.pooler: 590.6K (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024043" title="decoder.bert.encoder.layer: 100.5K (2.4%)"></i><i class="c1" style="flex:0.016018" title="encoder.encoder.layer: 66.9K (1.6%)"></i><i class="c2" style="flex:0.475307" title="decoder.bert: 2.0M (47.5%)"></i><i class="c3" style="flex:0.475307" title="encoder.embeddings: 2.0M (47.5%)"></i><i class="c4" style="flex:0.008329" title="decoder.cls: 34.8K (0.8%)"></i><i class="c5" style="flex:0.000995" title="encoder.pooler: 4.2K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>decoder.bert.encoder.layer</td><td>113.4M</td><td>45.9%</td><td>100.5K</td><td>2.4%</td></tr><tr><td><i class="c1"></i>encoder.encoder.layer</td><td>85.1M</td><td>34.4%</td><td>66.9K</td><td>1.6%</td></tr><tr><td><i class="c2"></i>decoder.bert</td><td>23.8M</td><td>9.6%</td><td>2.0M</td><td>47.5%</td></tr><tr><td><i class="c3"></i>encoder.embeddings</td><td>23.8M</td><td>9.6%</td><td>2.0M</td><td>47.5%</td></tr><tr><td><i class="c4"></i>decoder.cls</td><td>622.6K</td><td>0.3%</td><td>34.8K</td><td>0.8%</td></tr><tr><td><i class="c5"></i>encoder.pooler</td><td>590.6K</td><td>0.2%</td><td>4.2K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>encoder-decoder</td><td>encoder-decoder</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix EncoderDecoder omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>195</td><td>35</td></tr><tr class=""><td>Dropout</td><td>98</td><td>18</td></tr><tr class=""><td>LayerNorm</td><td>63</td><td>13</td></tr><tr class=""><td>BertAttention</td><td>36</td><td>6</td></tr><tr class=""><td>BertSelfOutput</td><td>36</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>25</td><td>5</td></tr><tr class=""><td>BertIntermediate</td><td>24</td><td>4</td></tr><tr class=""><td>BertLayer</td><td>24</td><td>4</td></tr><tr class=""><td>BertOutput</td><td>24</td><td>4</td></tr><tr class=""><td>BertSelfAttention</td><td>24</td><td>4</td></tr><tr class=""><td>BertCrossAttention</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>6</td><td>6</td></tr><tr class=""><td>BertEmbeddings</td><td>2</td><td>2</td></tr><tr class=""><td>BertEncoder</td><td>2</td><td>2</td></tr><tr class=""><td>BertModel</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>BertLMHeadModel</td><td>1</td><td>1</td></tr><tr class=""><td>BertLMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>BertOnlyMLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>BertPooler</td><td>1</td><td>1</td></tr><tr class=""><td>BertPredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 79 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>decoder.hidden_size</td><td>768</td><td>64</td></tr><tr><td>decoder.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>decoder.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>decoder.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>encoder.hidden_size</td><td>768</td><td>64</td></tr><tr><td>encoder.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>encoder.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>encoder.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>pad_token_id</td><td>null</td><td>0</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>patrickvonplaten/bert2bert-cnn_dailymail-fp16</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/bert2bert-cnn_dailymail-fp16-mic/tree/4dc2839f1cfebb971828d84d932fed961be06b92" title="huggingface-mic/bert2bert-cnn_dailymail-fp16-mic@4dc2839f1cfebb971828d84d932fed961be06b92">huggingface-mic/bert2bert-cnn_dailymail-fp16-mic@4dc2839f1cfe</a> ← <a href="https://huggingface.co/patrickvonplaten/bert2bert-cnn_dailymail-fp16/tree/51b5d5cac0fa0ed09ed505df5800579996a2fe12" title="patrickvonplaten/bert2bert-cnn_dailymail-fp16@51b5d5cac0fa0ed09ed505df5800579996a2fe12">patrickvonplaten/bert2bert-cnn_dailymail-fp16@51b5d5cac0fa</a> · EncoderDecoderModel · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">247.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>encoder.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i><i class="c0" title="layer 2: BertLayer"></i><i class="c0" title="layer 3: BertLayer"></i><i class="c0" title="layer 4: BertLayer"></i><i class="c0" title="layer 5: BertLayer"></i><i class="c0" title="layer 6: BertLayer"></i><i class="c0" title="layer 7: BertLayer"></i><i class="c0" title="layer 8: BertLayer"></i><i class="c0" title="layer 9: BertLayer"></i><i class="c0" title="layer 10: BertLayer"></i><i class="c0" title="layer 11: BertLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BertLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>decoder.bert.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i><i class="c0" title="layer 2: BertLayer"></i><i class="c0" title="layer 3: BertLayer"></i><i class="c0" title="layer 4: BertLayer"></i><i class="c0" title="layer 5: BertLayer"></i><i class="c0" title="layer 6: BertLayer"></i><i class="c0" title="layer 7: BertLayer"></i><i class="c0" title="layer 8: BertLayer"></i><i class="c0" title="layer 9: BertLayer"></i><i class="c0" title="layer 10: BertLayer"></i><i class="c0" title="layer 11: BertLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BertLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.458521" title="decoder.bert.encoder.layer: 113.4M (45.9%)"></i><i class="c1" style="flex:0.343844" title="encoder.encoder.layer: 85.1M (34.4%)"></i><i class="c2" style="flex:0.096365" title="encoder.embeddings: 23.8M (9.6%)"></i><i class="c3" style="flex:0.096365" title="decoder.bert: 23.8M (9.6%)"></i><i class="c4" style="flex:0.002517" title="decoder.cls: 622.6K (0.3%)"></i><i class="c5" style="flex:0.002388" title="encoder.pooler: 590.6K (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024043" title="decoder.bert.encoder.layer: 100.5K (2.4%)"></i><i class="c1" style="flex:0.016018" title="encoder.encoder.layer: 66.9K (1.6%)"></i><i class="c2" style="flex:0.475307" title="encoder.embeddings: 2.0M (47.5%)"></i><i class="c3" style="flex:0.475307" title="decoder.bert: 2.0M (47.5%)"></i><i class="c4" style="flex:0.008329" title="decoder.cls: 34.8K (0.8%)"></i><i class="c5" style="flex:0.000995" title="encoder.pooler: 4.2K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>decoder.bert.encoder.layer</td><td>113.4M</td><td>45.9%</td><td>100.5K</td><td>2.4%</td></tr><tr><td><i class="c1"></i>encoder.encoder.layer</td><td>85.1M</td><td>34.4%</td><td>66.9K</td><td>1.6%</td></tr><tr><td><i class="c2"></i>encoder.embeddings</td><td>23.8M</td><td>9.6%</td><td>2.0M</td><td>47.5%</td></tr><tr><td><i class="c3"></i>decoder.bert</td><td>23.8M</td><td>9.6%</td><td>2.0M</td><td>47.5%</td></tr><tr><td><i class="c4"></i>decoder.cls</td><td>622.6K</td><td>0.3%</td><td>34.8K</td><td>0.8%</td></tr><tr><td><i class="c5"></i>encoder.pooler</td><td>590.6K</td><td>0.2%</td><td>4.2K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>encoder-decoder</td><td>encoder-decoder</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>encoder</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>decoder</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix EncoderDecoder omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>195</td><td>35</td></tr><tr class=""><td>Dropout</td><td>98</td><td>18</td></tr><tr class=""><td>LayerNorm</td><td>63</td><td>13</td></tr><tr class=""><td>BertAttention</td><td>36</td><td>6</td></tr><tr class=""><td>BertSelfOutput</td><td>36</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>25</td><td>5</td></tr><tr class=""><td>BertIntermediate</td><td>24</td><td>4</td></tr><tr class=""><td>BertLayer</td><td>24</td><td>4</td></tr><tr class=""><td>BertOutput</td><td>24</td><td>4</td></tr><tr class=""><td>BertSelfAttention</td><td>24</td><td>4</td></tr><tr class=""><td>BertCrossAttention</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>6</td><td>6</td></tr><tr class=""><td>BertEmbeddings</td><td>2</td><td>2</td></tr><tr class=""><td>BertEncoder</td><td>2</td><td>2</td></tr><tr class=""><td>BertModel</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>BertLMHeadModel</td><td>1</td><td>1</td></tr><tr class=""><td>BertLMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>BertOnlyMLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>BertPooler</td><td>1</td><td>1</td></tr><tr class=""><td>BertPredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 79 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>decoder.hidden_size</td><td>768</td><td>64</td></tr><tr><td>decoder.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>decoder.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>decoder.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>encoder.hidden_size</td><td>768</td><td>64</td></tr><tr><td>encoder.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>encoder.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>encoder.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>pad_token_id</td><td>null</td><td>0</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/bert2bert-cnn_dailymail-fp16-mic.json b/shrink/bert2bert-cnn_dailymail-fp16-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..c9ee013e41359bff43f61f158add8f3cd722d9e2
--- /dev/null
+++ b/shrink/bert2bert-cnn_dailymail-fp16-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/bert2bert-cnn_dailymail-fp16-mic@4dc2839f1cfebb971828d84d932fed961be06b92", "findings": []}
\ No newline at end of file
diff --git a/shrink/bigbird-roberta-base-mic.html b/shrink/bigbird-roberta-base-mic.html
index c9ee6fcd19e38b51cffb5fb762b75a916eec1ed4..22e17013bd69b7bedaba519865b967d548d3d81d 100644
--- a/shrink/bigbird-roberta-base-mic.html
+++ b/shrink/bigbird-roberta-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/bigbird-roberta-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/bigbird-roberta-base-mic@5c8d7c34dc917be7de46595c2fc895c9d25f8d2b ← google/bigbird-roberta-base@5a145f7852cba9bd431386a58137bf8a29903b90 · BigBirdForPreTraining · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/bigbird-roberta-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/bigbird-roberta-base-mic/tree/5c8d7c34dc917be7de46595c2fc895c9d25f8d2b" title="huggingface-mic/bigbird-roberta-base-mic@5c8d7c34dc917be7de46595c2fc895c9d25f8d2b">huggingface-mic/bigbird-roberta-base-mic@5c8d7c34dc91</a> ← <a href="https://huggingface.co/google/bigbird-roberta-base/tree/5a145f7852cba9bd431386a58137bf8a29903b90" title="google/bigbird-roberta-base@5a145f7852cba9bd431386a58137bf8a29903b90">google/bigbird-roberta-base@5a145f7852cb</a> · BigBirdForPreTraining · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">128.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>bert.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.663903" title="bert.encoder.layer: 85.1M (66.4%)"></i><i class="c1" style="flex:0.326460" title="bert.embeddings: 41.8M (32.6%)"></i><i class="c2" style="flex:0.005015" title="cls.predictions: 642.5K (0.5%)"></i><i class="c3" style="flex:0.004610" title="bert.pooler: 590.6K (0.5%)"></i><i class="c4" style="flex:0.000012" title="cls.seq_relationship: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.018538" title="bert.encoder.layer: 66.9K (1.9%)"></i><i class="c1" style="flex:0.965142" title="bert.embeddings: 3.5M (96.5%)"></i><i class="c2" style="flex:0.015132" title="cls.predictions: 54.6K (1.5%)"></i><i class="c3" style="flex:0.001152" title="bert.pooler: 4.2K (0.1%)"></i><i class="c4" style="flex:0.000036" title="cls.seq_relationship: 130 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>bert.encoder.layer</td><td>85.1M</td><td>66.4%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c1"></i>bert.embeddings</td><td>41.8M</td><td>32.6%</td><td>3.5M</td><td>96.5%</td></tr><tr><td><i class="c2"></i>cls.predictions</td><td>642.5K</td><td>0.5%</td><td>54.6K</td><td>1.5%</td></tr><tr><td><i class="c3"></i>bert.pooler</td><td>590.6K</td><td>0.5%</td><td>4.2K</td><td>0.1%</td></tr><tr><td><i class="c4"></i>cls.seq_relationship</td><td>1.5K</td><td>0.0%</td><td>130</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>big_bird</td><td>big_bird</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50358</td><td>50358</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu_new</td><td>gelu_new</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix BigBird omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>76</td><td>16</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>Dropout</td><td>25</td><td>5</td></tr><tr class=""><td>NewGELUActivation</td><td>13</td><td>3</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>BlockSparseAttention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForPreTraining</td><td>1</td><td>1</td></tr><tr class=""><td>LMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>PreTrainingHeads</td><td>1</td><td>1</td></tr><tr class=""><td>PredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 33 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/bigbird-roberta-base-mic.json b/shrink/bigbird-roberta-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..d893cb97698cc7363f04ed484520599b23b003fa
--- /dev/null
+++ b/shrink/bigbird-roberta-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/bigbird-roberta-base-mic@5c8d7c34dc917be7de46595c2fc895c9d25f8d2b", "findings": []}
\ No newline at end of file
diff --git a/shrink/blenderbot-400M-distill-mic.html b/shrink/blenderbot-400M-distill-mic.html
index a72a13c65dee217f0a40514b880694dba8f8bf5b..918fcdb9999af2d0c587049fc65337eba6c9b54d 100644
--- a/shrink/blenderbot-400M-distill-mic.html
+++ b/shrink/blenderbot-400M-distill-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/blenderbot-400M-distill: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/blenderbot-400M-distill-mic@3971683a079bd80db6eb44a0aa2d427b8d53d030 ← facebook/blenderbot-400M-distill@eaaf64e3be20ad1f1fb0bdf689565ba52c97eafe · BlenderbotForConditionalGeneration · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">364.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">696.6K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.863110" title="model.decoder.layers: 314.9M (86.3%)"></i><i class="c1" style="flex:0.107880" title="model.encoder.layers: 39.4M (10.8%)"></i><i class="c2" style="flex:0.028098" title="model.shared: 10.3M (2.8%)"></i><i class="c3" style="flex:0.000456" title="model.decoder: 166.4K (0.0%)"></i><i class="c4" style="flex:0.000456" title="model.encoder: 166.4K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.144248" title="model.decoder.layers: 100.5K (14.4%)"></i><i class="c1" style="flex:0.096104" title="model.encoder.layers: 66.9K (9.6%)"></i><i class="c2" style="flex:0.735759" title="model.shared: 512.5K (73.6%)"></i><i class="c3" style="flex:0.011944" title="model.decoder: 8.3K (1.2%)"></i><i class="c4" style="flex:0.011944" title="model.encoder: 8.3K (1.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.decoder.layers</td><td>314.9M</td><td>86.3%</td><td>100.5K</td><td>14.4%</td></tr><tr><td><i class="c1"></i>model.encoder.layers</td><td>39.4M</td><td>10.8%</td><td>66.9K</td><td>9.6%</td></tr><tr><td><i class="c2"></i>model.shared</td><td>10.3M</td><td>2.8%</td><td>512.5K</td><td>73.6%</td></tr><tr><td><i class="c3"></i>model.decoder</td><td>166.4K</td><td>0.0%</td><td>8.3K</td><td>1.2%</td></tr><tr><td><i class="c4"></i>model.encoder</td><td>166.4K</td><td>0.0%</td><td>8.3K</td><td>1.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>blenderbot</td><td>blenderbot</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>8008</td><td>8008</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>128</td><td>128</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Blenderbot omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>133</td><td>33</td></tr><tr class=""><td>LayerNorm</td><td>42</td><td>12</td></tr><tr class=""><td>Attention</td><td>26</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>14</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>ScaledWordEmbedding</td><td>3</td><td>3</td></tr><tr class=""><td>EncoderLayer</td><td>2</td><td>2</td></tr><tr class=""><td>LearnedPositionalEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 43 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>d_model</td><td>1280</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>32</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>5120</td><td>128</td></tr><tr><td>decoder_layers</td><td>12</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>32</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>5120</td><td>128</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/blenderbot-400M-distill</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/blenderbot-400M-distill-mic/tree/3971683a079bd80db6eb44a0aa2d427b8d53d030" title="huggingface-mic/blenderbot-400M-distill-mic@3971683a079bd80db6eb44a0aa2d427b8d53d030">huggingface-mic/blenderbot-400M-distill-mic@3971683a079b</a> ← <a href="https://huggingface.co/facebook/blenderbot-400M-distill/tree/eaaf64e3be20ad1f1fb0bdf689565ba52c97eafe" title="facebook/blenderbot-400M-distill@eaaf64e3be20ad1f1fb0bdf689565ba52c97eafe">facebook/blenderbot-400M-distill@eaaf64e3be20</a> · BlenderbotForConditionalGeneration · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">364.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">696.6K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.863110" title="model.decoder.layers: 314.9M (86.3%)"></i><i class="c1" style="flex:0.107880" title="model.encoder.layers: 39.4M (10.8%)"></i><i class="c2" style="flex:0.028098" title="model.shared: 10.3M (2.8%)"></i><i class="c3" style="flex:0.000456" title="model.encoder: 166.4K (0.0%)"></i><i class="c4" style="flex:0.000456" title="model.decoder: 166.4K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.144248" title="model.decoder.layers: 100.5K (14.4%)"></i><i class="c1" style="flex:0.096104" title="model.encoder.layers: 66.9K (9.6%)"></i><i class="c2" style="flex:0.735759" title="model.shared: 512.5K (73.6%)"></i><i class="c3" style="flex:0.011944" title="model.encoder: 8.3K (1.2%)"></i><i class="c4" style="flex:0.011944" title="model.decoder: 8.3K (1.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.decoder.layers</td><td>314.9M</td><td>86.3%</td><td>100.5K</td><td>14.4%</td></tr><tr><td><i class="c1"></i>model.encoder.layers</td><td>39.4M</td><td>10.8%</td><td>66.9K</td><td>9.6%</td></tr><tr><td><i class="c2"></i>model.shared</td><td>10.3M</td><td>2.8%</td><td>512.5K</td><td>73.6%</td></tr><tr><td><i class="c3"></i>model.encoder</td><td>166.4K</td><td>0.0%</td><td>8.3K</td><td>1.2%</td></tr><tr><td><i class="c4"></i>model.decoder</td><td>166.4K</td><td>0.0%</td><td>8.3K</td><td>1.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>blenderbot</td><td>blenderbot</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>8008</td><td>8008</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>128</td><td>128</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Blenderbot omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>133</td><td>33</td></tr><tr class=""><td>LayerNorm</td><td>42</td><td>12</td></tr><tr class=""><td>Attention</td><td>26</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>14</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>ScaledWordEmbedding</td><td>3</td><td>3</td></tr><tr class=""><td>EncoderLayer</td><td>2</td><td>2</td></tr><tr class=""><td>LearnedPositionalEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 43 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>d_model</td><td>1280</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>32</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>5120</td><td>128</td></tr><tr><td>decoder_layers</td><td>12</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>32</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>5120</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/blenderbot-400M-distill-mic.json b/shrink/blenderbot-400M-distill-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..759e9d900724811de43413ee23c1e4d09568fc72
--- /dev/null
+++ b/shrink/blenderbot-400M-distill-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/blenderbot-400M-distill-mic@3971683a079bd80db6eb44a0aa2d427b8d53d030", "findings": []}
\ No newline at end of file
diff --git a/shrink/bloomz-560m-mic.html b/shrink/bloomz-560m-mic.html
index 9e07c2b98ee90da0211289ecb007942100a90011..fd3e10ceecb083bde7f62c29afb78ccea9be1394 100644
--- a/shrink/bloomz-560m-mic.html
+++ b/shrink/bloomz-560m-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>bigscience/bloomz-560m: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/bloomz-560m-mic@1fa586c833eadf22565c7300a184de964d067039 ← bigscience/bloomz-560m@a2845d7e13dd12efae154a9f1c63fcc2e0cc4b05 · BloomForCausalLM · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">559.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>transformer.h</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Block"></i><i class="c0" title="layer 1: Block"></i><i class="c0" title="layer 2: Block"></i><i class="c0" title="layer 3: Block"></i><i class="c0" title="layer 4: Block"></i><i class="c0" title="layer 5: Block"></i><i class="c0" title="layer 6: Block"></i><i class="c0" title="layer 7: Block"></i><i class="c0" title="layer 8: Block"></i><i class="c0" title="layer 9: Block"></i><i class="c0" title="layer 10: Block"></i><i class="c0" title="layer 11: Block"></i><i class="c0" title="layer 12: Block"></i><i class="c0" title="layer 13: Block"></i><i class="c0" title="layer 14: Block"></i><i class="c0" title="layer 15: Block"></i><i class="c0" title="layer 16: Block"></i><i class="c0" title="layer 17: Block"></i><i class="c0" title="layer 18: Block"></i><i class="c0" title="layer 19: Block"></i><i class="c0" title="layer 20: Block"></i><i class="c0" title="layer 21: Block"></i><i class="c0" title="layer 22: Block"></i><i class="c0" title="layer 23: Block"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Block"></i><i class="c0" title="layer 1: Block"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Block</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.540596" title="transformer.h: 302.3M (54.1%)"></i><i class="c1" style="flex:0.459396" title="transformer.word_embeddings: 256.9M (45.9%)"></i><i class="c2" style="flex:0.000004" title="transformer.word_embeddings_layernorm: 2.0K (0.0%)"></i><i class="c3" style="flex:0.000004" title="transformer.ln_f: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006187" title="transformer.h: 100.0K (0.6%)"></i><i class="c1" style="flex:0.993797" title="transformer.word_embeddings: 16.1M (99.4%)"></i><i class="c2" style="flex:0.000008" title="transformer.word_embeddings_layernorm: 128 (0.0%)"></i><i class="c3" style="flex:0.000008" title="transformer.ln_f: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>transformer.h</td><td>302.3M</td><td>54.1%</td><td>100.0K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>transformer.word_embeddings</td><td>256.9M</td><td>45.9%</td><td>16.1M</td><td>99.4%</td></tr><tr><td><i class="c2"></i>transformer.word_embeddings_layernorm</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c3"></i>transformer.ln_f</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>bloom</td><td>bloom</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>250880</td><td>250880</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Bloom omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>97</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>50</td><td>6</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Block</td><td>24</td><td>2</td></tr><tr class=""><td>Gelu</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>Dropout</td><td>24</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">3 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>n_head</td><td>16</td><td>4</td></tr><tr><td>n_layer</td><td>24</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>bigscience/bloomz-560m</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/bloomz-560m-mic/tree/1fa586c833eadf22565c7300a184de964d067039" title="huggingface-mic/bloomz-560m-mic@1fa586c833eadf22565c7300a184de964d067039">huggingface-mic/bloomz-560m-mic@1fa586c833ea</a> ← <a href="https://huggingface.co/bigscience/bloomz-560m/tree/a2845d7e13dd12efae154a9f1c63fcc2e0cc4b05" title="bigscience/bloomz-560m@a2845d7e13dd12efae154a9f1c63fcc2e0cc4b05">bigscience/bloomz-560m@a2845d7e13dd</a> · BloomForCausalLM · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">559.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>transformer.h</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Block"></i><i class="c0" title="layer 1: Block"></i><i class="c0" title="layer 2: Block"></i><i class="c0" title="layer 3: Block"></i><i class="c0" title="layer 4: Block"></i><i class="c0" title="layer 5: Block"></i><i class="c0" title="layer 6: Block"></i><i class="c0" title="layer 7: Block"></i><i class="c0" title="layer 8: Block"></i><i class="c0" title="layer 9: Block"></i><i class="c0" title="layer 10: Block"></i><i class="c0" title="layer 11: Block"></i><i class="c0" title="layer 12: Block"></i><i class="c0" title="layer 13: Block"></i><i class="c0" title="layer 14: Block"></i><i class="c0" title="layer 15: Block"></i><i class="c0" title="layer 16: Block"></i><i class="c0" title="layer 17: Block"></i><i class="c0" title="layer 18: Block"></i><i class="c0" title="layer 19: Block"></i><i class="c0" title="layer 20: Block"></i><i class="c0" title="layer 21: Block"></i><i class="c0" title="layer 22: Block"></i><i class="c0" title="layer 23: Block"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Block"></i><i class="c0" title="layer 1: Block"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Block</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.540596" title="transformer.h: 302.3M (54.1%)"></i><i class="c1" style="flex:0.459396" title="transformer.word_embeddings: 256.9M (45.9%)"></i><i class="c2" style="flex:0.000004" title="transformer.ln_f: 2.0K (0.0%)"></i><i class="c3" style="flex:0.000004" title="transformer.word_embeddings_layernorm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006187" title="transformer.h: 100.0K (0.6%)"></i><i class="c1" style="flex:0.993797" title="transformer.word_embeddings: 16.1M (99.4%)"></i><i class="c2" style="flex:0.000008" title="transformer.ln_f: 128 (0.0%)"></i><i class="c3" style="flex:0.000008" title="transformer.word_embeddings_layernorm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>transformer.h</td><td>302.3M</td><td>54.1%</td><td>100.0K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>transformer.word_embeddings</td><td>256.9M</td><td>45.9%</td><td>16.1M</td><td>99.4%</td></tr><tr><td><i class="c2"></i>transformer.ln_f</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c3"></i>transformer.word_embeddings_layernorm</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>bloom</td><td>bloom</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>250880</td><td>250880</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Bloom omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>97</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>50</td><td>6</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Block</td><td>24</td><td>2</td></tr><tr class=""><td>Gelu</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>Dropout</td><td>24</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">3 fields changed, 31 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>n_head</td><td>16</td><td>4</td></tr><tr><td>n_layer</td><td>24</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/bloomz-560m-mic.json b/shrink/bloomz-560m-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..695667536a58aa99b6718bef12835b14e2305106
--- /dev/null
+++ b/shrink/bloomz-560m-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/bloomz-560m-mic@1fa586c833eadf22565c7300a184de964d067039", "findings": []}
\ No newline at end of file
diff --git a/shrink/camembert-base-mic.html b/shrink/camembert-base-mic.html
index 4659603d4b2a5da3f1e0f0d2e613494d63a56a6e..6844ed1ae901908077bd37a3e5aa5a859b99c13b 100644
--- a/shrink/camembert-base-mic.html
+++ b/shrink/camembert-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>almanach/camembert-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/camembert-base-mic@20f8ae1a74c39ea94985475b75af352499fe19ff ← almanach/camembert-base@a75967561c78f2aa81cc41045378d3b4ee25af9e · CamembertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>almanach/camembert-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/camembert-base-mic/tree/20f8ae1a74c39ea94985475b75af352499fe19ff" title="huggingface-mic/camembert-base-mic@20f8ae1a74c39ea94985475b75af352499fe19ff">huggingface-mic/camembert-base-mic@20f8ae1a74c3</a> ← <a href="https://huggingface.co/almanach/camembert-base/tree/a75967561c78f2aa81cc41045378d3b4ee25af9e" title="almanach/camembert-base@a75967561c78f2aa81cc41045378d3b4ee25af9e">almanach/camembert-base@a75967561c78</a> · CamembertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">110.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>roberta.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.768642" title="roberta.encoder.layer: 85.1M (76.9%)"></i><i class="c1" style="flex:0.222129" title="lm_head.decoder: 24.6M (22.2%)"></i><i class="c2" style="flex:0.005337" title="lm_head.dense: 590.6K (0.5%)"></i><i class="c3" style="flex:0.003588" title="roberta.embeddings: 397.1K (0.4%)"></i><i class="c4" style="flex:0.000289" title="lm_head: 32.0K (0.0%)"></i><i class="c5" style="flex:0.000014" title="lm_head.layer_norm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.030643" title="roberta.encoder.layer: 66.9K (3.1%)"></i><i class="c1" style="flex:0.937599" title="lm_head.decoder: 2.0M (93.8%)"></i><i class="c2" style="flex:0.001904" title="lm_head.dense: 4.2K (0.2%)"></i><i class="c3" style="flex:0.015146" title="roberta.embeddings: 33.1K (1.5%)"></i><i class="c4" style="flex:0.014650" title="lm_head: 32.0K (1.5%)"></i><i class="c5" style="flex:0.000059" title="lm_head.layer_norm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>roberta.encoder.layer</td><td>85.1M</td><td>76.9%</td><td>66.9K</td><td>3.1%</td></tr><tr><td><i class="c1"></i>lm_head.decoder</td><td>24.6M</td><td>22.2%</td><td>2.0M</td><td>93.8%</td></tr><tr><td><i class="c2"></i>lm_head.dense</td><td>590.6K</td><td>0.5%</td><td>4.2K</td><td>0.2%</td></tr><tr><td><i class="c3"></i>roberta.embeddings</td><td>397.1K</td><td>0.4%</td><td>33.1K</td><td>1.5%</td></tr><tr><td><i class="c4"></i>lm_head</td><td>32.0K</td><td>0.0%</td><td>32.0K</td><td>1.5%</td></tr><tr><td><i class="c5"></i>lm_head.layer_norm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>camembert</td><td>camembert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32005</td><td>32005</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>514</td><td>514</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Camembert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>74</td><td>14</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>LMHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/camembert-base-mic.json b/shrink/camembert-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..2bfb0e0f76c3ac9d66f72ff7c769ebe597cd72b7
--- /dev/null
+++ b/shrink/camembert-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/camembert-base-mic@20f8ae1a74c39ea94985475b75af352499fe19ff", "findings": []}
\ No newline at end of file
diff --git a/shrink/canine-c-mic.html b/shrink/canine-c-mic.html
index e7692c3e8d857fc364e9371a80805af1385f3d90..c288a116be8fab9bb68937b3d1fc404983214d74 100644
--- a/shrink/canine-c-mic.html
+++ b/shrink/canine-c-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/canine-c: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/canine-c-mic@74effe6821673648251eccbcce73a671448d42f9 ← google/canine-c@dc0eaffdff3fa9161613311c7096eeb3e133ee19 · CanineModel · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">132.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>initial_char_encoder.layer</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>final_char_encoder.layer</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.643947" title="encoder.layer: 85.1M (64.4%)"></i><i class="c1" style="flex:0.095265" title="char_embeddings.char_position_embeddings: 12.6M (9.5%)"></i><i class="c2" style="flex:0.053662" title="initial_char_encoder.layer: 7.1M (5.4%)"></i><i class="c3" style="flex:0.053662" title="final_char_encoder.layer: 7.1M (5.4%)"></i><i class="c4" style="flex:0.035730" title="projection.conv: 4.7M (3.6%)"></i><i class="c5" style="flex:0.017868" title="chars_to_molecules.conv: 2.4M (1.8%)"></i><i class="c6" style="flex:0.011908" title="char_embeddings.HashBucketCodepointEmbedder_1: 1.6M (1.2%)"></i><i class="c7" style="flex:0.087956" title="12 others: 11.6M (8.8%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.029286" title="encoder.layer: 66.9K (2.9%)"></i><i class="c1" style="flex:0.458717" title="char_embeddings.char_position_embeddings: 1.0M (45.9%)"></i><i class="c2" style="flex:0.014643" title="initial_char_encoder.layer: 33.5K (1.5%)"></i><i class="c3" style="flex:0.014643" title="final_char_encoder.layer: 33.5K (1.5%)"></i><i class="c4" style="flex:0.014363" title="projection.conv: 32.8K (1.4%)"></i><i class="c5" style="flex:0.007195" title="chars_to_molecules.conv: 16.4K (0.7%)"></i><i class="c6" style="flex:0.057340" title="char_embeddings.HashBucketCodepointEmbedder_1: 131.1K (5.7%)"></i><i class="c7" style="flex:0.403813" title="12 others: 923.1K (40.4%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>encoder.layer</td><td>85.1M</td><td>64.4%</td><td>66.9K</td><td>2.9%</td></tr><tr><td><i class="c1"></i>char_embeddings.char_position_embeddings</td><td>12.6M</td><td>9.5%</td><td>1.0M</td><td>45.9%</td></tr><tr><td><i class="c2"></i>initial_char_encoder.layer</td><td>7.1M</td><td>5.4%</td><td>33.5K</td><td>1.5%</td></tr><tr><td><i class="c3"></i>final_char_encoder.layer</td><td>7.1M</td><td>5.4%</td><td>33.5K</td><td>1.5%</td></tr><tr><td><i class="c4"></i>projection.conv</td><td>4.7M</td><td>3.6%</td><td>32.8K</td><td>1.4%</td></tr><tr><td><i class="c5"></i>chars_to_molecules.conv</td><td>2.4M</td><td>1.8%</td><td>16.4K</td><td>0.7%</td></tr><tr><td><i class="c6"></i>char_embeddings.HashBucketCodepointEmbedder_1</td><td>1.6M</td><td>1.2%</td><td>131.1K</td><td>5.7%</td></tr><tr><td><i class="c7"></i>12 others</td><td>11.6M</td><td>8.8%</td><td>923.1K</td><td>40.4%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>canine</td><td>canine</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>16384</td><td>16384</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Canine omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>85</td><td>25</td></tr><tr class=""><td>Dropout</td><td>44</td><td>14</td></tr><tr class=""><td>LayerNorm</td><td>31</td><td>11</td></tr><tr class=""><td>GELUActivation</td><td>16</td><td>6</td></tr><tr class=""><td>Attention</td><td>14</td><td>4</td></tr><tr class=""><td>Intermediate</td><td>14</td><td>4</td></tr><tr class=""><td>Layer</td><td>14</td><td>4</td></tr><tr class=""><td>Output</td><td>14</td><td>4</td></tr><tr class=""><td>SelfAttention</td><td>14</td><td>4</td></tr><tr class=""><td>SelfOutput</td><td>14</td><td>4</td></tr><tr class=""><td>Embedding</td><td>10</td><td>10</td></tr><tr class=""><td>Encoder</td><td>3</td><td>3</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Conv1d</td><td>2</td><td>2</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr><tr class=""><td>CharactersToMolecules</td><td>1</td><td>1</td></tr><tr class=""><td>ConvProjection</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 24 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/canine-c</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/canine-c-mic/tree/74effe6821673648251eccbcce73a671448d42f9" title="huggingface-mic/canine-c-mic@74effe6821673648251eccbcce73a671448d42f9">huggingface-mic/canine-c-mic@74effe682167</a> ← <a href="https://huggingface.co/google/canine-c/tree/dc0eaffdff3fa9161613311c7096eeb3e133ee19" title="google/canine-c@dc0eaffdff3fa9161613311c7096eeb3e133ee19">google/canine-c@dc0eaffdff3f</a> · CanineModel · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">132.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>initial_char_encoder.layer</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>final_char_encoder.layer</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.643947" title="encoder.layer: 85.1M (64.4%)"></i><i class="c1" style="flex:0.095265" title="char_embeddings.char_position_embeddings: 12.6M (9.5%)"></i><i class="c2" style="flex:0.053662" title="final_char_encoder.layer: 7.1M (5.4%)"></i><i class="c3" style="flex:0.053662" title="initial_char_encoder.layer: 7.1M (5.4%)"></i><i class="c4" style="flex:0.035730" title="projection.conv: 4.7M (3.6%)"></i><i class="c5" style="flex:0.017868" title="chars_to_molecules.conv: 2.4M (1.8%)"></i><i class="c6" style="flex:0.011908" title="char_embeddings.HashBucketCodepointEmbedder_5: 1.6M (1.2%)"></i><i class="c7" style="flex:0.087956" title="12 others: 11.6M (8.8%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.029286" title="encoder.layer: 66.9K (2.9%)"></i><i class="c1" style="flex:0.458717" title="char_embeddings.char_position_embeddings: 1.0M (45.9%)"></i><i class="c2" style="flex:0.014643" title="final_char_encoder.layer: 33.5K (1.5%)"></i><i class="c3" style="flex:0.014643" title="initial_char_encoder.layer: 33.5K (1.5%)"></i><i class="c4" style="flex:0.014363" title="projection.conv: 32.8K (1.4%)"></i><i class="c5" style="flex:0.007195" title="chars_to_molecules.conv: 16.4K (0.7%)"></i><i class="c6" style="flex:0.057340" title="char_embeddings.HashBucketCodepointEmbedder_5: 131.1K (5.7%)"></i><i class="c7" style="flex:0.403813" title="12 others: 923.1K (40.4%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>encoder.layer</td><td>85.1M</td><td>64.4%</td><td>66.9K</td><td>2.9%</td></tr><tr><td><i class="c1"></i>char_embeddings.char_position_embeddings</td><td>12.6M</td><td>9.5%</td><td>1.0M</td><td>45.9%</td></tr><tr><td><i class="c2"></i>final_char_encoder.layer</td><td>7.1M</td><td>5.4%</td><td>33.5K</td><td>1.5%</td></tr><tr><td><i class="c3"></i>initial_char_encoder.layer</td><td>7.1M</td><td>5.4%</td><td>33.5K</td><td>1.5%</td></tr><tr><td><i class="c4"></i>projection.conv</td><td>4.7M</td><td>3.6%</td><td>32.8K</td><td>1.4%</td></tr><tr><td><i class="c5"></i>chars_to_molecules.conv</td><td>2.4M</td><td>1.8%</td><td>16.4K</td><td>0.7%</td></tr><tr><td><i class="c6"></i>char_embeddings.HashBucketCodepointEmbedder_5</td><td>1.6M</td><td>1.2%</td><td>131.1K</td><td>5.7%</td></tr><tr><td><i class="c7"></i>12 others</td><td>11.6M</td><td>8.8%</td><td>923.1K</td><td>40.4%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>canine</td><td>canine</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>16384</td><td>16384</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Canine omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>85</td><td>25</td></tr><tr class=""><td>Dropout</td><td>44</td><td>14</td></tr><tr class=""><td>LayerNorm</td><td>31</td><td>11</td></tr><tr class=""><td>GELUActivation</td><td>16</td><td>6</td></tr><tr class=""><td>Attention</td><td>14</td><td>4</td></tr><tr class=""><td>Intermediate</td><td>14</td><td>4</td></tr><tr class=""><td>Layer</td><td>14</td><td>4</td></tr><tr class=""><td>Output</td><td>14</td><td>4</td></tr><tr class=""><td>SelfAttention</td><td>14</td><td>4</td></tr><tr class=""><td>SelfOutput</td><td>14</td><td>4</td></tr><tr class=""><td>Embedding</td><td>10</td><td>10</td></tr><tr class=""><td>Encoder</td><td>3</td><td>3</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Conv1d</td><td>2</td><td>2</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr><tr class=""><td>CharactersToMolecules</td><td>1</td><td>1</td></tr><tr class=""><td>ConvProjection</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 24 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/canine-c-mic.json b/shrink/canine-c-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..5154d858ae94fd976cf3e72cd31992d59b4a6516
--- /dev/null
+++ b/shrink/canine-c-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/canine-c-mic@74effe6821673648251eccbcce73a671448d42f9", "findings": []}
\ No newline at end of file
diff --git a/shrink/chinese-clip-vit-base-patch16-mic.html b/shrink/chinese-clip-vit-base-patch16-mic.html
index caa66ff032180fac2d4ee19f25a96cec823b6ab2..e3881ef752306f579192d1600402541e5219e8ee 100644
--- a/shrink/chinese-clip-vit-base-patch16-mic.html
+++ b/shrink/chinese-clip-vit-base-patch16-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>OFA-Sys/chinese-clip-vit-base-patch16: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/chinese-clip-vit-base-patch16-mic@e4f19250757a82c065b189e77bbffde2acdfe746 ← OFA-Sys/chinese-clip-vit-base-patch16@36e679e65c2a2fead755ae21162091293ad37834 · ChineseCLIPModel · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">188.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">1.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i><i class="c0" title="layer 2: TextLayer"></i><i class="c0" title="layer 3: TextLayer"></i><i class="c0" title="layer 4: TextLayer"></i><i class="c0" title="layer 5: TextLayer"></i><i class="c0" title="layer 6: TextLayer"></i><i class="c0" title="layer 7: TextLayer"></i><i class="c0" title="layer 8: TextLayer"></i><i class="c0" title="layer 9: TextLayer"></i><i class="c0" title="layer 10: TextLayer"></i><i class="c0" title="layer 11: TextLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: VisionLayer"></i><i class="c0" title="layer 1: VisionLayer"></i><i class="c0" title="layer 2: VisionLayer"></i><i class="c0" title="layer 3: VisionLayer"></i><i class="c0" title="layer 4: VisionLayer"></i><i class="c0" title="layer 5: VisionLayer"></i><i class="c0" title="layer 6: VisionLayer"></i><i class="c0" title="layer 7: VisionLayer"></i><i class="c0" title="layer 8: VisionLayer"></i><i class="c0" title="layer 9: VisionLayer"></i><i class="c0" title="layer 10: VisionLayer"></i><i class="c0" title="layer 11: VisionLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionLayer"></i><i class="c0" title="layer 1: VisionLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.451786" title="text_model.encoder.layer: 85.1M (45.2%)"></i><i class="c1" style="flex:0.451786" title="vision_model.encoder.layers: 85.1M (45.2%)"></i><i class="c2" style="flex:0.088295" title="text_model.embeddings: 16.6M (8.8%)"></i><i class="c3" style="flex:0.003941" title="vision_model.embeddings: 741.9K (0.4%)"></i><i class="c4" style="flex:0.002089" title="text_projection: 393.2K (0.2%)"></i><i class="c5" style="flex:0.002089" title="visual_projection: 393.2K (0.2%)"></i><i class="c6" style="flex:0.000008" title="vision_model.pre_layrnorm: 1.5K (0.0%)"></i><i class="c7" style="flex:0.000008" title="2 others: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.042228" title="text_model.encoder.layer: 66.9K (4.2%)"></i><i class="c1" style="flex:0.042228" title="vision_model.encoder.layers: 66.9K (4.2%)"></i><i class="c2" style="flex:0.873798" title="text_model.embeddings: 1.4M (87.4%)"></i><i class="c3" style="flex:0.038999" title="vision_model.embeddings: 61.8K (3.9%)"></i><i class="c4" style="flex:0.001292" title="text_projection: 2.0K (0.1%)"></i><i class="c5" style="flex:0.001292" title="visual_projection: 2.0K (0.1%)"></i><i class="c6" style="flex:0.000081" title="vision_model.pre_layrnorm: 128 (0.0%)"></i><i class="c7" style="flex:0.000081" title="2 others: 129 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>text_model.encoder.layer</td><td>85.1M</td><td>45.2%</td><td>66.9K</td><td>4.2%</td></tr><tr><td><i class="c1"></i>vision_model.encoder.layers</td><td>85.1M</td><td>45.2%</td><td>66.9K</td><td>4.2%</td></tr><tr><td><i class="c2"></i>text_model.embeddings</td><td>16.6M</td><td>8.8%</td><td>1.4M</td><td>87.4%</td></tr><tr><td><i class="c3"></i>vision_model.embeddings</td><td>741.9K</td><td>0.4%</td><td>61.8K</td><td>3.9%</td></tr><tr><td><i class="c4"></i>text_projection</td><td>393.2K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>visual_projection</td><td>393.2K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>vision_model.pre_layrnorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c7"></i>2 others</td><td>1.5K</td><td>0.0%</td><td>129</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>chinese_clip</td><td>chinese_clip</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>chinese_clip_text_model</td><td>chinese_clip_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>21128</td><td>21128</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>chinese_clip_vision_model</td><td>chinese_clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix ChineseCLIP omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>26</td></tr><tr class=""><td>LayerNorm</td><td>51</td><td>11</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>TextAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextIntermediate</td><td>12</td><td>2</td></tr><tr class=""><td>TextLayer</td><td>12</td><td>2</td></tr><tr class=""><td>TextOutput</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>12</td><td>2</td></tr><tr class=""><td>VisionLayer</td><td>12</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>QuickGELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 74 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>512</td><td>32</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>OFA-Sys/chinese-clip-vit-base-patch16</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/chinese-clip-vit-base-patch16-mic/tree/e4f19250757a82c065b189e77bbffde2acdfe746" title="huggingface-mic/chinese-clip-vit-base-patch16-mic@e4f19250757a82c065b189e77bbffde2acdfe746">huggingface-mic/chinese-clip-vit-base-patch16-mic@e4f19250757a</a> ← <a href="https://huggingface.co/OFA-Sys/chinese-clip-vit-base-patch16/tree/36e679e65c2a2fead755ae21162091293ad37834" title="OFA-Sys/chinese-clip-vit-base-patch16@36e679e65c2a2fead755ae21162091293ad37834">OFA-Sys/chinese-clip-vit-base-patch16@36e679e65c2a</a> · ChineseCLIPModel · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">188.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">1.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i><i class="c0" title="layer 2: TextLayer"></i><i class="c0" title="layer 3: TextLayer"></i><i class="c0" title="layer 4: TextLayer"></i><i class="c0" title="layer 5: TextLayer"></i><i class="c0" title="layer 6: TextLayer"></i><i class="c0" title="layer 7: TextLayer"></i><i class="c0" title="layer 8: TextLayer"></i><i class="c0" title="layer 9: TextLayer"></i><i class="c0" title="layer 10: TextLayer"></i><i class="c0" title="layer 11: TextLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: VisionLayer"></i><i class="c0" title="layer 1: VisionLayer"></i><i class="c0" title="layer 2: VisionLayer"></i><i class="c0" title="layer 3: VisionLayer"></i><i class="c0" title="layer 4: VisionLayer"></i><i class="c0" title="layer 5: VisionLayer"></i><i class="c0" title="layer 6: VisionLayer"></i><i class="c0" title="layer 7: VisionLayer"></i><i class="c0" title="layer 8: VisionLayer"></i><i class="c0" title="layer 9: VisionLayer"></i><i class="c0" title="layer 10: VisionLayer"></i><i class="c0" title="layer 11: VisionLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionLayer"></i><i class="c0" title="layer 1: VisionLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.451786" title="vision_model.encoder.layers: 85.1M (45.2%)"></i><i class="c1" style="flex:0.451786" title="text_model.encoder.layer: 85.1M (45.2%)"></i><i class="c2" style="flex:0.088295" title="text_model.embeddings: 16.6M (8.8%)"></i><i class="c3" style="flex:0.003941" title="vision_model.embeddings: 741.9K (0.4%)"></i><i class="c4" style="flex:0.002089" title="visual_projection: 393.2K (0.2%)"></i><i class="c5" style="flex:0.002089" title="text_projection: 393.2K (0.2%)"></i><i class="c6" style="flex:0.000008" title="vision_model.post_layernorm: 1.5K (0.0%)"></i><i class="c7" style="flex:0.000008" title="2 others: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.042228" title="vision_model.encoder.layers: 66.9K (4.2%)"></i><i class="c1" style="flex:0.042228" title="text_model.encoder.layer: 66.9K (4.2%)"></i><i class="c2" style="flex:0.873798" title="text_model.embeddings: 1.4M (87.4%)"></i><i class="c3" style="flex:0.038999" title="vision_model.embeddings: 61.8K (3.9%)"></i><i class="c4" style="flex:0.001292" title="visual_projection: 2.0K (0.1%)"></i><i class="c5" style="flex:0.001292" title="text_projection: 2.0K (0.1%)"></i><i class="c6" style="flex:0.000081" title="vision_model.post_layernorm: 128 (0.0%)"></i><i class="c7" style="flex:0.000081" title="2 others: 129 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>vision_model.encoder.layers</td><td>85.1M</td><td>45.2%</td><td>66.9K</td><td>4.2%</td></tr><tr><td><i class="c1"></i>text_model.encoder.layer</td><td>85.1M</td><td>45.2%</td><td>66.9K</td><td>4.2%</td></tr><tr><td><i class="c2"></i>text_model.embeddings</td><td>16.6M</td><td>8.8%</td><td>1.4M</td><td>87.4%</td></tr><tr><td><i class="c3"></i>vision_model.embeddings</td><td>741.9K</td><td>0.4%</td><td>61.8K</td><td>3.9%</td></tr><tr><td><i class="c4"></i>visual_projection</td><td>393.2K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>text_projection</td><td>393.2K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>vision_model.post_layernorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c7"></i>2 others</td><td>1.5K</td><td>0.0%</td><td>129</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>chinese_clip</td><td>chinese_clip</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>chinese_clip_text_model</td><td>chinese_clip_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>21128</td><td>21128</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>chinese_clip_vision_model</td><td>chinese_clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix ChineseCLIP omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>26</td></tr><tr class=""><td>LayerNorm</td><td>51</td><td>11</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>TextAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextIntermediate</td><td>12</td><td>2</td></tr><tr class=""><td>TextLayer</td><td>12</td><td>2</td></tr><tr class=""><td>TextOutput</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>12</td><td>2</td></tr><tr class=""><td>VisionLayer</td><td>12</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>QuickGELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 74 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>512</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/chinese-clip-vit-base-patch16-mic.json b/shrink/chinese-clip-vit-base-patch16-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a0f716f798a9b087218b8b63fbb474231fb6824a
--- /dev/null
+++ b/shrink/chinese-clip-vit-base-patch16-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/chinese-clip-vit-base-patch16-mic@e4f19250757a82c065b189e77bbffde2acdfe746", "findings": []}
\ No newline at end of file
diff --git a/shrink/clap-htsat-fused-mic.html b/shrink/clap-htsat-fused-mic.html
index 9ee12b4cf7066e99c91e06b5c76e67da655401b6..24ff8f9cbaf542f55939e7394d338464a8adc8fb 100644
--- a/shrink/clap-htsat-fused-mic.html
+++ b/shrink/clap-htsat-fused-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>laion/clap-htsat-fused: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/clap-htsat-fused-mic@fb8962e8a171e9824f1f2b920fe20ea46a28895c ← laion/clap-htsat-fused@365dea6ef167def6676140ed93bbc43f84dabb28 · ClapModel · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">153.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i><i class="c0" title="layer 2: TextLayer"></i><i class="c0" title="layer 3: TextLayer"></i><i class="c0" title="layer 4: TextLayer"></i><i class="c0" title="layer 5: TextLayer"></i><i class="c0" title="layer 6: TextLayer"></i><i class="c0" title="layer 7: TextLayer"></i><i class="c0" title="layer 8: TextLayer"></i><i class="c0" title="layer 9: TextLayer"></i><i class="c0" title="layer 10: TextLayer"></i><i class="c0" title="layer 11: TextLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>audio_model.audio_encoder.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: AudioPatchMerging"></i><i class="c0" title="layer 1: AudioPatchMerging"></i><i class="c0" title="layer 2: AudioPatchMerging"></i><i class="c1" title="layer 3: no AudioPatchMerging"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: AudioPatchMerging"></i><i class="c0" title="layer 1: AudioPatchMerging"></i><i class="c0" title="layer 2: AudioPatchMerging"></i><i class="c1" title="layer 3: no AudioPatchMerging"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioPatchMerging</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>no AudioPatchMerging</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.554074" title="text_model.encoder.layer: 85.1M (55.4%)"></i><i class="c1" style="flex:0.254063" title="text_model.embeddings: 39.0M (25.4%)"></i><i class="c2" style="flex:0.179346" title="audio_model.audio_encoder.layers: 27.5M (17.9%)"></i><i class="c3" style="flex:0.003847" title="text_model.pooler: 590.6K (0.4%)"></i><i class="c4" style="flex:0.002565" title="audio_projection.linear1: 393.7K (0.3%)"></i><i class="c5" style="flex:0.002565" title="text_projection.linear1: 393.7K (0.3%)"></i><i class="c6" style="flex:0.001711" title="text_projection.linear2: 262.7K (0.2%)"></i><i class="c7" style="flex:0.001829" title="4 others: 280.8K (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.018326" title="text_model.encoder.layer: 66.9K (1.8%)"></i><i class="c1" style="flex:0.889701" title="text_model.embeddings: 3.3M (89.0%)"></i><i class="c2" style="flex:0.088050" title="audio_model.audio_encoder.layers: 321.6K (8.8%)"></i><i class="c3" style="flex:0.001139" title="text_model.pooler: 4.2K (0.1%)"></i><i class="c4" style="flex:0.001130" title="audio_projection.linear1: 4.1K (0.1%)"></i><i class="c5" style="flex:0.000569" title="text_projection.linear1: 2.1K (0.1%)"></i><i class="c6" style="flex:0.000289" title="text_projection.linear2: 1.1K (0.0%)"></i><i class="c7" style="flex:0.000796" title="4 others: 2.9K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>text_model.encoder.layer</td><td>85.1M</td><td>55.4%</td><td>66.9K</td><td>1.8%</td></tr><tr><td><i class="c1"></i>text_model.embeddings</td><td>39.0M</td><td>25.4%</td><td>3.3M</td><td>89.0%</td></tr><tr><td><i class="c2"></i>audio_model.audio_encoder.layers</td><td>27.5M</td><td>17.9%</td><td>321.6K</td><td>8.8%</td></tr><tr><td><i class="c3"></i>text_model.pooler</td><td>590.6K</td><td>0.4%</td><td>4.2K</td><td>0.1%</td></tr><tr><td><i class="c4"></i>audio_projection.linear1</td><td>393.7K</td><td>0.3%</td><td>4.1K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>text_projection.linear1</td><td>393.7K</td><td>0.3%</td><td>2.1K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>text_projection.linear2</td><td>262.7K</td><td>0.2%</td><td>1.1K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>4 others</td><td>280.8K</td><td>0.2%</td><td>2.9K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clap</td><td>clap</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>clap_text_model</td><td>clap_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>50265</td><td>50265</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>514</td><td>514</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>clap_audio_model</td><td>clap_audio_model</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>patch_size</td><td>4</td><td>4</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>64</td><td>64</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Clap omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>152</td><td>44</td></tr><tr class=""><td>Dropout</td><td>73</td><td>19</td></tr><tr class=""><td>LayerNorm</td><td>54</td><td>18</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>6</td></tr><tr class=""><td>AudioAttention</td><td>12</td><td>4</td></tr><tr class=""><td>AudioIntermediate</td><td>12</td><td>4</td></tr><tr class=""><td>AudioLayer</td><td>12</td><td>4</td></tr><tr class=""><td>AudioOutput</td><td>12</td><td>4</td></tr><tr class=""><td>AudioSelfAttention</td><td>12</td><td>4</td></tr><tr class=""><td>AudioSelfOutput</td><td>12</td><td>4</td></tr><tr class=""><td>TextAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextIntermediate</td><td>12</td><td>2</td></tr><tr class=""><td>TextLayer</td><td>12</td><td>2</td></tr><tr class=""><td>TextOutput</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Identity</td><td>12</td><td>4</td></tr><tr class=""><td>Conv2d</td><td>6</td><td>6</td></tr><tr class=""><td>ModuleList</td><td>6</td><td>6</td></tr><tr class=""><td>BatchNorm2d</td><td>5</td><td>5</td></tr><tr class=""><td>AudioStage</td><td>4</td><td>4</td></tr><tr class=""><td>ReLU</td><td>4</td><td>4</td></tr><tr class=""><td>AudioPatchMerging</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>ProjectionLayer</td><td>2</td><td>2</td></tr><tr class=""><td>Sequential</td><td>2</td><td>2</td></tr><tr class=""><td>AdaptiveAvgPool1d</td><td>1</td><td>1</td></tr><tr class=""><td>AdaptiveAvgPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>AudioAFFBlock</td><td>1</td><td>1</td></tr><tr class=""><td>AudioEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>AudioPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextPooler</td><td>1</td><td>1</td></tr><tr class=""><td>Sigmoid</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 106 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.depths</td><td>[2, 2, 6, 2]</td><td>[1, 1, 1, 1]</td></tr><tr><td>audio_config.hidden_size</td><td>768</td><td>128</td></tr><tr><td>audio_config.patch_embeds_hidden_size</td><td>96</td><td>16</td></tr><tr><td>audio_config.projection_dim</td><td>512</td><td>32</td></tr><tr><td>audio_config.projection_hidden_size</td><td>768</td><td>128</td></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>6</td></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.projection_dim</td><td>512</td><td>32</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>laion/clap-htsat-fused</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/clap-htsat-fused-mic/tree/fb8962e8a171e9824f1f2b920fe20ea46a28895c" title="huggingface-mic/clap-htsat-fused-mic@fb8962e8a171e9824f1f2b920fe20ea46a28895c">huggingface-mic/clap-htsat-fused-mic@fb8962e8a171</a> ← <a href="https://huggingface.co/laion/clap-htsat-fused/tree/365dea6ef167def6676140ed93bbc43f84dabb28" title="laion/clap-htsat-fused@365dea6ef167def6676140ed93bbc43f84dabb28">laion/clap-htsat-fused@365dea6ef167</a> · ClapModel · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">153.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i><i class="c0" title="layer 2: TextLayer"></i><i class="c0" title="layer 3: TextLayer"></i><i class="c0" title="layer 4: TextLayer"></i><i class="c0" title="layer 5: TextLayer"></i><i class="c0" title="layer 6: TextLayer"></i><i class="c0" title="layer 7: TextLayer"></i><i class="c0" title="layer 8: TextLayer"></i><i class="c0" title="layer 9: TextLayer"></i><i class="c0" title="layer 10: TextLayer"></i><i class="c0" title="layer 11: TextLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextLayer"></i><i class="c0" title="layer 1: TextLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>audio_model.audio_encoder.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: AudioPatchMerging"></i><i class="c0" title="layer 1: AudioPatchMerging"></i><i class="c0" title="layer 2: AudioPatchMerging"></i><i class="c1" title="layer 3: no AudioPatchMerging"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: AudioPatchMerging"></i><i class="c0" title="layer 1: AudioPatchMerging"></i><i class="c0" title="layer 2: AudioPatchMerging"></i><i class="c1" title="layer 3: no AudioPatchMerging"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioPatchMerging</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>no AudioPatchMerging</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.554074" title="text_model.encoder.layer: 85.1M (55.4%)"></i><i class="c1" style="flex:0.254063" title="text_model.embeddings: 39.0M (25.4%)"></i><i class="c2" style="flex:0.179346" title="audio_model.audio_encoder.layers: 27.5M (17.9%)"></i><i class="c3" style="flex:0.003847" title="text_model.pooler: 590.6K (0.4%)"></i><i class="c4" style="flex:0.002565" title="text_projection.linear1: 393.7K (0.3%)"></i><i class="c5" style="flex:0.002565" title="audio_projection.linear1: 393.7K (0.3%)"></i><i class="c6" style="flex:0.001711" title="text_projection.linear2: 262.7K (0.2%)"></i><i class="c7" style="flex:0.001829" title="4 others: 280.8K (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.018326" title="text_model.encoder.layer: 66.9K (1.8%)"></i><i class="c1" style="flex:0.889701" title="text_model.embeddings: 3.3M (89.0%)"></i><i class="c2" style="flex:0.088050" title="audio_model.audio_encoder.layers: 321.6K (8.8%)"></i><i class="c3" style="flex:0.001139" title="text_model.pooler: 4.2K (0.1%)"></i><i class="c4" style="flex:0.000569" title="text_projection.linear1: 2.1K (0.1%)"></i><i class="c5" style="flex:0.001130" title="audio_projection.linear1: 4.1K (0.1%)"></i><i class="c6" style="flex:0.000289" title="text_projection.linear2: 1.1K (0.0%)"></i><i class="c7" style="flex:0.000796" title="4 others: 2.9K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>text_model.encoder.layer</td><td>85.1M</td><td>55.4%</td><td>66.9K</td><td>1.8%</td></tr><tr><td><i class="c1"></i>text_model.embeddings</td><td>39.0M</td><td>25.4%</td><td>3.3M</td><td>89.0%</td></tr><tr><td><i class="c2"></i>audio_model.audio_encoder.layers</td><td>27.5M</td><td>17.9%</td><td>321.6K</td><td>8.8%</td></tr><tr><td><i class="c3"></i>text_model.pooler</td><td>590.6K</td><td>0.4%</td><td>4.2K</td><td>0.1%</td></tr><tr><td><i class="c4"></i>text_projection.linear1</td><td>393.7K</td><td>0.3%</td><td>2.1K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>audio_projection.linear1</td><td>393.7K</td><td>0.3%</td><td>4.1K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>text_projection.linear2</td><td>262.7K</td><td>0.2%</td><td>1.1K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>4 others</td><td>280.8K</td><td>0.2%</td><td>2.9K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clap</td><td>clap</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>clap_text_model</td><td>clap_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>50265</td><td>50265</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>514</td><td>514</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>clap_audio_model</td><td>clap_audio_model</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>patch_size</td><td>4</td><td>4</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>num_mel_bins</td><td>64</td><td>64</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Clap omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>152</td><td>44</td></tr><tr class=""><td>Dropout</td><td>73</td><td>19</td></tr><tr class=""><td>LayerNorm</td><td>54</td><td>18</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>6</td></tr><tr class=""><td>AudioAttention</td><td>12</td><td>4</td></tr><tr class=""><td>AudioIntermediate</td><td>12</td><td>4</td></tr><tr class=""><td>AudioLayer</td><td>12</td><td>4</td></tr><tr class=""><td>AudioOutput</td><td>12</td><td>4</td></tr><tr class=""><td>AudioSelfAttention</td><td>12</td><td>4</td></tr><tr class=""><td>AudioSelfOutput</td><td>12</td><td>4</td></tr><tr class=""><td>TextAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextIntermediate</td><td>12</td><td>2</td></tr><tr class=""><td>TextLayer</td><td>12</td><td>2</td></tr><tr class=""><td>TextOutput</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>TextSelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Identity</td><td>12</td><td>4</td></tr><tr class=""><td>Conv2d</td><td>6</td><td>6</td></tr><tr class=""><td>ModuleList</td><td>6</td><td>6</td></tr><tr class=""><td>BatchNorm2d</td><td>5</td><td>5</td></tr><tr class=""><td>AudioStage</td><td>4</td><td>4</td></tr><tr class=""><td>ReLU</td><td>4</td><td>4</td></tr><tr class=""><td>AudioPatchMerging</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>ProjectionLayer</td><td>2</td><td>2</td></tr><tr class=""><td>Sequential</td><td>2</td><td>2</td></tr><tr class=""><td>AdaptiveAvgPool1d</td><td>1</td><td>1</td></tr><tr class=""><td>AdaptiveAvgPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>AudioAFFBlock</td><td>1</td><td>1</td></tr><tr class=""><td>AudioEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>AudioPatchEmbed</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextPooler</td><td>1</td><td>1</td></tr><tr class=""><td>Sigmoid</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">13 fields changed, 106 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.depths</td><td>[2, 2, 6, 2]</td><td>[1, 1, 1, 1]</td></tr><tr><td>audio_config.hidden_size</td><td>768</td><td>128</td></tr><tr><td>audio_config.patch_embeds_hidden_size</td><td>96</td><td>16</td></tr><tr><td>audio_config.projection_dim</td><td>512</td><td>32</td></tr><tr><td>audio_config.projection_hidden_size</td><td>768</td><td>128</td></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>6</td></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.projection_dim</td><td>512</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/clap-htsat-fused-mic.json b/shrink/clap-htsat-fused-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..99cdb2f4f2fd16f378052cd5a13638194e4b9a4d
--- /dev/null
+++ b/shrink/clap-htsat-fused-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/clap-htsat-fused-mic@fb8962e8a171e9824f1f2b920fe20ea46a28895c", "findings": []}
\ No newline at end of file
diff --git a/shrink/clip-mic.html b/shrink/clip-mic.html
index 890c00fc01b609f6fc104b1128ba2188c8aa680c..46029abe963a62a9d490cfdf98b96f7d96aea797 100644
--- a/shrink/clip-mic.html
+++ b/shrink/clip-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>openai/clip-vit-base-patch32: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/clip-mic@44c8d8042c6d0355f856d0b8549f5f8270ab9640 ← openai/clip-vit-base-patch32@3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268 · CLIPModel · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>openai/clip-vit-base-patch32</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/clip-mic/tree/44c8d8042c6d0355f856d0b8549f5f8270ab9640" title="huggingface-mic/clip-mic@44c8d8042c6d0355f856d0b8549f5f8270ab9640">huggingface-mic/clip-mic@44c8d8042c6d</a> ← <a href="https://huggingface.co/openai/clip-vit-base-patch32/tree/3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268" title="openai/clip-vit-base-patch32@3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268">openai/clip-vit-base-patch32@3d74acf9a28c</a> · CLIPModel · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">151.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.562242" title="vision_model.encoder.layers: 85.1M (56.2%)"></i><i class="c1" style="flex:0.250061" title="text_model.encoder.layers: 37.8M (25.0%)"></i><i class="c2" style="flex:0.167483" title="text_model.embeddings: 25.3M (16.7%)"></i><i class="c3" style="flex:0.015855" title="vision_model.embeddings: 2.4M (1.6%)"></i><i class="c4" style="flex:0.002599" title="visual_projection: 393.2K (0.3%)"></i><i class="c5" style="flex:0.001733" title="text_projection: 262.1K (0.2%)"></i><i class="c6" style="flex:0.000010" title="vision_model.pre_layrnorm: 1.5K (0.0%)"></i><i class="c7" style="flex:0.000017" title="3 others: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.019098" title="vision_model.encoder.layers: 66.9K (1.9%)"></i><i class="c1" style="flex:0.019098" title="text_model.encoder.layers: 66.9K (1.9%)"></i><i class="c2" style="flex:0.903505" title="text_model.embeddings: 3.2M (90.4%)"></i><i class="c3" style="flex:0.057020" title="vision_model.embeddings: 199.9K (5.7%)"></i><i class="c4" style="flex:0.000584" title="visual_projection: 2.0K (0.1%)"></i><i class="c5" style="flex:0.000584" title="text_projection: 2.0K (0.1%)"></i><i class="c6" style="flex:0.000037" title="vision_model.pre_layrnorm: 128 (0.0%)"></i><i class="c7" style="flex:0.000073" title="3 others: 257 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>vision_model.encoder.layers</td><td>85.1M</td><td>56.2%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c1"></i>text_model.encoder.layers</td><td>37.8M</td><td>25.0%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c2"></i>text_model.embeddings</td><td>25.3M</td><td>16.7%</td><td>3.2M</td><td>90.4%</td></tr><tr><td><i class="c3"></i>vision_model.embeddings</td><td>2.4M</td><td>1.6%</td><td>199.9K</td><td>5.7%</td></tr><tr><td><i class="c4"></i>visual_projection</td><td>393.2K</td><td>0.3%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>text_projection</td><td>262.1K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>vision_model.pre_layrnorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c7"></i>3 others</td><td>2.6K</td><td>0.0%</td><td>257</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clip</td><td>clip</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>clip_text_model</td><td>clip_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>49408</td><td>49408</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>77</td><td>77</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>32</td><td>32</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix CLIP omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>26</td></tr><tr class=""><td>LayerNorm</td><td>51</td><td>11</td></tr><tr class=""><td>Attention</td><td>24</td><td>4</td></tr><tr class=""><td>EncoderLayer</td><td>24</td><td>4</td></tr><tr class=""><td>MLP</td><td>24</td><td>4</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>4</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Encoder</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 80 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.projection_dim</td><td>512</td><td>32</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>512</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/clip-mic.json b/shrink/clip-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..c1560517d4a4e27c52ebfaebe241d72b45309286
--- /dev/null
+++ b/shrink/clip-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/clip-mic@44c8d8042c6d0355f856d0b8549f5f8270ab9640", "findings": []}
\ No newline at end of file
diff --git a/shrink/clip-vit-base-patch32-mic.html b/shrink/clip-vit-base-patch32-mic.html
index 69945ba23e38fb02d26e87da91828784ccfebc4b..82561dfcc363d4e535a3b20cf03652b3ce2f26fe 100644
--- a/shrink/clip-vit-base-patch32-mic.html
+++ b/shrink/clip-vit-base-patch32-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>openai/clip-vit-base-patch32: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/clip-vit-base-patch32-mic@51ae90d9c34bf49fafe02004f9b4b1c2062cd278 ← openai/clip-vit-base-patch32@3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268 · CLIPModel · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">151.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.562242" title="vision_model.encoder.layers: 85.1M (56.2%)"></i><i class="c1" style="flex:0.250061" title="text_model.encoder.layers: 37.8M (25.0%)"></i><i class="c2" style="flex:0.167483" title="text_model.embeddings: 25.3M (16.7%)"></i><i class="c3" style="flex:0.015855" title="vision_model.embeddings: 2.4M (1.6%)"></i><i class="c4" style="flex:0.002599" title="visual_projection: 393.2K (0.3%)"></i><i class="c5" style="flex:0.001733" title="text_projection: 262.1K (0.2%)"></i><i class="c6" style="flex:0.000010" title="vision_model.pre_layrnorm: 1.5K (0.0%)"></i><i class="c7" style="flex:0.000017" title="3 others: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.019098" title="vision_model.encoder.layers: 66.9K (1.9%)"></i><i class="c1" style="flex:0.019098" title="text_model.encoder.layers: 66.9K (1.9%)"></i><i class="c2" style="flex:0.903505" title="text_model.embeddings: 3.2M (90.4%)"></i><i class="c3" style="flex:0.057020" title="vision_model.embeddings: 199.9K (5.7%)"></i><i class="c4" style="flex:0.000584" title="visual_projection: 2.0K (0.1%)"></i><i class="c5" style="flex:0.000584" title="text_projection: 2.0K (0.1%)"></i><i class="c6" style="flex:0.000037" title="vision_model.pre_layrnorm: 128 (0.0%)"></i><i class="c7" style="flex:0.000073" title="3 others: 257 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>vision_model.encoder.layers</td><td>85.1M</td><td>56.2%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c1"></i>text_model.encoder.layers</td><td>37.8M</td><td>25.0%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c2"></i>text_model.embeddings</td><td>25.3M</td><td>16.7%</td><td>3.2M</td><td>90.4%</td></tr><tr><td><i class="c3"></i>vision_model.embeddings</td><td>2.4M</td><td>1.6%</td><td>199.9K</td><td>5.7%</td></tr><tr><td><i class="c4"></i>visual_projection</td><td>393.2K</td><td>0.3%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>text_projection</td><td>262.1K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>vision_model.pre_layrnorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c7"></i>3 others</td><td>2.6K</td><td>0.0%</td><td>257</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clip</td><td>clip</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>clip_text_model</td><td>clip_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>49408</td><td>49408</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>77</td><td>77</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>32</td><td>32</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix CLIP omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>26</td></tr><tr class=""><td>LayerNorm</td><td>51</td><td>11</td></tr><tr class=""><td>Attention</td><td>24</td><td>4</td></tr><tr class=""><td>EncoderLayer</td><td>24</td><td>4</td></tr><tr class=""><td>MLP</td><td>24</td><td>4</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>4</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Encoder</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 80 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.projection_dim</td><td>512</td><td>32</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>512</td><td>32</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>openai/clip-vit-base-patch32</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/clip-vit-base-patch32-mic/tree/51ae90d9c34bf49fafe02004f9b4b1c2062cd278" title="huggingface-mic/clip-vit-base-patch32-mic@51ae90d9c34bf49fafe02004f9b4b1c2062cd278">huggingface-mic/clip-vit-base-patch32-mic@51ae90d9c34b</a> ← <a href="https://huggingface.co/openai/clip-vit-base-patch32/tree/3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268" title="openai/clip-vit-base-patch32@3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268">openai/clip-vit-base-patch32@3d74acf9a28c</a> · CLIPModel · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">151.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>text_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.562242" title="vision_model.encoder.layers: 85.1M (56.2%)"></i><i class="c1" style="flex:0.250061" title="text_model.encoder.layers: 37.8M (25.0%)"></i><i class="c2" style="flex:0.167483" title="text_model.embeddings: 25.3M (16.7%)"></i><i class="c3" style="flex:0.015855" title="vision_model.embeddings: 2.4M (1.6%)"></i><i class="c4" style="flex:0.002599" title="visual_projection: 393.2K (0.3%)"></i><i class="c5" style="flex:0.001733" title="text_projection: 262.1K (0.2%)"></i><i class="c6" style="flex:0.000010" title="vision_model.post_layernorm: 1.5K (0.0%)"></i><i class="c7" style="flex:0.000017" title="3 others: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.019098" title="vision_model.encoder.layers: 66.9K (1.9%)"></i><i class="c1" style="flex:0.019098" title="text_model.encoder.layers: 66.9K (1.9%)"></i><i class="c2" style="flex:0.903505" title="text_model.embeddings: 3.2M (90.4%)"></i><i class="c3" style="flex:0.057020" title="vision_model.embeddings: 199.9K (5.7%)"></i><i class="c4" style="flex:0.000584" title="visual_projection: 2.0K (0.1%)"></i><i class="c5" style="flex:0.000584" title="text_projection: 2.0K (0.1%)"></i><i class="c6" style="flex:0.000037" title="vision_model.post_layernorm: 128 (0.0%)"></i><i class="c7" style="flex:0.000073" title="3 others: 257 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>vision_model.encoder.layers</td><td>85.1M</td><td>56.2%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c1"></i>text_model.encoder.layers</td><td>37.8M</td><td>25.0%</td><td>66.9K</td><td>1.9%</td></tr><tr><td><i class="c2"></i>text_model.embeddings</td><td>25.3M</td><td>16.7%</td><td>3.2M</td><td>90.4%</td></tr><tr><td><i class="c3"></i>vision_model.embeddings</td><td>2.4M</td><td>1.6%</td><td>199.9K</td><td>5.7%</td></tr><tr><td><i class="c4"></i>visual_projection</td><td>393.2K</td><td>0.3%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c5"></i>text_projection</td><td>262.1K</td><td>0.2%</td><td>2.0K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>vision_model.post_layernorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c7"></i>3 others</td><td>2.6K</td><td>0.0%</td><td>257</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clip</td><td>clip</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>clip_text_model</td><td>clip_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>49408</td><td>49408</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>77</td><td>77</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>32</td><td>32</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix CLIP omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>26</td></tr><tr class=""><td>LayerNorm</td><td>51</td><td>11</td></tr><tr class=""><td>Attention</td><td>24</td><td>4</td></tr><tr class=""><td>EncoderLayer</td><td>24</td><td>4</td></tr><tr class=""><td>MLP</td><td>24</td><td>4</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>4</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Encoder</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 80 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.projection_dim</td><td>512</td><td>32</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>512</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/clip-vit-base-patch32-mic.json b/shrink/clip-vit-base-patch32-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..0fbe84fca03a18e6ae3b6ef5c8bf3fd3c2f72b42
--- /dev/null
+++ b/shrink/clip-vit-base-patch32-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/clip-vit-base-patch32-mic@51ae90d9c34bf49fafe02004f9b4b1c2062cd278", "findings": []}
\ No newline at end of file
diff --git a/shrink/clip-vit-base-patch32-vision-mic.html b/shrink/clip-vit-base-patch32-vision-mic.html
index a7f21c695960ce9066a7d9a0e99c20decf316c74..b8861fb9d78e620795438c56ba55eeb1f6a3a62a 100644
--- a/shrink/clip-vit-base-patch32-vision-mic.html
+++ b/shrink/clip-vit-base-patch32-vision-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>kunalkanaujiya/clip-vit-base-patch32-vision: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/clip-vit-base-patch32-vision-mic@b1084cb8a6e683e107ed57c9657c25ddbf107137 ← kunalkanaujiya/clip-vit-base-patch32-vision@4c3ca43dcae1f63f1cda22eeabd99674c894d2b7 · CLIPVisionModelWithProjection · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">87.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">269.1K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i><i class="c0" title="layer 2: CLIPEncoderLayer"></i><i class="c0" title="layer 3: CLIPEncoderLayer"></i><i class="c0" title="layer 4: CLIPEncoderLayer"></i><i class="c0" title="layer 5: CLIPEncoderLayer"></i><i class="c0" title="layer 6: CLIPEncoderLayer"></i><i class="c0" title="layer 7: CLIPEncoderLayer"></i><i class="c0" title="layer 8: CLIPEncoderLayer"></i><i class="c0" title="layer 9: CLIPEncoderLayer"></i><i class="c0" title="layer 10: CLIPEncoderLayer"></i><i class="c0" title="layer 11: CLIPEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>CLIPEncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.968187" title="vision_model.encoder.layers: 85.1M (96.8%)"></i><i class="c1" style="flex:0.027302" title="vision_model.embeddings: 2.4M (2.7%)"></i><i class="c2" style="flex:0.004476" title="visual_projection: 393.2K (0.4%)"></i><i class="c3" style="flex:0.000017" title="vision_model.post_layernorm: 1.5K (0.0%)"></i><i class="c4" style="flex:0.000017" title="vision_model.pre_layrnorm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.248751" title="vision_model.encoder.layers: 66.9K (24.9%)"></i><i class="c1" style="flex:0.742687" title="vision_model.embeddings: 199.9K (74.3%)"></i><i class="c2" style="flex:0.007610" title="visual_projection: 2.0K (0.8%)"></i><i class="c3" style="flex:0.000476" title="vision_model.post_layernorm: 128 (0.0%)"></i><i class="c4" style="flex:0.000476" title="vision_model.pre_layrnorm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>vision_model.encoder.layers</td><td>85.1M</td><td>96.8%</td><td>66.9K</td><td>24.9%</td></tr><tr><td><i class="c1"></i>vision_model.embeddings</td><td>2.4M</td><td>2.7%</td><td>199.9K</td><td>74.3%</td></tr><tr><td><i class="c2"></i>visual_projection</td><td>393.2K</td><td>0.4%</td><td>2.0K</td><td>0.8%</td></tr><tr><td><i class="c3"></i>vision_model.post_layernorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c4"></i>vision_model.pre_layrnorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>32</td><td>32</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix CLIPVision omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>73</td><td>13</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>CLIPAttention</td><td>12</td><td>2</td></tr><tr class=""><td>CLIPEncoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>CLIPMLP</td><td>12</td><td>2</td></tr><tr class=""><td>QuickGELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>CLIPEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModelWithProjection</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>kunalkanaujiya/clip-vit-base-patch32-vision</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/clip-vit-base-patch32-vision-mic/tree/b1084cb8a6e683e107ed57c9657c25ddbf107137" title="huggingface-mic/clip-vit-base-patch32-vision-mic@b1084cb8a6e683e107ed57c9657c25ddbf107137">huggingface-mic/clip-vit-base-patch32-vision-mic@b1084cb8a6e6</a> ← <a href="https://huggingface.co/kunalkanaujiya/clip-vit-base-patch32-vision/tree/4c3ca43dcae1f63f1cda22eeabd99674c894d2b7" title="kunalkanaujiya/clip-vit-base-patch32-vision@4c3ca43dcae1f63f1cda22eeabd99674c894d2b7">kunalkanaujiya/clip-vit-base-patch32-vision@4c3ca43dcae1</a> · CLIPVisionModelWithProjection · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">87.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">269.1K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i><i class="c0" title="layer 2: CLIPEncoderLayer"></i><i class="c0" title="layer 3: CLIPEncoderLayer"></i><i class="c0" title="layer 4: CLIPEncoderLayer"></i><i class="c0" title="layer 5: CLIPEncoderLayer"></i><i class="c0" title="layer 6: CLIPEncoderLayer"></i><i class="c0" title="layer 7: CLIPEncoderLayer"></i><i class="c0" title="layer 8: CLIPEncoderLayer"></i><i class="c0" title="layer 9: CLIPEncoderLayer"></i><i class="c0" title="layer 10: CLIPEncoderLayer"></i><i class="c0" title="layer 11: CLIPEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>CLIPEncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.968187" title="vision_model.encoder.layers: 85.1M (96.8%)"></i><i class="c1" style="flex:0.027302" title="vision_model.embeddings: 2.4M (2.7%)"></i><i class="c2" style="flex:0.004476" title="visual_projection: 393.2K (0.4%)"></i><i class="c3" style="flex:0.000017" title="vision_model.pre_layrnorm: 1.5K (0.0%)"></i><i class="c4" style="flex:0.000017" title="vision_model.post_layernorm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.248751" title="vision_model.encoder.layers: 66.9K (24.9%)"></i><i class="c1" style="flex:0.742687" title="vision_model.embeddings: 199.9K (74.3%)"></i><i class="c2" style="flex:0.007610" title="visual_projection: 2.0K (0.8%)"></i><i class="c3" style="flex:0.000476" title="vision_model.pre_layrnorm: 128 (0.0%)"></i><i class="c4" style="flex:0.000476" title="vision_model.post_layernorm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>vision_model.encoder.layers</td><td>85.1M</td><td>96.8%</td><td>66.9K</td><td>24.9%</td></tr><tr><td><i class="c1"></i>vision_model.embeddings</td><td>2.4M</td><td>2.7%</td><td>199.9K</td><td>74.3%</td></tr><tr><td><i class="c2"></i>visual_projection</td><td>393.2K</td><td>0.4%</td><td>2.0K</td><td>0.8%</td></tr><tr><td><i class="c3"></i>vision_model.pre_layrnorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c4"></i>vision_model.post_layernorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>32</td><td>32</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix CLIPVision omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>73</td><td>13</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>CLIPAttention</td><td>12</td><td>2</td></tr><tr class=""><td>CLIPEncoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>CLIPMLP</td><td>12</td><td>2</td></tr><tr class=""><td>QuickGELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>CLIPEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModelWithProjection</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/clip-vit-base-patch32-vision-mic.json b/shrink/clip-vit-base-patch32-vision-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..8023b1f2a2dad157015cc61ce50e3cd5af7613b1
--- /dev/null
+++ b/shrink/clip-vit-base-patch32-vision-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/clip-vit-base-patch32-vision-mic@b1084cb8a6e683e107ed57c9657c25ddbf107137", "findings": []}
\ No newline at end of file
diff --git a/shrink/clipseg-rd64-refined-mic.html b/shrink/clipseg-rd64-refined-mic.html
index 450a8ba0290472569f27d5df117a90e9233eaf2e..4fb8eaa51d20842bc69a4173a5c6bb1490817427 100644
--- a/shrink/clipseg-rd64-refined-mic.html
+++ b/shrink/clipseg-rd64-refined-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>CIDAS/clipseg-rd64-refined: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/clipseg-rd64-refined-mic@5be0dc861834231292b5d365fffc3d3e1d268cd0 ← CIDAS/clipseg-rd64-refined@999e0328d9e10b484360c477313983f9afdd7050 · CLIPSegForImageSegmentation · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>CIDAS/clipseg-rd64-refined</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/clipseg-rd64-refined-mic/tree/5be0dc861834231292b5d365fffc3d3e1d268cd0" title="huggingface-mic/clipseg-rd64-refined-mic@5be0dc861834231292b5d365fffc3d3e1d268cd0">huggingface-mic/clipseg-rd64-refined-mic@5be0dc861834</a> ← <a href="https://huggingface.co/CIDAS/clipseg-rd64-refined/tree/999e0328d9e10b484360c477313983f9afdd7050" title="CIDAS/clipseg-rd64-refined@999e0328d9e10b484360c477313983f9afdd7050">CIDAS/clipseg-rd64-refined@999e0328d9e1</a> · CLIPSegForImageSegmentation · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">150.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>clip.text_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>clip.vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>decoder.layers</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.564217" title="clip.vision_model.encoder.layers: 85.1M (56.4%)"></i><i class="c1" style="flex:0.250940" title="clip.text_model.encoder.layers: 37.8M (25.1%)"></i><i class="c2" style="flex:0.168078" title="clip.text_model: 25.3M (16.8%)"></i><i class="c3" style="flex:0.005595" title="decoder.layers: 843.5K (0.6%)"></i><i class="c4" style="flex:0.004942" title="clip.vision_model: 745.0K (0.5%)"></i><i class="c5" style="flex:0.002608" title="clip.visual_projection: 393.2K (0.3%)"></i><i class="c6" style="flex:0.001739" title="clip.text_projection: 262.1K (0.2%)"></i><i class="c7" style="flex:0.001881" title="5 others: 283.6K (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.015577" title="clip.vision_model.encoder.layers: 66.9K (1.6%)"></i><i class="c1" style="flex:0.015577" title="clip.text_model.encoder.layers: 66.9K (1.6%)"></i><i class="c2" style="flex:0.736956" title="clip.text_model: 3.2M (73.7%)"></i><i class="c3" style="flex:0.196261" title="decoder.layers: 843.5K (19.6%)"></i><i class="c4" style="flex:0.014445" title="clip.vision_model: 62.1K (1.4%)"></i><i class="c5" style="flex:0.000477" title="clip.visual_projection: 2.0K (0.0%)"></i><i class="c6" style="flex:0.000477" title="clip.text_projection: 2.0K (0.0%)"></i><i class="c7" style="flex:0.020231" title="5 others: 86.9K (2.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>clip.vision_model.encoder.layers</td><td>85.1M</td><td>56.4%</td><td>66.9K</td><td>1.6%</td></tr><tr><td><i class="c1"></i>clip.text_model.encoder.layers</td><td>37.8M</td><td>25.1%</td><td>66.9K</td><td>1.6%</td></tr><tr><td><i class="c2"></i>clip.text_model</td><td>25.3M</td><td>16.8%</td><td>3.2M</td><td>73.7%</td></tr><tr><td><i class="c3"></i>decoder.layers</td><td>843.5K</td><td>0.6%</td><td>843.5K</td><td>19.6%</td></tr><tr><td><i class="c4"></i>clip.vision_model</td><td>745.0K</td><td>0.5%</td><td>62.1K</td><td>1.4%</td></tr><tr><td><i class="c5"></i>clip.visual_projection</td><td>393.2K</td><td>0.3%</td><td>2.0K</td><td>0.0%</td></tr><tr><td><i class="c6"></i>clip.text_projection</td><td>262.1K</td><td>0.2%</td><td>2.0K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>5 others</td><td>283.6K</td><td>0.2%</td><td>86.9K</td><td>2.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>clipseg</td><td>clipseg</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>clipseg_text_model</td><td>clipseg_text_model</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>49408</td><td>49408</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>77</td><td>77</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>clipseg_vision_model</td><td>clipseg_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix CLIPSeg omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>169</td><td>49</td></tr><tr class=""><td>LayerNorm</td><td>57</td><td>17</td></tr><tr class=""><td>Attention</td><td>27</td><td>7</td></tr><tr class=""><td>MLP</td><td>27</td><td>7</td></tr><tr class=""><td>EncoderLayer</td><td>24</td><td>4</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>4</td></tr><tr class=""><td>ReLU</td><td>5</td><td>5</td></tr><tr class=""><td>ModuleList</td><td>4</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Encoder</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>2</td><td>2</td></tr><tr class=""><td>ConvTranspose2d</td><td>2</td><td>2</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForImageSegmentation</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Sequential</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 89 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>extract_layers</td><td>[3, 6, 9]</td><td>[0, 1, 1]</td></tr><tr><td>projection_dim</td><td>512</td><td>32</td></tr><tr><td>text_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/clipseg-rd64-refined-mic.json b/shrink/clipseg-rd64-refined-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e9f286278d789dab522010200ae52de856195970
--- /dev/null
+++ b/shrink/clipseg-rd64-refined-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/clipseg-rd64-refined-mic@5be0dc861834231292b5d365fffc3d3e1d268cd0", "findings": []}
\ No newline at end of file
diff --git a/shrink/conv-bert-base-mic.html b/shrink/conv-bert-base-mic.html
index f187c8b42013eb8df3f3c72ae4b143cfa71e5fb6..d41b45dd05ff3da59d627c96311f6c11908a1227 100644
--- a/shrink/conv-bert-base-mic.html
+++ b/shrink/conv-bert-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>YituTech/conv-bert-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/conv-bert-base-mic@939ec1a4a34866f8fe836950538ad5ba620139ef ← YituTech/conv-bert-base@5cb451936b5c4a96562d8b146de85f64f9cf2c22 · ConvBertModel · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>YituTech/conv-bert-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/conv-bert-base-mic/tree/939ec1a4a34866f8fe836950538ad5ba620139ef" title="huggingface-mic/conv-bert-base-mic@939ec1a4a34866f8fe836950538ad5ba620139ef">huggingface-mic/conv-bert-base-mic@939ec1a4a348</a> ← <a href="https://huggingface.co/YituTech/conv-bert-base/tree/5cb451936b5c4a96562d8b146de85f64f9cf2c22" title="YituTech/conv-bert-base@5cb451936b5c4a96562d8b146de85f64f9cf2c22">YituTech/conv-bert-base@5cb451936b5c</a> · ConvBertModel · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">105.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">24.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.774441" title="encoder.layer: 81.8M (77.4%)"></i><i class="c1" style="flex:0.221809" title="embeddings.word_embeddings: 23.4M (22.2%)"></i><i class="c2" style="flex:0.003721" title="embeddings.position_embeddings: 393.2K (0.4%)"></i><i class="c3" style="flex:0.000015" title="embeddings.token_type_embeddings: 1.5K (0.0%)"></i><i class="c4" style="flex:0.000015" title="embeddings.LayerNorm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.002719" title="encoder.layer: 65.1K (0.3%)"></i><i class="c1" style="flex:0.978681" title="embeddings.word_embeddings: 23.4M (97.9%)"></i><i class="c2" style="flex:0.016417" title="embeddings.position_embeddings: 393.2K (1.6%)"></i><i class="c3" style="flex:0.000064" title="embeddings.token_type_embeddings: 1.5K (0.0%)"></i><i class="c4" style="flex:0.000064" title="embeddings.LayerNorm: 1.5K (0.0%)"></i><i class="c5" style="flex:0.002055" title="embeddings_project: 49.2K (0.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>encoder.layer</td><td>81.8M</td><td>77.4%</td><td>65.1K</td><td>0.3%</td></tr><tr><td><i class="c1"></i>embeddings.word_embeddings</td><td>23.4M</td><td>22.2%</td><td>23.4M</td><td>97.9%</td></tr><tr><td><i class="c2"></i>embeddings.position_embeddings</td><td>393.2K</td><td>0.4%</td><td>393.2K</td><td>1.6%</td></tr><tr><td><i class="c3"></i>embeddings.token_type_embeddings</td><td>1.5K</td><td>0.0%</td><td>1.5K</td><td>0.0%</td></tr><tr><td><i class="c4"></i>embeddings.LayerNorm</td><td>1.5K</td><td>0.0%</td><td>1.5K</td><td>0.0%</td></tr><tr><td><i class="c5"></i>embeddings_project</td><td>0</td><td>0.0%</td><td>49.2K</td><td>0.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>convbert</td><td>convbert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix ConvBert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>96</td><td>17</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>Conv1d</td><td>24</td><td>4</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>SeparableConv1D</td><td>12</td><td>2</td></tr><tr class=""><td>Unfold</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 28 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/conv-bert-base-mic.json b/shrink/conv-bert-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a84e265714484a21681fcaf83fe7aed2bdd50293
--- /dev/null
+++ b/shrink/conv-bert-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/conv-bert-base-mic@939ec1a4a34866f8fe836950538ad5ba620139ef", "findings": []}
\ No newline at end of file
diff --git a/shrink/convnext-tiny-224-mic.html b/shrink/convnext-tiny-224-mic.html
index 96c413cea742ccd44d8dce19f40b083b1eb2a4fa..619d8ac3ed2b9e594bd129cbe775c9ffebeb475c 100644
--- a/shrink/convnext-tiny-224-mic.html
+++ b/shrink/convnext-tiny-224-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/convnext-tiny-224: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/convnext-tiny-224-mic@f5dae79dc4935e97ae29a9414541fa8d790454ee ← facebook/convnext-tiny-224@6166b7613034066690a621d8bf25ffdf181a34f0 · ConvNextForImageClassification · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/convnext-tiny-224</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/convnext-tiny-224-mic/tree/f5dae79dc4935e97ae29a9414541fa8d790454ee" title="huggingface-mic/convnext-tiny-224-mic@f5dae79dc4935e97ae29a9414541fa8d790454ee">huggingface-mic/convnext-tiny-224-mic@f5dae79dc493</a> ← <a href="https://huggingface.co/facebook/convnext-tiny-224/tree/6166b7613034066690a621d8bf25ffdf181a34f0" title="facebook/convnext-tiny-224@6166b7613034066690a621d8bf25ffdf181a34f0">facebook/convnext-tiny-224@6166b7613034</a> · ConvNextForImageClassification · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">28.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">361.5K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>convnext.encoder.stages</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.972877" title="convnext.encoder.stages: 27.8M (97.3%)"></i><i class="c1" style="flex:0.026898" title="classifier: 769.0K (2.7%)"></i><i class="c2" style="flex:0.000171" title="convnext.embeddings: 4.9K (0.0%)"></i><i class="c3" style="flex:0.000054" title="convnext.layernorm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.640216" title="convnext.encoder.stages: 231.5K (64.0%)"></i><i class="c1" style="flex:0.356819" title="classifier: 129.0K (35.7%)"></i><i class="c2" style="flex:0.002257" title="convnext.embeddings: 816 (0.2%)"></i><i class="c3" style="flex:0.000708" title="convnext.layernorm: 256 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>convnext.encoder.stages</td><td>27.8M</td><td>97.3%</td><td>231.5K</td><td>64.0%</td></tr><tr><td><i class="c1"></i>classifier</td><td>769.0K</td><td>2.7%</td><td>129.0K</td><td>35.7%</td></tr><tr><td><i class="c2"></i>convnext.embeddings</td><td>4.9K</td><td>0.0%</td><td>816</td><td>0.2%</td></tr><tr><td><i class="c3"></i>convnext.layernorm</td><td>1.5K</td><td>0.0%</td><td>256</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>convnext</td><td>convnext</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>4</td><td>4</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix ConvNext omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>37</td><td>9</td></tr><tr class=""><td>Conv2d</td><td>22</td><td>8</td></tr><tr class=""><td>LayerNorm</td><td>22</td><td>8</td></tr><tr class=""><td>Layer</td><td>18</td><td>4</td></tr><tr class=""><td>GELUActivation</td><td>18</td><td>4</td></tr><tr class=""><td>Identity</td><td>18</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>9</td><td>9</td></tr><tr class=""><td>Stage</td><td>4</td><td>4</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForImageClassification</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>LayerNorm</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">2 fields changed, 21 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>depths</td><td>[3, 3, 9, 3]</td><td>[1, 1, 1, 1]</td></tr><tr><td>hidden_sizes</td><td>[96, 192, 384, 768]</td><td>[16, 32, 64, 128]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/convnext-tiny-224-mic.json b/shrink/convnext-tiny-224-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a375507a475632f20a3b6253d621ebef642e36d6
--- /dev/null
+++ b/shrink/convnext-tiny-224-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/convnext-tiny-224-mic@f5dae79dc4935e97ae29a9414541fa8d790454ee", "findings": []}
\ No newline at end of file
diff --git a/shrink/convnextv2-nano-1k-224-mic.html b/shrink/convnextv2-nano-1k-224-mic.html
index 8a82dda42f7e0f5f2666be925c3bb5fa47679545..f15f89f14b151c74612e466061505bd383e89357 100644
--- a/shrink/convnextv2-nano-1k-224-mic.html
+++ b/shrink/convnextv2-nano-1k-224-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/convnextv2-nano-1k-224: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/convnextv2-nano-1k-224-mic@df320ec01c275577e0266bbffdecd6eec693d41b ← facebook/convnextv2-nano-1k-224@5fd38629f57580835f733e36a7da810d5c8df9ec · ConvNextV2ForImageClassification · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/convnextv2-nano-1k-224</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/convnextv2-nano-1k-224-mic/tree/df320ec01c275577e0266bbffdecd6eec693d41b" title="huggingface-mic/convnextv2-nano-1k-224-mic@df320ec01c275577e0266bbffdecd6eec693d41b">huggingface-mic/convnextv2-nano-1k-224-mic@df320ec01c27</a> ← <a href="https://huggingface.co/facebook/convnextv2-nano-1k-224/tree/5fd38629f57580835f733e36a7da810d5c8df9ec" title="facebook/convnextv2-nano-1k-224@5fd38629f57580835f733e36a7da810d5c8df9ec">facebook/convnextv2-nano-1k-224@5fd38629f575</a> · ConvNextV2ForImageClassification · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">15.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">363.2K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>convnextv2.encoder.stages</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.958630" title="convnextv2.encoder.stages: 15.0M (95.9%)"></i><i class="c1" style="flex:0.041027" title="classifier: 641.0K (4.1%)"></i><i class="c2" style="flex:0.000261" title="convnextv2.embeddings: 4.1K (0.0%)"></i><i class="c3" style="flex:0.000082" title="convnextv2.layernorm: 1.3K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.641880" title="convnextv2.encoder.stages: 233.1K (64.2%)"></i><i class="c1" style="flex:0.355168" title="classifier: 129.0K (35.5%)"></i><i class="c2" style="flex:0.002247" title="convnextv2.embeddings: 816 (0.2%)"></i><i class="c3" style="flex:0.000705" title="convnextv2.layernorm: 256 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>convnextv2.encoder.stages</td><td>15.0M</td><td>95.9%</td><td>233.1K</td><td>64.2%</td></tr><tr><td><i class="c1"></i>classifier</td><td>641.0K</td><td>4.1%</td><td>129.0K</td><td>35.5%</td></tr><tr><td><i class="c2"></i>convnextv2.embeddings</td><td>4.1K</td><td>0.0%</td><td>816</td><td>0.2%</td></tr><tr><td><i class="c3"></i>convnextv2.layernorm</td><td>1.3K</td><td>0.0%</td><td>256</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>convnextv2</td><td>convnextv2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>4</td><td>4</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix ConvNextV2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>29</td><td>9</td></tr><tr class=""><td>Conv2d</td><td>18</td><td>8</td></tr><tr class=""><td>LayerNorm</td><td>18</td><td>8</td></tr><tr class=""><td>GRN</td><td>14</td><td>4</td></tr><tr class=""><td>Layer</td><td>14</td><td>4</td></tr><tr class=""><td>GELUActivation</td><td>14</td><td>4</td></tr><tr class=""><td>Identity</td><td>14</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>9</td><td>9</td></tr><tr class=""><td>Stage</td><td>4</td><td>4</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForImageClassification</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>LayerNorm</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">2 fields changed, 20 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>depths</td><td>[2, 2, 8, 2]</td><td>[1, 1, 1, 1]</td></tr><tr><td>hidden_sizes</td><td>[80, 160, 320, 640]</td><td>[16, 32, 64, 128]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/convnextv2-nano-1k-224-mic.json b/shrink/convnextv2-nano-1k-224-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a405aafb378576b650e322dd9e559cdff914b6b1
--- /dev/null
+++ b/shrink/convnextv2-nano-1k-224-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/convnextv2-nano-1k-224-mic@df320ec01c275577e0266bbffdecd6eec693d41b", "findings": []}
\ No newline at end of file
diff --git a/shrink/csm-1b-mic.html b/shrink/csm-1b-mic.html
index 14ac9a1a9fa957b7d2c7dac67239d6729b92f961..d379018ff35ef185afea9180dc95523410b2ba0c 100644
--- a/shrink/csm-1b-mic.html
+++ b/shrink/csm-1b-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>unsloth/csm-1b: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/csm-1b-mic@ceb9f9aebdb8fe8d8aca3a3c51091ae2527893ca ← unsloth/csm-1b@39e8c756bec133ad7eb9ea1097c84a2bd891c949 · CsmForConditionalGeneration · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.6B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">39.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">6</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>backbone_model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>depth_decoder.model.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>4</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.encoder.encoder_transformer.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i><i class="c0" title="layer 2: MimiTransformerLayer"></i><i class="c0" title="layer 3: MimiTransformerLayer"></i><i class="c0" title="layer 4: MimiTransformerLayer"></i><i class="c0" title="layer 5: MimiTransformerLayer"></i><i class="c0" title="layer 6: MimiTransformerLayer"></i><i class="c0" title="layer 7: MimiTransformerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiTransformerLayer</td><td>8</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.decoder.decoder_transformer.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i><i class="c0" title="layer 2: MimiTransformerLayer"></i><i class="c0" title="layer 3: MimiTransformerLayer"></i><i class="c0" title="layer 4: MimiTransformerLayer"></i><i class="c0" title="layer 5: MimiTransformerLayer"></i><i class="c0" title="layer 6: MimiTransformerLayer"></i><i class="c0" title="layer 7: MimiTransformerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiTransformerLayer</td><td>8</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.quantizer.semantic_residual_vector_quantizer.layers</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiEuclideanCodebook</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.quantizer.acoustic_residual_vector_quantizer.layers</h3><div class="row"><span class="lab">Original · 31</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i><i class="c0" title="layer 1: MimiEuclideanCodebook"></i><i class="c0" title="layer 2: MimiEuclideanCodebook"></i><i class="c0" title="layer 3: MimiEuclideanCodebook"></i><i class="c0" title="layer 4: MimiEuclideanCodebook"></i><i class="c0" title="layer 5: MimiEuclideanCodebook"></i><i class="c0" title="layer 6: MimiEuclideanCodebook"></i><i class="c0" title="layer 7: MimiEuclideanCodebook"></i><i class="c0" title="layer 8: MimiEuclideanCodebook"></i><i class="c0" title="layer 9: MimiEuclideanCodebook"></i><i class="c0" title="layer 10: MimiEuclideanCodebook"></i><i class="c0" title="layer 11: MimiEuclideanCodebook"></i><i class="c0" title="layer 12: MimiEuclideanCodebook"></i><i class="c0" title="layer 13: MimiEuclideanCodebook"></i><i class="c0" title="layer 14: MimiEuclideanCodebook"></i><i class="c0" title="layer 15: MimiEuclideanCodebook"></i><i class="c0" title="layer 16: MimiEuclideanCodebook"></i><i class="c0" title="layer 17: MimiEuclideanCodebook"></i><i class="c0" title="layer 18: MimiEuclideanCodebook"></i><i class="c0" title="layer 19: MimiEuclideanCodebook"></i><i class="c0" title="layer 20: MimiEuclideanCodebook"></i><i class="c0" title="layer 21: MimiEuclideanCodebook"></i><i class="c0" title="layer 22: MimiEuclideanCodebook"></i><i class="c0" title="layer 23: MimiEuclideanCodebook"></i><i class="c0" title="layer 24: MimiEuclideanCodebook"></i><i class="c0" title="layer 25: MimiEuclideanCodebook"></i><i class="c0" title="layer 26: MimiEuclideanCodebook"></i><i class="c0" title="layer 27: MimiEuclideanCodebook"></i><i class="c0" title="layer 28: MimiEuclideanCodebook"></i><i class="c0" title="layer 29: MimiEuclideanCodebook"></i><i class="c0" title="layer 30: MimiEuclideanCodebook"></i></div></div><div class="row"><span class="lab">Fixture · 31</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i><i class="c0" title="layer 1: MimiEuclideanCodebook"></i><i class="c0" title="layer 2: MimiEuclideanCodebook"></i><i class="c0" title="layer 3: MimiEuclideanCodebook"></i><i class="c0" title="layer 4: MimiEuclideanCodebook"></i><i class="c0" title="layer 5: MimiEuclideanCodebook"></i><i class="c0" title="layer 6: MimiEuclideanCodebook"></i><i class="c0" title="layer 7: MimiEuclideanCodebook"></i><i class="c0" title="layer 8: MimiEuclideanCodebook"></i><i class="c0" title="layer 9: MimiEuclideanCodebook"></i><i class="c0" title="layer 10: MimiEuclideanCodebook"></i><i class="c0" title="layer 11: MimiEuclideanCodebook"></i><i class="c0" title="layer 12: MimiEuclideanCodebook"></i><i class="c0" title="layer 13: MimiEuclideanCodebook"></i><i class="c0" title="layer 14: MimiEuclideanCodebook"></i><i class="c0" title="layer 15: MimiEuclideanCodebook"></i><i class="c0" title="layer 16: MimiEuclideanCodebook"></i><i class="c0" title="layer 17: MimiEuclideanCodebook"></i><i class="c0" title="layer 18: MimiEuclideanCodebook"></i><i class="c0" title="layer 19: MimiEuclideanCodebook"></i><i class="c0" title="layer 20: MimiEuclideanCodebook"></i><i class="c0" title="layer 21: MimiEuclideanCodebook"></i><i class="c0" title="layer 22: MimiEuclideanCodebook"></i><i class="c0" title="layer 23: MimiEuclideanCodebook"></i><i class="c0" title="layer 24: MimiEuclideanCodebook"></i><i class="c0" title="layer 25: MimiEuclideanCodebook"></i><i class="c0" title="layer 26: MimiEuclideanCodebook"></i><i class="c0" title="layer 27: MimiEuclideanCodebook"></i><i class="c0" title="layer 28: MimiEuclideanCodebook"></i><i class="c0" title="layer 29: MimiEuclideanCodebook"></i><i class="c0" title="layer 30: MimiEuclideanCodebook"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiEuclideanCodebook</td><td>31</td><td>31</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.596253" title="backbone_model.layers: 973.1M (59.6%)"></i><i class="c1" style="flex:0.160939" title="embed_text_tokens: 262.7M (16.1%)"></i><i class="c2" style="flex:0.082357" title="backbone_model.embed_tokens: 134.4M (8.2%)"></i><i class="c3" style="flex:0.068107" title="depth_decoder.model.layers: 111.2M (6.8%)"></i><i class="c4" style="flex:0.039892" title="depth_decoder.codebooks_head: 65.1M (4.0%)"></i><i class="c5" style="flex:0.015434" title="codec_model.encoder.encoder_transformer.layers: 25.2M (1.5%)"></i><i class="c6" style="flex:0.015434" title="codec_model.decoder.decoder_transformer.layers: 25.2M (1.5%)"></i><i class="c7" style="flex:0.021585" title="6 others: 35.2M (2.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.001862" title="backbone_model.layers: 74.0K (0.2%)"></i><i class="c1" style="flex:0.206546" title="embed_text_tokens: 8.2M (20.7%)"></i><i class="c2" style="flex:0.105695" title="backbone_model.embed_tokens: 4.2M (10.6%)"></i><i class="c3" style="flex:0.001862" title="depth_decoder.model.layers: 74.0K (0.2%)"></i><i class="c4" style="flex:0.102392" title="depth_decoder.codebooks_head: 4.1M (10.2%)"></i><i class="c5" style="flex:0.001668" title="codec_model.encoder.encoder_transformer.layers: 66.3K (0.2%)"></i><i class="c6" style="flex:0.001668" title="codec_model.decoder.decoder_transformer.layers: 66.3K (0.2%)"></i><i class="c7" style="flex:0.578307" title="6 others: 23.0M (57.8%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>backbone_model.layers</td><td>973.1M</td><td>59.6%</td><td>74.0K</td><td>0.2%</td></tr><tr><td><i class="c1"></i>embed_text_tokens</td><td>262.7M</td><td>16.1%</td><td>8.2M</td><td>20.7%</td></tr><tr><td><i class="c2"></i>backbone_model.embed_tokens</td><td>134.4M</td><td>8.2%</td><td>4.2M</td><td>10.6%</td></tr><tr><td><i class="c3"></i>depth_decoder.model.layers</td><td>111.2M</td><td>6.8%</td><td>74.0K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>depth_decoder.codebooks_head</td><td>65.1M</td><td>4.0%</td><td>4.1M</td><td>10.2%</td></tr><tr><td><i class="c5"></i>codec_model.encoder.encoder_transformer.layers</td><td>25.2M</td><td>1.5%</td><td>66.3K</td><td>0.2%</td></tr><tr><td><i class="c6"></i>codec_model.decoder.decoder_transformer.layers</td><td>25.2M</td><td>1.5%</td><td>66.3K</td><td>0.2%</td></tr><tr><td><i class="c7"></i>6 others</td><td>35.2M</td><td>2.2%</td><td>23.0M</td><td>57.8%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>llama3</td><td>llama3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>csm</td><td>csm</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>2051</td><td>2051</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>RoPE type</td><td>llama3</td><td>llama3</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>model_type</td><td>csm_depth_decoder_model</td><td>csm_depth_decoder_model</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>vocab_size</td><td>2051</td><td>2051</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>max_position_embeddings</td><td>33</td><td>33</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>attention style</td><td>MHA (8 heads, 8 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>model_type</td><td>mimi</td><td>mimi</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>max_position_embeddings</td><td>8000</td><td>8000</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>sampling_rate</td><td>24000</td><td>24000</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Csm omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>238</td><td>54</td></tr><tr class=""><td>RMSNorm</td><td>42</td><td>10</td></tr><tr class=""><td>LayerNorm</td><td>32</td><td>8</td></tr><tr class=""><td>MimiEuclideanCodebook</td><td>32</td><td>32</td></tr><tr class=""><td>MimiLayerScale</td><td>32</td><td>8</td></tr><tr class=""><td>MimiVectorQuantization</td><td>32</td><td>32</td></tr><tr class=""><td>Conv1d</td><td>29</td><td>29</td></tr><tr class=""><td>ELU</td><td>26</td><td>26</td></tr><tr class=""><td>MimiConv1d</td><td>25</td><td>25</td></tr><tr class=""><td>Attention</td><td>20</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>20</td><td>4</td></tr><tr class=""><td>MLP</td><td>20</td><td>4</td></tr><tr class=""><td>SiLUActivation</td><td>20</td><td>4</td></tr><tr class=""><td>GELUActivation</td><td>16</td><td>4</td></tr><tr class=""><td>MimiAttention</td><td>16</td><td>4</td></tr><tr class=""><td>MimiMLP</td><td>16</td><td>4</td></tr><tr class=""><td>MimiTransformerLayer</td><td>16</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>16</td><td>16</td></tr><tr class=""><td>Identity</td><td>8</td><td>8</td></tr><tr class=""><td>MimiResnetBlock</td><td>8</td><td>8</td></tr><tr class=""><td>ConvTranspose1d</td><td>5</td><td>5</td></tr><tr class=""><td>MimiConvTranspose1d</td><td>5</td><td>5</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>RotaryEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>MimiResidualVectorQuantizer</td><td>2</td><td>2</td></tr><tr class=""><td>MimiRotaryEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>MimiTransformerModel</td><td>2</td><td>2</td></tr><tr class=""><td>BackboneModel</td><td>1</td><td>1</td></tr><tr class=""><td>BackboneModelEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>CodebooksHead</td><td>1</td><td>1</td></tr><tr class=""><td>DepthDecoderForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>DepthDecoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>MimiDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>MimiDecoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>MimiEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>MimiEncoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>MimiModel</td><td>1</td><td>1</td></tr><tr class=""><td>MimiSplitResidualVectorQuantizer</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">19 fields changed, 96 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>codec_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>codec_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>codec_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>codec_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>codec_config.num_hidden_layers</td><td>8</td><td>2</td></tr><tr><td>codec_config.num_key_value_heads</td><td>8</td><td>4</td></tr><tr><td>codec_config.upsample_groups</td><td>512</td><td>64</td></tr><tr><td>depth_decoder_config.backbone_hidden_size</td><td>2048</td><td>64</td></tr><tr><td>depth_decoder_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>depth_decoder_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>depth_decoder_config.intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>depth_decoder_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>depth_decoder_config.num_hidden_layers</td><td>4</td><td>2</td></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>unsloth/csm-1b</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/csm-1b-mic/tree/ceb9f9aebdb8fe8d8aca3a3c51091ae2527893ca" title="huggingface-mic/csm-1b-mic@ceb9f9aebdb8fe8d8aca3a3c51091ae2527893ca">huggingface-mic/csm-1b-mic@ceb9f9aebdb8</a> ← <a href="https://huggingface.co/unsloth/csm-1b/tree/39e8c756bec133ad7eb9ea1097c84a2bd891c949" title="unsloth/csm-1b@39e8c756bec133ad7eb9ea1097c84a2bd891c949">unsloth/csm-1b@39e8c756bec1</a> · CsmForConditionalGeneration · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.6B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">39.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">6</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>backbone_model.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>depth_decoder.model.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>4</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.encoder.encoder_transformer.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i><i class="c0" title="layer 2: MimiTransformerLayer"></i><i class="c0" title="layer 3: MimiTransformerLayer"></i><i class="c0" title="layer 4: MimiTransformerLayer"></i><i class="c0" title="layer 5: MimiTransformerLayer"></i><i class="c0" title="layer 6: MimiTransformerLayer"></i><i class="c0" title="layer 7: MimiTransformerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiTransformerLayer</td><td>8</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.decoder.decoder_transformer.layers</h3><div class="row"><span class="lab">Original · 8</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i><i class="c0" title="layer 2: MimiTransformerLayer"></i><i class="c0" title="layer 3: MimiTransformerLayer"></i><i class="c0" title="layer 4: MimiTransformerLayer"></i><i class="c0" title="layer 5: MimiTransformerLayer"></i><i class="c0" title="layer 6: MimiTransformerLayer"></i><i class="c0" title="layer 7: MimiTransformerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: MimiTransformerLayer"></i><i class="c0" title="layer 1: MimiTransformerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiTransformerLayer</td><td>8</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.quantizer.semantic_residual_vector_quantizer.layers</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiEuclideanCodebook</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>codec_model.quantizer.acoustic_residual_vector_quantizer.layers</h3><div class="row"><span class="lab">Original · 31</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i><i class="c0" title="layer 1: MimiEuclideanCodebook"></i><i class="c0" title="layer 2: MimiEuclideanCodebook"></i><i class="c0" title="layer 3: MimiEuclideanCodebook"></i><i class="c0" title="layer 4: MimiEuclideanCodebook"></i><i class="c0" title="layer 5: MimiEuclideanCodebook"></i><i class="c0" title="layer 6: MimiEuclideanCodebook"></i><i class="c0" title="layer 7: MimiEuclideanCodebook"></i><i class="c0" title="layer 8: MimiEuclideanCodebook"></i><i class="c0" title="layer 9: MimiEuclideanCodebook"></i><i class="c0" title="layer 10: MimiEuclideanCodebook"></i><i class="c0" title="layer 11: MimiEuclideanCodebook"></i><i class="c0" title="layer 12: MimiEuclideanCodebook"></i><i class="c0" title="layer 13: MimiEuclideanCodebook"></i><i class="c0" title="layer 14: MimiEuclideanCodebook"></i><i class="c0" title="layer 15: MimiEuclideanCodebook"></i><i class="c0" title="layer 16: MimiEuclideanCodebook"></i><i class="c0" title="layer 17: MimiEuclideanCodebook"></i><i class="c0" title="layer 18: MimiEuclideanCodebook"></i><i class="c0" title="layer 19: MimiEuclideanCodebook"></i><i class="c0" title="layer 20: MimiEuclideanCodebook"></i><i class="c0" title="layer 21: MimiEuclideanCodebook"></i><i class="c0" title="layer 22: MimiEuclideanCodebook"></i><i class="c0" title="layer 23: MimiEuclideanCodebook"></i><i class="c0" title="layer 24: MimiEuclideanCodebook"></i><i class="c0" title="layer 25: MimiEuclideanCodebook"></i><i class="c0" title="layer 26: MimiEuclideanCodebook"></i><i class="c0" title="layer 27: MimiEuclideanCodebook"></i><i class="c0" title="layer 28: MimiEuclideanCodebook"></i><i class="c0" title="layer 29: MimiEuclideanCodebook"></i><i class="c0" title="layer 30: MimiEuclideanCodebook"></i></div></div><div class="row"><span class="lab">Fixture · 31</span><div class="cells"><i class="c0" title="layer 0: MimiEuclideanCodebook"></i><i class="c0" title="layer 1: MimiEuclideanCodebook"></i><i class="c0" title="layer 2: MimiEuclideanCodebook"></i><i class="c0" title="layer 3: MimiEuclideanCodebook"></i><i class="c0" title="layer 4: MimiEuclideanCodebook"></i><i class="c0" title="layer 5: MimiEuclideanCodebook"></i><i class="c0" title="layer 6: MimiEuclideanCodebook"></i><i class="c0" title="layer 7: MimiEuclideanCodebook"></i><i class="c0" title="layer 8: MimiEuclideanCodebook"></i><i class="c0" title="layer 9: MimiEuclideanCodebook"></i><i class="c0" title="layer 10: MimiEuclideanCodebook"></i><i class="c0" title="layer 11: MimiEuclideanCodebook"></i><i class="c0" title="layer 12: MimiEuclideanCodebook"></i><i class="c0" title="layer 13: MimiEuclideanCodebook"></i><i class="c0" title="layer 14: MimiEuclideanCodebook"></i><i class="c0" title="layer 15: MimiEuclideanCodebook"></i><i class="c0" title="layer 16: MimiEuclideanCodebook"></i><i class="c0" title="layer 17: MimiEuclideanCodebook"></i><i class="c0" title="layer 18: MimiEuclideanCodebook"></i><i class="c0" title="layer 19: MimiEuclideanCodebook"></i><i class="c0" title="layer 20: MimiEuclideanCodebook"></i><i class="c0" title="layer 21: MimiEuclideanCodebook"></i><i class="c0" title="layer 22: MimiEuclideanCodebook"></i><i class="c0" title="layer 23: MimiEuclideanCodebook"></i><i class="c0" title="layer 24: MimiEuclideanCodebook"></i><i class="c0" title="layer 25: MimiEuclideanCodebook"></i><i class="c0" title="layer 26: MimiEuclideanCodebook"></i><i class="c0" title="layer 27: MimiEuclideanCodebook"></i><i class="c0" title="layer 28: MimiEuclideanCodebook"></i><i class="c0" title="layer 29: MimiEuclideanCodebook"></i><i class="c0" title="layer 30: MimiEuclideanCodebook"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MimiEuclideanCodebook</td><td>31</td><td>31</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.596253" title="backbone_model.layers: 973.1M (59.6%)"></i><i class="c1" style="flex:0.160939" title="embed_text_tokens: 262.7M (16.1%)"></i><i class="c2" style="flex:0.082357" title="backbone_model.embed_tokens: 134.4M (8.2%)"></i><i class="c3" style="flex:0.068107" title="depth_decoder.model.layers: 111.2M (6.8%)"></i><i class="c4" style="flex:0.039892" title="depth_decoder.codebooks_head: 65.1M (4.0%)"></i><i class="c5" style="flex:0.015434" title="codec_model.decoder.decoder_transformer.layers: 25.2M (1.5%)"></i><i class="c6" style="flex:0.015434" title="codec_model.encoder.encoder_transformer.layers: 25.2M (1.5%)"></i><i class="c7" style="flex:0.021585" title="6 others: 35.2M (2.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.001862" title="backbone_model.layers: 74.0K (0.2%)"></i><i class="c1" style="flex:0.206546" title="embed_text_tokens: 8.2M (20.7%)"></i><i class="c2" style="flex:0.105695" title="backbone_model.embed_tokens: 4.2M (10.6%)"></i><i class="c3" style="flex:0.001862" title="depth_decoder.model.layers: 74.0K (0.2%)"></i><i class="c4" style="flex:0.102392" title="depth_decoder.codebooks_head: 4.1M (10.2%)"></i><i class="c5" style="flex:0.001668" title="codec_model.decoder.decoder_transformer.layers: 66.3K (0.2%)"></i><i class="c6" style="flex:0.001668" title="codec_model.encoder.encoder_transformer.layers: 66.3K (0.2%)"></i><i class="c7" style="flex:0.578307" title="6 others: 23.0M (57.8%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>backbone_model.layers</td><td>973.1M</td><td>59.6%</td><td>74.0K</td><td>0.2%</td></tr><tr><td><i class="c1"></i>embed_text_tokens</td><td>262.7M</td><td>16.1%</td><td>8.2M</td><td>20.7%</td></tr><tr><td><i class="c2"></i>backbone_model.embed_tokens</td><td>134.4M</td><td>8.2%</td><td>4.2M</td><td>10.6%</td></tr><tr><td><i class="c3"></i>depth_decoder.model.layers</td><td>111.2M</td><td>6.8%</td><td>74.0K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>depth_decoder.codebooks_head</td><td>65.1M</td><td>4.0%</td><td>4.1M</td><td>10.2%</td></tr><tr><td><i class="c5"></i>codec_model.decoder.decoder_transformer.layers</td><td>25.2M</td><td>1.5%</td><td>66.3K</td><td>0.2%</td></tr><tr><td><i class="c6"></i>codec_model.encoder.encoder_transformer.layers</td><td>25.2M</td><td>1.5%</td><td>66.3K</td><td>0.2%</td></tr><tr><td><i class="c7"></i>6 others</td><td>35.2M</td><td>2.2%</td><td>23.0M</td><td>57.8%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>llama3</td><td>llama3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>csm</td><td>csm</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>2051</td><td>2051</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>RoPE type</td><td>llama3</td><td>llama3</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>model_type</td><td>csm_depth_decoder_model</td><td>csm_depth_decoder_model</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>vocab_size</td><td>2051</td><td>2051</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>depth_decoder_config</td><td>max_position_embeddings</td><td>33</td><td>33</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>attention style</td><td>MHA (8 heads, 8 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>model_type</td><td>mimi</td><td>mimi</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>max_position_embeddings</td><td>8000</td><td>8000</td><td>✓ kept</td></tr><tr class=""><td>codec_config</td><td>sampling_rate</td><td>24000</td><td>24000</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Csm omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>238</td><td>54</td></tr><tr class=""><td>RMSNorm</td><td>42</td><td>10</td></tr><tr class=""><td>LayerNorm</td><td>32</td><td>8</td></tr><tr class=""><td>MimiEuclideanCodebook</td><td>32</td><td>32</td></tr><tr class=""><td>MimiLayerScale</td><td>32</td><td>8</td></tr><tr class=""><td>MimiVectorQuantization</td><td>32</td><td>32</td></tr><tr class=""><td>Conv1d</td><td>29</td><td>29</td></tr><tr class=""><td>ELU</td><td>26</td><td>26</td></tr><tr class=""><td>MimiConv1d</td><td>25</td><td>25</td></tr><tr class=""><td>Attention</td><td>20</td><td>4</td></tr><tr class=""><td>DecoderLayer</td><td>20</td><td>4</td></tr><tr class=""><td>MLP</td><td>20</td><td>4</td></tr><tr class=""><td>SiLUActivation</td><td>20</td><td>4</td></tr><tr class=""><td>GELUActivation</td><td>16</td><td>4</td></tr><tr class=""><td>MimiAttention</td><td>16</td><td>4</td></tr><tr class=""><td>MimiMLP</td><td>16</td><td>4</td></tr><tr class=""><td>MimiTransformerLayer</td><td>16</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>16</td><td>16</td></tr><tr class=""><td>Identity</td><td>8</td><td>8</td></tr><tr class=""><td>MimiResnetBlock</td><td>8</td><td>8</td></tr><tr class=""><td>ConvTranspose1d</td><td>5</td><td>5</td></tr><tr class=""><td>MimiConvTranspose1d</td><td>5</td><td>5</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>RotaryEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>MimiResidualVectorQuantizer</td><td>2</td><td>2</td></tr><tr class=""><td>MimiRotaryEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>MimiTransformerModel</td><td>2</td><td>2</td></tr><tr class=""><td>BackboneModel</td><td>1</td><td>1</td></tr><tr class=""><td>BackboneModelEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>CodebooksHead</td><td>1</td><td>1</td></tr><tr class=""><td>DepthDecoderForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>DepthDecoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>MimiDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>MimiDecoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>MimiEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>MimiEncoderModel</td><td>1</td><td>1</td></tr><tr class=""><td>MimiModel</td><td>1</td><td>1</td></tr><tr class=""><td>MimiSplitResidualVectorQuantizer</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">19 fields changed, 96 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>codec_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>codec_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>codec_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>codec_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>codec_config.num_hidden_layers</td><td>8</td><td>2</td></tr><tr><td>codec_config.num_key_value_heads</td><td>8</td><td>4</td></tr><tr><td>codec_config.upsample_groups</td><td>512</td><td>64</td></tr><tr><td>depth_decoder_config.backbone_hidden_size</td><td>2048</td><td>64</td></tr><tr><td>depth_decoder_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>depth_decoder_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>depth_decoder_config.intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>depth_decoder_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>depth_decoder_config.num_hidden_layers</td><td>4</td><td>2</td></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/csm-1b-mic.json b/shrink/csm-1b-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..7d5c7de75615242ccea9edf624f7e62726437b80
--- /dev/null
+++ b/shrink/csm-1b-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/csm-1b-mic@ceb9f9aebdb8fe8d8aca3a3c51091ae2527893ca", "findings": []}
\ No newline at end of file
diff --git a/shrink/deberta-large-mnli-mic.html b/shrink/deberta-large-mnli-mic.html
index 470e83f0a51f7f70ed6ae6cfb245351ce12c2c60..5d35838044e69a4d890f9ffe47272631bebb60a6 100644
--- a/shrink/deberta-large-mnli-mic.html
+++ b/shrink/deberta-large-mnli-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>microsoft/deberta-large-mnli: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/deberta-large-mnli-mic@58e79989780438ce43238a30ec6c3453ea288907 ← microsoft/deberta-large-mnli@7296194b9009373def4f7c5dad292651e4b5cf4e · DebertaForSequenceClassification · overfit weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>microsoft/deberta-large-mnli</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/deberta-large-mnli-mic/tree/58e79989780438ce43238a30ec6c3453ea288907" title="huggingface-mic/deberta-large-mnli-mic@58e79989780438ce43238a30ec6c3453ea288907">huggingface-mic/deberta-large-mnli-mic@58e799897804</a> ← <a href="https://huggingface.co/microsoft/deberta-large-mnli/tree/7296194b9009373def4f7c5dad292651e4b5cf4e" title="microsoft/deberta-large-mnli@7296194b9009373def4f7c5dad292651e4b5cf4e">microsoft/deberta-large-mnli@7296194b9009</a> · DebertaForSequenceClassification · overfit weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">406.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>deberta.encoder.layer</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.868113" title="deberta.encoder.layer: 352.6M (86.8%)"></i><i class="c1" style="flex:0.126714" title="deberta.embeddings: 51.5M (12.7%)"></i><i class="c2" style="flex:0.002584" title="pooler.dense: 1.0M (0.3%)"></i><i class="c3" style="flex:0.002581" title="deberta.encoder: 1.0M (0.3%)"></i><i class="c4" style="flex:0.000008" title="classifier: 3.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.024724" title="deberta.encoder.layer: 83.3K (2.5%)"></i><i class="c1" style="flex:0.954539" title="deberta.embeddings: 3.2M (95.5%)"></i><i class="c2" style="flex:0.001234" title="pooler.dense: 4.2K (0.1%)"></i><i class="c3" style="flex:0.019445" title="deberta.encoder: 65.5K (1.9%)"></i><i class="c4" style="flex:0.000058" title="classifier: 195 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>deberta.encoder.layer</td><td>352.6M</td><td>86.8%</td><td>83.3K</td><td>2.5%</td></tr><tr><td><i class="c1"></i>deberta.embeddings</td><td>51.5M</td><td>12.7%</td><td>3.2M</td><td>95.5%</td></tr><tr><td><i class="c2"></i>pooler.dense</td><td>1.0M</td><td>0.3%</td><td>4.2K</td><td>0.1%</td></tr><tr><td><i class="c3"></i>deberta.encoder</td><td>1.0M</td><td>0.3%</td><td>65.5K</td><td>1.9%</td></tr><tr><td><i class="c4"></i>classifier</td><td>3.1K</td><td>0.0%</td><td>195</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>deberta</td><td>deberta</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50265</td><td>50265</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Deberta omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>14</td></tr><tr class=""><td>Dropout</td><td>99</td><td>11</td></tr><tr class=""><td>LayerNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>24</td><td>2</td></tr><tr class=""><td>Layer</td><td>24</td><td>2</td></tr><tr class=""><td>Output</td><td>24</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>24</td><td>2</td></tr><tr class=""><td>DisentangledSelfAttention</td><td>24</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ContextPooler</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForSequenceClassification</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 28 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>pooler_hidden_size</td><td>1024</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/deberta-large-mnli-mic.json b/shrink/deberta-large-mnli-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..9f88508d704e1ef7ded43f94a376f645c00a06f2
--- /dev/null
+++ b/shrink/deberta-large-mnli-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/deberta-large-mnli-mic@58e79989780438ce43238a30ec6c3453ea288907", "findings": []}
\ No newline at end of file
diff --git a/shrink/detr-mic.html b/shrink/detr-mic.html
index 2886151593bd808df0539005f80a16ebbbb6d01b..44f7dddcdcacfd9f4fdab4b8444b77ebfb7b77a6 100644
--- a/shrink/detr-mic.html
+++ b/shrink/detr-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/detr-resnet-50: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/detr-mic@47543c8af7646760a74482fedf00ec00fd8735e1 ← facebook/detr-resnet-50@1d5f47bd3bdd2c4bbfa585418ffe6da5028b4c0b · DetrForObjectDetection · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">41.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">23.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.564841" title="model.backbone: 23.5M (56.5%)"></i><i class="c1" style="flex:0.228117" title="model.decoder.layers: 9.5M (22.8%)"></i><i class="c2" style="flex:0.190017" title="model.encoder.layers: 7.9M (19.0%)"></i><i class="c3" style="flex:0.012632" title="model.input_projection: 524.5K (1.3%)"></i><i class="c4" style="flex:0.003194" title="bbox_predictor.layers: 132.6K (0.3%)"></i><i class="c5" style="flex:0.000616" title="model.query_position_embeddings: 25.6K (0.1%)"></i><i class="c6" style="flex:0.000569" title="class_labels_classifier: 23.6K (0.1%)"></i><i class="c7" style="flex:0.000012" title="1 others: 512 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.986555" title="model.backbone: 23.5M (98.7%)"></i><i class="c1" style="flex:0.004226" title="model.decoder.layers: 100.5K (0.4%)"></i><i class="c2" style="flex:0.002816" title="model.encoder.layers: 66.9K (0.3%)"></i><i class="c3" style="flex:0.005516" title="model.input_projection: 131.1K (0.6%)"></i><i class="c4" style="flex:0.000361" title="bbox_predictor.layers: 8.6K (0.0%)"></i><i class="c5" style="flex:0.000269" title="model.query_position_embeddings: 6.4K (0.0%)"></i><i class="c6" style="flex:0.000252" title="class_labels_classifier: 6.0K (0.0%)"></i><i class="c7" style="flex:0.000005" title="1 others: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.backbone</td><td>23.5M</td><td>56.5%</td><td>23.5M</td><td>98.7%</td></tr><tr><td><i class="c1"></i>model.decoder.layers</td><td>9.5M</td><td>22.8%</td><td>100.5K</td><td>0.4%</td></tr><tr><td><i class="c2"></i>model.encoder.layers</td><td>7.9M</td><td>19.0%</td><td>66.9K</td><td>0.3%</td></tr><tr><td><i class="c3"></i>model.input_projection</td><td>524.5K</td><td>1.3%</td><td>131.1K</td><td>0.6%</td></tr><tr><td><i class="c4"></i>bbox_predictor.layers</td><td>132.6K</td><td>0.3%</td><td>8.6K</td><td>0.0%</td></tr><tr><td><i class="c5"></i>model.query_position_embeddings</td><td>25.6K</td><td>0.1%</td><td>6.4K</td><td>0.0%</td></tr><tr><td><i class="c6"></i>class_labels_classifier</td><td>23.6K</td><td>0.1%</td><td>6.0K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>1 others</td><td>512</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>detr</td><td>detr</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>1024</td><td>1024</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>model_type</td><td>timm_backbone</td><td>timm_backbone</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Detr omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>100</td><td>36</td></tr><tr class=""><td>ReLU</td><td>61</td><td>53</td></tr><tr class=""><td>Conv2d</td><td>54</td><td>54</td></tr><tr class=""><td>FrozenBatchNorm2d</td><td>53</td><td>53</td></tr><tr class=""><td>Identity</td><td>32</td><td>32</td></tr><tr class=""><td>LayerNorm</td><td>31</td><td>11</td></tr><tr class=""><td>Bottleneck</td><td>16</td><td>16</td></tr><tr class=""><td>MLP</td><td>12</td><td>4</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>4</td></tr><tr class=""><td>Sequential</td><td>8</td><td>8</td></tr><tr class=""><td>CrossAttention</td><td>6</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>ConvEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForObjectDetection</td><td>1</td><td>1</td></tr><tr class=""><td>MLPPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SinePositionEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>FeatureListNet</td><td>1</td><td>1</td></tr><tr class=""><td>MaxPool2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 48 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>backbone_config.out_features</td><td>null</td><td>[&quot;layer4&quot;]</td></tr><tr><td>backbone_config.out_indices</td><td>[1, 2, 3, 4]</td><td>[-1]</td></tr><tr><td>backbone_config.stage_names</td><td>—</td><td>[&quot;act1&quot;, &quot;layer1&quot;, &quot;layer2&quot;, &quot;layer3&quot;, &quot;layer4&quot;]</td></tr><tr><td>d_model</td><td>256</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>2048</td><td>128</td></tr><tr><td>decoder_layers</td><td>6</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>2048</td><td>128</td></tr><tr><td>encoder_layers</td><td>6</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/detr-resnet-50</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/detr-mic/tree/47543c8af7646760a74482fedf00ec00fd8735e1" title="huggingface-mic/detr-mic@47543c8af7646760a74482fedf00ec00fd8735e1">huggingface-mic/detr-mic@47543c8af764</a> ← <a href="https://huggingface.co/facebook/detr-resnet-50/tree/1d5f47bd3bdd2c4bbfa585418ffe6da5028b4c0b" title="facebook/detr-resnet-50@1d5f47bd3bdd2c4bbfa585418ffe6da5028b4c0b">facebook/detr-resnet-50@1d5f47bd3bdd</a> · DetrForObjectDetection · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">41.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">23.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>the fixture keeps 57% of the original&#x27;s 41.5M parameters</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.564841" title="model.backbone: 23.5M (56.5%)"></i><i class="c1" style="flex:0.228117" title="model.decoder.layers: 9.5M (22.8%)"></i><i class="c2" style="flex:0.190017" title="model.encoder.layers: 7.9M (19.0%)"></i><i class="c3" style="flex:0.012632" title="model.input_projection: 524.5K (1.3%)"></i><i class="c4" style="flex:0.003194" title="bbox_predictor.layers: 132.6K (0.3%)"></i><i class="c5" style="flex:0.000616" title="model.query_position_embeddings: 25.6K (0.1%)"></i><i class="c6" style="flex:0.000569" title="class_labels_classifier: 23.6K (0.1%)"></i><i class="c7" style="flex:0.000012" title="1 others: 512 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.986555" title="model.backbone: 23.5M (98.7%)"></i><i class="c1" style="flex:0.004226" title="model.decoder.layers: 100.5K (0.4%)"></i><i class="c2" style="flex:0.002816" title="model.encoder.layers: 66.9K (0.3%)"></i><i class="c3" style="flex:0.005516" title="model.input_projection: 131.1K (0.6%)"></i><i class="c4" style="flex:0.000361" title="bbox_predictor.layers: 8.6K (0.0%)"></i><i class="c5" style="flex:0.000269" title="model.query_position_embeddings: 6.4K (0.0%)"></i><i class="c6" style="flex:0.000252" title="class_labels_classifier: 6.0K (0.0%)"></i><i class="c7" style="flex:0.000005" title="1 others: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.backbone</td><td>23.5M</td><td>56.5%</td><td>23.5M</td><td>98.7%</td></tr><tr><td><i class="c1"></i>model.decoder.layers</td><td>9.5M</td><td>22.8%</td><td>100.5K</td><td>0.4%</td></tr><tr><td><i class="c2"></i>model.encoder.layers</td><td>7.9M</td><td>19.0%</td><td>66.9K</td><td>0.3%</td></tr><tr><td><i class="c3"></i>model.input_projection</td><td>524.5K</td><td>1.3%</td><td>131.1K</td><td>0.6%</td></tr><tr><td><i class="c4"></i>bbox_predictor.layers</td><td>132.6K</td><td>0.3%</td><td>8.6K</td><td>0.0%</td></tr><tr><td><i class="c5"></i>model.query_position_embeddings</td><td>25.6K</td><td>0.1%</td><td>6.4K</td><td>0.0%</td></tr><tr><td><i class="c6"></i>class_labels_classifier</td><td>23.6K</td><td>0.1%</td><td>6.0K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>1 others</td><td>512</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>detr</td><td>detr</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>1024</td><td>1024</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>model_type</td><td>timm_backbone</td><td>timm_backbone</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Detr omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>100</td><td>36</td></tr><tr class=""><td>ReLU</td><td>61</td><td>53</td></tr><tr class=""><td>Conv2d</td><td>54</td><td>54</td></tr><tr class=""><td>FrozenBatchNorm2d</td><td>53</td><td>53</td></tr><tr class=""><td>Identity</td><td>32</td><td>32</td></tr><tr class=""><td>LayerNorm</td><td>31</td><td>11</td></tr><tr class=""><td>Bottleneck</td><td>16</td><td>16</td></tr><tr class=""><td>MLP</td><td>12</td><td>4</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>4</td></tr><tr class=""><td>Sequential</td><td>8</td><td>8</td></tr><tr class=""><td>CrossAttention</td><td>6</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>ConvEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForObjectDetection</td><td>1</td><td>1</td></tr><tr class=""><td>MLPPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SinePositionEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>FeatureListNet</td><td>1</td><td>1</td></tr><tr class=""><td>MaxPool2d</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 48 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>backbone_config.out_features</td><td>null</td><td>[&quot;layer4&quot;]</td></tr><tr><td>backbone_config.out_indices</td><td>[1, 2, 3, 4]</td><td>[-1]</td></tr><tr><td>backbone_config.stage_names</td><td>—</td><td>[&quot;act1&quot;, &quot;layer1&quot;, &quot;layer2&quot;, &quot;layer3&quot;, &quot;layer4&quot;]</td></tr><tr><td>d_model</td><td>256</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>2048</td><td>128</td></tr><tr><td>decoder_layers</td><td>6</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>2048</td><td>128</td></tr><tr><td>encoder_layers</td><td>6</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/detr-mic.json b/shrink/detr-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..b0d52a4c4e2a0086d9dc30a2c05d2f17cb7be72a
--- /dev/null
+++ b/shrink/detr-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/detr-mic@47543c8af7646760a74482fedf00ec00fd8735e1", "findings": ["the fixture keeps 57% of the original's 41.5M parameters"]}
\ No newline at end of file
diff --git a/shrink/dinov2-small-mic.html b/shrink/dinov2-small-mic.html
index 0a197f34b46ad830643925490d21fa4376bb35dd..6e9f322554729c44e99f455872d9887d98c0d700 100644
--- a/shrink/dinov2-small-mic.html
+++ b/shrink/dinov2-small-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/dinov2-small: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/dinov2-small-mic@770cc8e9a4eccc42eb2f4ea67371b9997d0395d5 ← facebook/dinov2-small@ed25f3a31f01632728cabb09d1542f84ab7b0056 · Dinov2Model · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/dinov2-small</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/dinov2-small-mic/tree/770cc8e9a4eccc42eb2f4ea67371b9997d0395d5" title="huggingface-mic/dinov2-small-mic@770cc8e9a4eccc42eb2f4ea67371b9997d0395d5">huggingface-mic/dinov2-small-mic@770cc8e9a4ec</a> ← <a href="https://huggingface.co/facebook/dinov2-small/tree/ed25f3a31f01632728cabb09d1542f84ab7b0056" title="facebook/dinov2-small@ed25f3a31f01632728cabb09d1542f84ab7b0056">facebook/dinov2-small@ed25f3a31f01</a> · Dinov2Model · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">22.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">225.9K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.965825" title="encoder.layer: 21.3M (96.6%)"></i><i class="c1" style="flex:0.023886" title="embeddings: 526.8K (2.4%)"></i><i class="c2" style="flex:0.010254" title="embeddings.patch_embeddings: 226.2K (1.0%)"></i><i class="c3" style="flex:0.000035" title="layernorm: 768 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.443752" title="encoder.layer: 100.2K (44.4%)"></i><i class="c1" style="flex:0.388779" title="embeddings: 87.8K (38.9%)"></i><i class="c2" style="flex:0.166903" title="embeddings.patch_embeddings: 37.7K (16.7%)"></i><i class="c3" style="flex:0.000567" title="layernorm: 128 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>encoder.layer</td><td>21.3M</td><td>96.6%</td><td>100.2K</td><td>44.4%</td></tr><tr><td><i class="c1"></i>embeddings</td><td>526.8K</td><td>2.4%</td><td>87.8K</td><td>38.9%</td></tr><tr><td><i class="c2"></i>embeddings.patch_embeddings</td><td>226.2K</td><td>1.0%</td><td>37.7K</td><td>16.7%</td></tr><tr><td><i class="c3"></i>layernorm</td><td>768</td><td>0.0%</td><td>128</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>dinov2</td><td>dinov2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>518</td><td>518</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Dinov2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>72</td><td>12</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>LayerScale</td><td>24</td><td>4</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>MLP</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Identity</td><td>12</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>PatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Dropout</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>384</td><td>64</td></tr><tr><td>num_attention_heads</td><td>6</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>out_features</td><td>[&quot;stage12&quot;]</td><td>[&quot;stage2&quot;]</td></tr><tr><td>out_indices</td><td>[12]</td><td>[2]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/dinov2-small-mic.json b/shrink/dinov2-small-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..5b1267687f95cab40ed2efd7c063481af0119663
--- /dev/null
+++ b/shrink/dinov2-small-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/dinov2-small-mic@770cc8e9a4eccc42eb2f4ea67371b9997d0395d5", "findings": []}
\ No newline at end of file
diff --git a/shrink/dinov2-with-registers-base-mic.html b/shrink/dinov2-with-registers-base-mic.html
index 19736b64d13838fb5b2121b1a461664a2a168c45..f45d58465aade6f1a5171839ed80bd809ef3b681 100644
--- a/shrink/dinov2-with-registers-base-mic.html
+++ b/shrink/dinov2-with-registers-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/dinov2-with-registers-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/dinov2-with-registers-base-mic@186e99a19d6165f6a42c2794175a5d87d8781984 ← facebook/dinov2-with-registers-base@a1d738ccfa7ae170945f210395d99dde8adb1805 · Dinov2WithRegistersModel · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/dinov2-with-registers-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/dinov2-with-registers-base-mic/tree/186e99a19d6165f6a42c2794175a5d87d8781984" title="huggingface-mic/dinov2-with-registers-base-mic@186e99a19d6165f6a42c2794175a5d87d8781984">huggingface-mic/dinov2-with-registers-base-mic@186e99a19d61</a> ← <a href="https://huggingface.co/facebook/dinov2-with-registers-base/tree/a1d738ccfa7ae170945f210395d99dde8adb1805" title="facebook/dinov2-with-registers-base@a1d738ccfa7ae170945f210395d99dde8adb1805">facebook/dinov2-with-registers-base@a1d738ccfa7a</a> · Dinov2WithRegistersModel · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">86.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">226.1K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.982553" title="encoder.layer: 85.1M (98.3%)"></i><i class="c1" style="flex:0.012205" title="embeddings: 1.1M (1.2%)"></i><i class="c2" style="flex:0.005224" title="embeddings.patch_embeddings: 452.4K (0.5%)"></i><i class="c3" style="flex:0.000018" title="layernorm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.443249" title="encoder.layer: 100.2K (44.3%)"></i><i class="c1" style="flex:0.389471" title="embeddings: 88.1K (38.9%)"></i><i class="c2" style="flex:0.166714" title="embeddings.patch_embeddings: 37.7K (16.7%)"></i><i class="c3" style="flex:0.000566" title="layernorm: 128 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>encoder.layer</td><td>85.1M</td><td>98.3%</td><td>100.2K</td><td>44.3%</td></tr><tr><td><i class="c1"></i>embeddings</td><td>1.1M</td><td>1.2%</td><td>88.1K</td><td>38.9%</td></tr><tr><td><i class="c2"></i>embeddings.patch_embeddings</td><td>452.4K</td><td>0.5%</td><td>37.7K</td><td>16.7%</td></tr><tr><td><i class="c3"></i>layernorm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>dinov2_with_registers</td><td>dinov2_with_registers</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>518</td><td>518</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Dinov2WithRegisters omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>72</td><td>12</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>LayerScale</td><td>24</td><td>4</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>MLP</td><td>12</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Identity</td><td>12</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>PatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Dropout</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 28 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>out_features</td><td>[&quot;stage12&quot;]</td><td>[&quot;stage2&quot;]</td></tr><tr><td>out_indices</td><td>[12]</td><td>[2]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/dinov2-with-registers-base-mic.json b/shrink/dinov2-with-registers-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e432faa6e6deba8af0a32e6a020768490ec7b286
--- /dev/null
+++ b/shrink/dinov2-with-registers-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/dinov2-with-registers-base-mic@186e99a19d6165f6a42c2794175a5d87d8781984", "findings": []}
\ No newline at end of file
diff --git a/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.html b/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.html
index 19e75966d1501e6818d3914e10df586c4ddf0398..68ca702733c4ec76cad36e4893d54da94b51c9c8 100644
--- a/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.html
+++ b/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>camenduru/dinov3-vitl16-pretrain-lvd1689m: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/dinov3-vitl16-pretrain-lvd1689m-mic@ba70f2bd5556b75212ed9bdd17efebacf33b37ae ← camenduru/dinov3-vitl16-pretrain-lvd1689m@3c276edd87d6f6e569ff0c4400e086807d0f3881 · DINOv3ViTModel · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>camenduru/dinov3-vitl16-pretrain-lvd1689m</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/dinov3-vitl16-pretrain-lvd1689m-mic/tree/ba70f2bd5556b75212ed9bdd17efebacf33b37ae" title="huggingface-mic/dinov3-vitl16-pretrain-lvd1689m-mic@ba70f2bd5556b75212ed9bdd17efebacf33b37ae">huggingface-mic/dinov3-vitl16-pretrain-lvd1689m-mic@ba70f2bd5556</a> ← <a href="https://huggingface.co/camenduru/dinov3-vitl16-pretrain-lvd1689m/tree/3c276edd87d6f6e569ff0c4400e086807d0f3881" title="camenduru/dinov3-vitl16-pretrain-lvd1689m@3c276edd87d6f6e569ff0c4400e086807d0f3881">camenduru/dinov3-vitl16-pretrain-lvd1689m@3c276edd87d6</a> · DINOv3ViTModel · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">303.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">116.8K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layer</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.997375" title="model.layer: 302.3M (99.7%)"></i><i class="c1" style="flex:0.002598" title="embeddings.patch_embeddings: 787.5K (0.3%)"></i><i class="c2" style="flex:0.000020" title="embeddings: 6.1K (0.0%)"></i><i class="c3" style="flex:0.000007" title="norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.574247" title="model.layer: 67.1K (57.4%)"></i><i class="c1" style="flex:0.421370" title="embeddings.patch_embeddings: 49.2K (42.1%)"></i><i class="c2" style="flex:0.003288" title="embeddings: 384 (0.3%)"></i><i class="c3" style="flex:0.001096" title="norm: 128 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layer</td><td>302.3M</td><td>99.7%</td><td>67.1K</td><td>57.4%</td></tr><tr><td><i class="c1"></i>embeddings.patch_embeddings</td><td>787.5K</td><td>0.3%</td><td>49.2K</td><td>42.1%</td></tr><tr><td><i class="c2"></i>embeddings</td><td>6.1K</td><td>0.0%</td><td>384</td><td>0.3%</td></tr><tr><td><i class="c3"></i>norm</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>dinov3_vit</td><td>dinov3_vit</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>rope_theta</td><td>100.0</td><td>100.0</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DINOv3ViT omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>144</td><td>12</td></tr><tr class=""><td>LayerNorm</td><td>49</td><td>5</td></tr><tr class=""><td>LayerScale</td><td>48</td><td>4</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Layer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Identity</td><td>24</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RopePositionEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 32 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>out_features</td><td>[&quot;stage24&quot;]</td><td>[&quot;stage2&quot;]</td></tr><tr><td>out_indices</td><td>[24]</td><td>[2]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.json b/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..f54fec38145686010a72c4fd9e7b5433573929b9
--- /dev/null
+++ b/shrink/dinov3-vitl16-pretrain-lvd1689m-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/dinov3-vitl16-pretrain-lvd1689m-mic@ba70f2bd5556b75212ed9bdd17efebacf33b37ae", "findings": []}
\ No newline at end of file
diff --git a/shrink/distilbert-base-uncased-mic.html b/shrink/distilbert-base-uncased-mic.html
index c28b7989061d081ff6f006362deeaf41024c20ff..683269a70f2adc040ce450b12e49524dde7ae592 100644
--- a/shrink/distilbert-base-uncased-mic.html
+++ b/shrink/distilbert-base-uncased-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>distilbert/distilbert-base-uncased: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/distilbert-base-uncased-mic@c65e2b7864ac5e15a46a61bdb7c1c89da10ab09c ← distilbert/distilbert-base-uncased@12040accade4e8a0f71eabdb258fecc2e7e948be · DistilBertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>distilbert/distilbert-base-uncased</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/distilbert-base-uncased-mic/tree/c65e2b7864ac5e15a46a61bdb7c1c89da10ab09c" title="huggingface-mic/distilbert-base-uncased-mic@c65e2b7864ac5e15a46a61bdb7c1c89da10ab09c">huggingface-mic/distilbert-base-uncased-mic@c65e2b7864ac</a> ← <a href="https://huggingface.co/distilbert/distilbert-base-uncased/tree/12040accade4e8a0f71eabdb258fecc2e7e948be" title="distilbert/distilbert-base-uncased@12040accade4e8a0f71eabdb258fecc2e7e948be">distilbert/distilbert-base-uncased@12040accade4</a> · DistilBertForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">67.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>distilbert.transformer.layer</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: TransformerBlock"></i><i class="c0" title="layer 1: TransformerBlock"></i><i class="c0" title="layer 2: TransformerBlock"></i><i class="c0" title="layer 3: TransformerBlock"></i><i class="c0" title="layer 4: TransformerBlock"></i><i class="c0" title="layer 5: TransformerBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TransformerBlock"></i><i class="c0" title="layer 1: TransformerBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TransformerBlock</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.634872" title="distilbert.transformer.layer: 42.5M (63.5%)"></i><i class="c1" style="flex:0.355833" title="distilbert.embeddings: 23.8M (35.6%)"></i><i class="c2" style="flex:0.008817" title="vocab_transform: 590.6K (0.9%)"></i><i class="c3" style="flex:0.000456" title="vocab_projector: 30.5K (0.0%)"></i><i class="c4" style="flex:0.000023" title="vocab_layer_norm: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.032060" title="distilbert.transformer.layer: 66.9K (3.2%)"></i><i class="c1" style="flex:0.951269" title="distilbert.embeddings: 2.0M (95.1%)"></i><i class="c2" style="flex:0.001992" title="vocab_transform: 4.2K (0.2%)"></i><i class="c3" style="flex:0.014617" title="vocab_projector: 30.5K (1.5%)"></i><i class="c4" style="flex:0.000061" title="vocab_layer_norm: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>distilbert.transformer.layer</td><td>42.5M</td><td>63.5%</td><td>66.9K</td><td>3.2%</td></tr><tr><td><i class="c1"></i>distilbert.embeddings</td><td>23.8M</td><td>35.6%</td><td>2.0M</td><td>95.1%</td></tr><tr><td><i class="c2"></i>vocab_transform</td><td>590.6K</td><td>0.9%</td><td>4.2K</td><td>0.2%</td></tr><tr><td><i class="c3"></i>vocab_projector</td><td>30.5K</td><td>0.0%</td><td>30.5K</td><td>1.5%</td></tr><tr><td><i class="c4"></i>vocab_layer_norm</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>distilbert</td><td>distilbert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DistilBert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>38</td><td>14</td></tr><tr class=""><td>LayerNorm</td><td>14</td><td>6</td></tr><tr class=""><td>Dropout</td><td>13</td><td>5</td></tr><tr class=""><td>GELUActivation</td><td>7</td><td>3</td></tr><tr class=""><td>SelfAttention</td><td>6</td><td>2</td></tr><tr class=""><td>FFN</td><td>6</td><td>2</td></tr><tr class=""><td>TransformerBlock</td><td>6</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>CrossEntropyLoss</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>Transformer</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 23 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>dim</td><td>768</td><td>64</td></tr><tr><td>hidden_dim</td><td>3072</td><td>128</td></tr><tr><td>n_heads</td><td>12</td><td>4</td></tr><tr><td>n_layers</td><td>6</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/distilbert-base-uncased-mic.json b/shrink/distilbert-base-uncased-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..14b2bee2c4210bef72fb1d8d60b51e81ed76d19f
--- /dev/null
+++ b/shrink/distilbert-base-uncased-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/distilbert-base-uncased-mic@c65e2b7864ac5e15a46a61bdb7c1c89da10ab09c", "findings": []}
\ No newline at end of file
diff --git a/shrink/docling-layout-heron-mic.html b/shrink/docling-layout-heron-mic.html
index 9edd384edb7bf99920a032106bde8482c37ee562..465fa3be0cdfe491c3b379823a961be44ca4c1b7 100644
--- a/shrink/docling-layout-heron-mic.html
+++ b/shrink/docling-layout-heron-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>docling-project/docling-layout-heron: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/docling-layout-heron-mic@a4c39589d12b25a98e2c975388b2629748602771 ← docling-project/docling-layout-heron@8f39ad3c0b4c58e9c2d2c84a38465abf757272d8 · RTDetrV2ForObjectDetection · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">42.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">10.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">11</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.backbone.model.encoder.stages</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: no AvgPool2d"></i><i class="c1" title="layer 1: AvgPool2d"></i><i class="c1" title="layer 2: AvgPool2d"></i><i class="c1" title="layer 3: AvgPool2d"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: no AvgPool2d"></i><i class="c1" title="layer 1: AvgPool2d"></i><i class="c1" title="layer 2: AvgPool2d"></i><i class="c1" title="layer 3: AvgPool2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no AvgPool2d</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>AvgPool2d</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder_input_proj</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv2d</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.aifi</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: AIFILayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: AIFILayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AIFILayer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.lateral_convs</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ConvNormLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.fpn_blocks</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: RepVggBlock"></i><i class="c0" title="layer 1: RepVggBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: RepVggBlock"></i><i class="c0" title="layer 1: RepVggBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>RepVggBlock</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.downsample_convs</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ConvNormLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.pan_blocks</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: RepVggBlock"></i><i class="c0" title="layer 1: RepVggBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: RepVggBlock"></i><i class="c0" title="layer 1: RepVggBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>RepVggBlock</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder_input_proj</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv2d</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.bbox_embed</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>bbox_embed</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.548280" title="model.backbone.model.encoder.stages: 23.4M (54.8%)"></i><i class="c1" style="flex:0.139733" title="model.decoder.layers: 6.0M (14.0%)"></i><i class="c2" style="flex:0.104407" title="model.encoder.fpn_blocks: 4.5M (10.4%)"></i><i class="c3" style="flex:0.104407" title="model.encoder.pan_blocks: 4.5M (10.4%)"></i><i class="c4" style="flex:0.027610" title="model.encoder.downsample_convs: 1.2M (2.8%)"></i><i class="c5" style="flex:0.021492" title="model.encoder_input_proj: 919.0K (2.1%)"></i><i class="c6" style="flex:0.018607" title="model.decoder.bbox_embed: 795.7K (1.9%)"></i><i class="c7" style="flex:0.035464" title="9 others: 1.5M (3.5%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.002976" title="model.backbone.model.encoder.stages: 32.4K (0.3%)"></i><i class="c1" style="flex:0.009420" title="model.decoder.layers: 102.6K (0.9%)"></i><i class="c2" style="flex:0.410073" title="model.encoder.fpn_blocks: 4.5M (41.0%)"></i><i class="c3" style="flex:0.410073" title="model.encoder.pan_blocks: 4.5M (41.0%)"></i><i class="c4" style="flex:0.108444" title="model.encoder.downsample_convs: 1.2M (10.8%)"></i><i class="c5" style="flex:0.005408" title="model.encoder_input_proj: 58.9K (0.5%)"></i><i class="c6" style="flex:0.001576" title="model.decoder.bbox_embed: 17.2K (0.2%)"></i><i class="c7" style="flex:0.052031" title="9 others: 566.5K (5.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.backbone.model.encoder.stages</td><td>23.4M</td><td>54.8%</td><td>32.4K</td><td>0.3%</td></tr><tr><td><i class="c1"></i>model.decoder.layers</td><td>6.0M</td><td>14.0%</td><td>102.6K</td><td>0.9%</td></tr><tr><td><i class="c2"></i>model.encoder.fpn_blocks</td><td>4.5M</td><td>10.4%</td><td>4.5M</td><td>41.0%</td></tr><tr><td><i class="c3"></i>model.encoder.pan_blocks</td><td>4.5M</td><td>10.4%</td><td>4.5M</td><td>41.0%</td></tr><tr><td><i class="c4"></i>model.encoder.downsample_convs</td><td>1.2M</td><td>2.8%</td><td>1.2M</td><td>10.8%</td></tr><tr><td><i class="c5"></i>model.encoder_input_proj</td><td>919.0K</td><td>2.1%</td><td>58.9K</td><td>0.5%</td></tr><tr><td><i class="c6"></i>model.decoder.bbox_embed</td><td>795.7K</td><td>1.9%</td><td>17.2K</td><td>0.2%</td></tr><tr><td><i class="c7"></i>9 others</td><td>1.5M</td><td>3.5%</td><td>566.5K</td><td>5.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>rt_detr_v2</td><td>rt_detr_v2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>model_type</td><td>rt_detr_resnet</td><td>rt_detr_resnet</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>hidden_act</td><td>relu</td><td>relu</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix RTDetrV2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>121</td><td>49</td></tr><tr class=""><td>Conv2d</td><td>97</td><td>61</td></tr><tr class=""><td>ReLU</td><td>57</td><td>17</td></tr><tr class=""><td>Identity</td><td>56</td><td>32</td></tr><tr class=""><td>FrozenBatchNorm2d</td><td>55</td><td>19</td></tr><tr class=""><td>RTDetrResNetConvLayer</td><td>51</td><td>15</td></tr><tr class=""><td>BatchNorm2d</td><td>42</td><td>42</td></tr><tr class=""><td>ConvNormLayer</td><td>36</td><td>36</td></tr><tr class=""><td>Sequential</td><td>35</td><td>23</td></tr><tr class=""><td>ModuleList</td><td>28</td><td>20</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>24</td></tr><tr class=""><td>LayerNorm</td><td>21</td><td>9</td></tr><tr class=""><td>RTDetrResNetBottleNeckLayer</td><td>16</td><td>4</td></tr><tr class=""><td>MLPPredictionHead</td><td>14</td><td>6</td></tr><tr class=""><td>RepVggBlock</td><td>12</td><td>12</td></tr><tr class=""><td>MLP</td><td>7</td><td>3</td></tr><tr class=""><td>SelfAttention</td><td>7</td><td>3</td></tr><tr class=""><td>DecoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>MultiscaleDeformableAttention</td><td>6</td><td>2</td></tr><tr class=""><td>RTDetrResNetShortCut</td><td>4</td><td>4</td></tr><tr class=""><td>RTDetrResNetStage</td><td>4</td><td>4</td></tr><tr class=""><td>CSPRepLayer</td><td>4</td><td>4</td></tr><tr class=""><td>AvgPool2d</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>MaxPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>RTDetrResNetBackbone</td><td>1</td><td>1</td></tr><tr class=""><td>RTDetrResNetEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>RTDetrResNetEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>AIFILayer</td><td>1</td><td>1</td></tr><tr class=""><td>ConvEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>EncoderLayer</td><td>1</td><td>1</td></tr><tr class=""><td>ForObjectDetection</td><td>1</td><td>1</td></tr><tr class=""><td>HybridEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SinePositionEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 74 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>backbone_config.depths</td><td>[3, 4, 6, 3]</td><td>[1, 1, 1, 1]</td></tr><tr><td>backbone_config.hidden_sizes</td><td>[256, 512, 1024, 2048]</td><td>[16, 32, 64, 128]</td></tr><tr><td>d_model</td><td>256</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>1024</td><td>128</td></tr><tr><td>decoder_layers</td><td>6</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>1024</td><td>128</td></tr><tr><td>encoder_in_channels</td><td>[512, 1024, 2048]</td><td>[32, 64, 128]</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>docling-project/docling-layout-heron</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/docling-layout-heron-mic/tree/a4c39589d12b25a98e2c975388b2629748602771" title="huggingface-mic/docling-layout-heron-mic@a4c39589d12b25a98e2c975388b2629748602771">huggingface-mic/docling-layout-heron-mic@a4c39589d12b</a> ← <a href="https://huggingface.co/docling-project/docling-layout-heron/tree/8f39ad3c0b4c58e9c2d2c84a38465abf757272d8" title="docling-project/docling-layout-heron@8f39ad3c0b4c58e9c2d2c84a38465abf757272d8">docling-project/docling-layout-heron@8f39ad3c0b4c</a> · RTDetrV2ForObjectDetection · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">42.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">10.9M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">11</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.backbone.model.encoder.stages</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: no AvgPool2d"></i><i class="c1" title="layer 1: AvgPool2d"></i><i class="c1" title="layer 2: AvgPool2d"></i><i class="c1" title="layer 3: AvgPool2d"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: no AvgPool2d"></i><i class="c1" title="layer 1: AvgPool2d"></i><i class="c1" title="layer 2: AvgPool2d"></i><i class="c1" title="layer 3: AvgPool2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>no AvgPool2d</td><td>1</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>AvgPool2d</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder_input_proj</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv2d</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.aifi</h3><div class="row"><span class="lab">Original · 1</span><div class="cells"><i class="c0" title="layer 0: AIFILayer"></i></div></div><div class="row"><span class="lab">Fixture · 1</span><div class="cells"><i class="c0" title="layer 0: AIFILayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AIFILayer</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.lateral_convs</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ConvNormLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.fpn_blocks</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: CSPRepLayer"></i><i class="c0" title="layer 1: CSPRepLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: CSPRepLayer"></i><i class="c0" title="layer 1: CSPRepLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>CSPRepLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.downsample_convs</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: ConvNormLayer"></i><i class="c0" title="layer 1: ConvNormLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ConvNormLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.pan_blocks</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: CSPRepLayer"></i><i class="c0" title="layer 1: CSPRepLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: CSPRepLayer"></i><i class="c0" title="layer 1: CSPRepLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>CSPRepLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder_input_proj</h3><div class="row"><span class="lab">Original · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><div class="row"><span class="lab">Fixture · 3</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv2d</td><td>3</td><td>3</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.bbox_embed</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>bbox_embed</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.548280" title="model.backbone.model.encoder.stages: 23.4M (54.8%)"></i><i class="c1" style="flex:0.139733" title="model.decoder.layers: 6.0M (14.0%)"></i><i class="c2" style="flex:0.104407" title="model.encoder.pan_blocks: 4.5M (10.4%)"></i><i class="c3" style="flex:0.104407" title="model.encoder.fpn_blocks: 4.5M (10.4%)"></i><i class="c4" style="flex:0.027610" title="model.encoder.downsample_convs: 1.2M (2.8%)"></i><i class="c5" style="flex:0.021492" title="model.encoder_input_proj: 919.0K (2.1%)"></i><i class="c6" style="flex:0.018607" title="model.decoder.bbox_embed: 795.7K (1.9%)"></i><i class="c7" style="flex:0.035464" title="9 others: 1.5M (3.5%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.002976" title="model.backbone.model.encoder.stages: 32.4K (0.3%)"></i><i class="c1" style="flex:0.009420" title="model.decoder.layers: 102.6K (0.9%)"></i><i class="c2" style="flex:0.410073" title="model.encoder.pan_blocks: 4.5M (41.0%)"></i><i class="c3" style="flex:0.410073" title="model.encoder.fpn_blocks: 4.5M (41.0%)"></i><i class="c4" style="flex:0.108444" title="model.encoder.downsample_convs: 1.2M (10.8%)"></i><i class="c5" style="flex:0.005408" title="model.encoder_input_proj: 58.9K (0.5%)"></i><i class="c6" style="flex:0.001576" title="model.decoder.bbox_embed: 17.2K (0.2%)"></i><i class="c7" style="flex:0.052031" title="9 others: 566.5K (5.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.backbone.model.encoder.stages</td><td>23.4M</td><td>54.8%</td><td>32.4K</td><td>0.3%</td></tr><tr><td><i class="c1"></i>model.decoder.layers</td><td>6.0M</td><td>14.0%</td><td>102.6K</td><td>0.9%</td></tr><tr><td><i class="c2"></i>model.encoder.pan_blocks</td><td>4.5M</td><td>10.4%</td><td>4.5M</td><td>41.0%</td></tr><tr><td><i class="c3"></i>model.encoder.fpn_blocks</td><td>4.5M</td><td>10.4%</td><td>4.5M</td><td>41.0%</td></tr><tr><td><i class="c4"></i>model.encoder.downsample_convs</td><td>1.2M</td><td>2.8%</td><td>1.2M</td><td>10.8%</td></tr><tr><td><i class="c5"></i>model.encoder_input_proj</td><td>919.0K</td><td>2.1%</td><td>58.9K</td><td>0.5%</td></tr><tr><td><i class="c6"></i>model.decoder.bbox_embed</td><td>795.7K</td><td>1.9%</td><td>17.2K</td><td>0.2%</td></tr><tr><td><i class="c7"></i>9 others</td><td>1.5M</td><td>3.5%</td><td>566.5K</td><td>5.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>rt_detr_v2</td><td>rt_detr_v2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>model_type</td><td>rt_detr_resnet</td><td>rt_detr_resnet</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>hidden_act</td><td>relu</td><td>relu</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix RTDetrV2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>121</td><td>49</td></tr><tr class=""><td>Conv2d</td><td>97</td><td>61</td></tr><tr class=""><td>ReLU</td><td>57</td><td>17</td></tr><tr class=""><td>Identity</td><td>56</td><td>32</td></tr><tr class=""><td>FrozenBatchNorm2d</td><td>55</td><td>19</td></tr><tr class=""><td>RTDetrResNetConvLayer</td><td>51</td><td>15</td></tr><tr class=""><td>BatchNorm2d</td><td>42</td><td>42</td></tr><tr class=""><td>ConvNormLayer</td><td>36</td><td>36</td></tr><tr class=""><td>Sequential</td><td>35</td><td>23</td></tr><tr class=""><td>ModuleList</td><td>28</td><td>20</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>24</td></tr><tr class=""><td>LayerNorm</td><td>21</td><td>9</td></tr><tr class=""><td>RTDetrResNetBottleNeckLayer</td><td>16</td><td>4</td></tr><tr class=""><td>MLPPredictionHead</td><td>14</td><td>6</td></tr><tr class=""><td>RepVggBlock</td><td>12</td><td>12</td></tr><tr class=""><td>MLP</td><td>7</td><td>3</td></tr><tr class=""><td>SelfAttention</td><td>7</td><td>3</td></tr><tr class=""><td>DecoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>MultiscaleDeformableAttention</td><td>6</td><td>2</td></tr><tr class=""><td>RTDetrResNetShortCut</td><td>4</td><td>4</td></tr><tr class=""><td>RTDetrResNetStage</td><td>4</td><td>4</td></tr><tr class=""><td>CSPRepLayer</td><td>4</td><td>4</td></tr><tr class=""><td>AvgPool2d</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>MaxPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>RTDetrResNetBackbone</td><td>1</td><td>1</td></tr><tr class=""><td>RTDetrResNetEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>RTDetrResNetEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>AIFILayer</td><td>1</td><td>1</td></tr><tr class=""><td>ConvEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>EncoderLayer</td><td>1</td><td>1</td></tr><tr class=""><td>ForObjectDetection</td><td>1</td><td>1</td></tr><tr class=""><td>HybridEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SinePositionEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 74 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>backbone_config.depths</td><td>[3, 4, 6, 3]</td><td>[1, 1, 1, 1]</td></tr><tr><td>backbone_config.hidden_sizes</td><td>[256, 512, 1024, 2048]</td><td>[16, 32, 64, 128]</td></tr><tr><td>d_model</td><td>256</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>1024</td><td>128</td></tr><tr><td>decoder_layers</td><td>6</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>1024</td><td>128</td></tr><tr><td>encoder_in_channels</td><td>[512, 1024, 2048]</td><td>[32, 64, 128]</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/docling-layout-heron-mic.json b/shrink/docling-layout-heron-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..5c7b5448d049617e69a29d764c3980d001585829
--- /dev/null
+++ b/shrink/docling-layout-heron-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/docling-layout-heron-mic@a4c39589d12b25a98e2c975388b2629748602771", "findings": []}
\ No newline at end of file
diff --git a/shrink/dpt-large-mic.html b/shrink/dpt-large-mic.html
index 29b3939f17a9357532f83137f01cdb30389f6bd3..3a680bf2c6b4e5b35196800a3d707dab7cadf505 100644
--- a/shrink/dpt-large-mic.html
+++ b/shrink/dpt-large-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>Intel/dpt-large: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/dpt-large-mic@fef05122ed4bc80c9abd66c8dfb177787005eb45 ← Intel/dpt-large@bc15f29aa3a80d532f2ed650b5e16ac48d8958f9 · DPTForDepthEstimation · overfit weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>Intel/dpt-large</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/dpt-large-mic/tree/fef05122ed4bc80c9abd66c8dfb177787005eb45" title="huggingface-mic/dpt-large-mic@fef05122ed4bc80c9abd66c8dfb177787005eb45">huggingface-mic/dpt-large-mic@fef05122ed4b</a> ← <a href="https://huggingface.co/Intel/dpt-large/tree/bc15f29aa3a80d532f2ed650b5e16ac48d8958f9" title="Intel/dpt-large@bc15f29aa3a80d532f2ed650b5e16ac48d8958f9">Intel/dpt-large@bc15f29aa3a8</a> · DPTForDepthEstimation · overfit weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">343.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">1.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">4</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>dpt.encoder.layer</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: ViTLayer"></i><i class="c0" title="layer 1: ViTLayer"></i><i class="c0" title="layer 2: ViTLayer"></i><i class="c0" title="layer 3: ViTLayer"></i><i class="c0" title="layer 4: ViTLayer"></i><i class="c0" title="layer 5: ViTLayer"></i><i class="c0" title="layer 6: ViTLayer"></i><i class="c0" title="layer 7: ViTLayer"></i><i class="c0" title="layer 8: ViTLayer"></i><i class="c0" title="layer 9: ViTLayer"></i><i class="c0" title="layer 10: ViTLayer"></i><i class="c0" title="layer 11: ViTLayer"></i><i class="c0" title="layer 12: ViTLayer"></i><i class="c0" title="layer 13: ViTLayer"></i><i class="c0" title="layer 14: ViTLayer"></i><i class="c0" title="layer 15: ViTLayer"></i><i class="c0" title="layer 16: ViTLayer"></i><i class="c0" title="layer 17: ViTLayer"></i><i class="c0" title="layer 18: ViTLayer"></i><i class="c0" title="layer 19: ViTLayer"></i><i class="c0" title="layer 20: ViTLayer"></i><i class="c0" title="layer 21: ViTLayer"></i><i class="c0" title="layer 22: ViTLayer"></i><i class="c0" title="layer 23: ViTLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: ViTLayer"></i><i class="c0" title="layer 1: ViTLayer"></i><i class="c0" title="layer 2: ViTLayer"></i><i class="c0" title="layer 3: ViTLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ViTLayer</td><td>24</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>neck.reassemble_stage.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: ConvTranspose2d"></i><i class="c0" title="layer 1: ConvTranspose2d"></i><i class="c1" title="layer 2: no ConvTranspose2d"></i><i class="c1" title="layer 3: no ConvTranspose2d"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: ConvTranspose2d"></i><i class="c0" title="layer 1: ConvTranspose2d"></i><i class="c1" title="layer 2: no ConvTranspose2d"></i><i class="c1" title="layer 3: no ConvTranspose2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ConvTranspose2d</td><td>2</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>no ConvTranspose2d</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>neck.reassemble_stage.readout_projects</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>neck.fusion_stage.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: FeatureFusionLayer"></i><i class="c0" title="layer 1: FeatureFusionLayer"></i><i class="c0" title="layer 2: FeatureFusionLayer"></i><i class="c0" title="layer 3: FeatureFusionLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: FeatureFusionLayer"></i><i class="c0" title="layer 1: FeatureFusionLayer"></i><i class="c0" title="layer 2: FeatureFusionLayer"></i><i class="c0" title="layer 3: FeatureFusionLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>FeatureFusionLayer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.881290" title="dpt.encoder.layer: 302.3M (88.1%)"></i><i class="c1" style="flex:0.042044" title="neck.reassemble_stage.layers: 14.4M (4.2%)"></i><i class="c2" style="flex:0.028290" title="neck.fusion_stage.layers: 9.7M (2.8%)"></i><i class="c3" style="flex:0.024466" title="neck.reassemble_stage.readout_projects: 8.4M (2.4%)"></i><i class="c4" style="flex:0.018914" title="neck.convs: 6.5M (1.9%)"></i><i class="c5" style="flex:0.004021" title="dpt.embeddings: 1.4M (0.4%)"></i><i class="c6" style="flex:0.000968" title="head.head: 332.0K (0.1%)"></i><i class="c7" style="flex:0.000006" title="1 others: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.127943" title="dpt.encoder.layer: 133.9K (12.8%)"></i><i class="c1" style="flex:0.054094" title="neck.reassemble_stage.layers: 56.6K (5.4%)"></i><i class="c2" style="flex:0.580514" title="neck.fusion_stage.layers: 607.5K (58.1%)"></i><i class="c3" style="flex:0.031558" title="neck.reassemble_stage.readout_projects: 33.0K (3.2%)"></i><i class="c4" style="flex:0.096875" title="neck.convs: 101.4K (9.7%)"></i><i class="c5" style="flex:0.082380" title="dpt.embeddings: 86.2K (8.2%)"></i><i class="c6" style="flex:0.026513" title="head.head: 27.7K (2.7%)"></i><i class="c7" style="flex:0.000122" title="1 others: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>dpt.encoder.layer</td><td>302.3M</td><td>88.1%</td><td>133.9K</td><td>12.8%</td></tr><tr><td><i class="c1"></i>neck.reassemble_stage.layers</td><td>14.4M</td><td>4.2%</td><td>56.6K</td><td>5.4%</td></tr><tr><td><i class="c2"></i>neck.fusion_stage.layers</td><td>9.7M</td><td>2.8%</td><td>607.5K</td><td>58.1%</td></tr><tr><td><i class="c3"></i>neck.reassemble_stage.readout_projects</td><td>8.4M</td><td>2.4%</td><td>33.0K</td><td>3.2%</td></tr><tr><td><i class="c4"></i>neck.convs</td><td>6.5M</td><td>1.9%</td><td>101.4K</td><td>9.7%</td></tr><tr><td><i class="c5"></i>dpt.embeddings</td><td>1.4M</td><td>0.4%</td><td>86.2K</td><td>8.2%</td></tr><tr><td><i class="c6"></i>head.head</td><td>332.0K</td><td>0.1%</td><td>27.7K</td><td>2.7%</td></tr><tr><td><i class="c7"></i>1 others</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>dpt</td><td>dpt</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>384</td><td>384</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix DPT omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>148</td><td>28</td></tr><tr class=""><td>Dropout</td><td>49</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>49</td><td>9</td></tr><tr class=""><td>Conv2d</td><td>33</td><td>33</td></tr><tr class=""><td>GELUActivation</td><td>28</td><td>8</td></tr><tr class=""><td>SelfAttention</td><td>24</td><td>4</td></tr><tr class=""><td>ViTAttention</td><td>24</td><td>4</td></tr><tr class=""><td>ViTIntermediate</td><td>24</td><td>4</td></tr><tr class=""><td>ViTLayer</td><td>24</td><td>4</td></tr><tr class=""><td>ViTOutput</td><td>24</td><td>4</td></tr><tr class=""><td>ViTSelfOutput</td><td>24</td><td>4</td></tr><tr class=""><td>ReLU</td><td>18</td><td>18</td></tr><tr class=""><td>PreActResidualLayer</td><td>8</td><td>8</td></tr><tr class=""><td>ModuleList</td><td>5</td><td>5</td></tr><tr class=""><td>Sequential</td><td>5</td><td>5</td></tr><tr class=""><td>FeatureFusionLayer</td><td>4</td><td>4</td></tr><tr class=""><td>ReassembleLayer</td><td>4</td><td>4</td></tr><tr class=""><td>ConvTranspose2d</td><td>2</td><td>2</td></tr><tr class=""><td>DepthEstimationHead</td><td>1</td><td>1</td></tr><tr class=""><td>FeatureFusionStage</td><td>1</td><td>1</td></tr><tr class=""><td>ForDepthEstimation</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Neck</td><td>1</td><td>1</td></tr><tr class=""><td>ReassembleStage</td><td>1</td><td>1</td></tr><tr class=""><td>ViTEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>ViTEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>ViTPatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Identity</td><td>1</td><td>1</td></tr><tr class=""><td>Upsample</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 34 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>backbone_out_indices</td><td>[5, 11, 17, 23]</td><td>[0, 1, 2, 3]</td></tr><tr><td>fusion_hidden_size</td><td>256</td><td>64</td></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>neck_hidden_sizes</td><td>[256, 512, 1024, 1024]</td><td>[16, 32, 64, 64]</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/dpt-large-mic.json b/shrink/dpt-large-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..cc4db33b360f2f4b214429f1227bfa32c20b4892
--- /dev/null
+++ b/shrink/dpt-large-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/dpt-large-mic@fef05122ed4bc80c9abd66c8dfb177787005eb45", "findings": []}
\ No newline at end of file
diff --git a/shrink/electra-small-discriminator-mic.html b/shrink/electra-small-discriminator-mic.html
index 1d1848adc3a0cc204a4e2fc8e7a3f6e3b858c729..a44fdfa5d8b3d5205c76fb8f75560d88b5ebc75e 100644
--- a/shrink/electra-small-discriminator-mic.html
+++ b/shrink/electra-small-discriminator-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/electra-small-discriminator: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/electra-small-discriminator-mic@ac7792ec804c827b79e2d41cc27764deb121a907 ← google/electra-small-discriminator@fa8239aadc095e9164941d05878b98afe9b953c3 · ElectraForPreTraining · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/electra-small-discriminator</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/electra-small-discriminator-mic/tree/ac7792ec804c827b79e2d41cc27764deb121a907" title="huggingface-mic/electra-small-discriminator-mic@ac7792ec804c827b79e2d41cc27764deb121a907">huggingface-mic/electra-small-discriminator-mic@ac7792ec804c</a> ← <a href="https://huggingface.co/google/electra-small-discriminator/tree/fa8239aadc095e9164941d05878b98afe9b953c3" title="google/electra-small-discriminator@fa8239aadc095e9164941d05878b98afe9b953c3">google/electra-small-discriminator@fa8239aadc09</a> · ElectraForPreTraining · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">13.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>electra.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.699467" title="electra.encoder.layer: 9.5M (69.9%)"></i><i class="c1" style="flex:0.293221" title="electra.embeddings: 4.0M (29.3%)"></i><i class="c2" style="flex:0.004856" title="discriminator_predictions.dense: 65.8K (0.5%)"></i><i class="c3" style="flex:0.002437" title="electra.embeddings_project: 33.0K (0.2%)"></i><i class="c4" style="flex:0.000019" title="discriminator_predictions.dense_prediction: 257 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.016520" title="electra.encoder.layer: 66.9K (1.7%)"></i><i class="c1" style="flex:0.980400" title="electra.embeddings: 4.0M (98.0%)"></i><i class="c2" style="flex:0.001027" title="discriminator_predictions.dense: 4.2K (0.1%)"></i><i class="c3" style="flex:0.002037" title="electra.embeddings_project: 8.3K (0.2%)"></i><i class="c4" style="flex:0.000016" title="discriminator_predictions.dense_prediction: 65 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>electra.encoder.layer</td><td>9.5M</td><td>69.9%</td><td>66.9K</td><td>1.7%</td></tr><tr><td><i class="c1"></i>electra.embeddings</td><td>4.0M</td><td>29.3%</td><td>4.0M</td><td>98.0%</td></tr><tr><td><i class="c2"></i>discriminator_predictions.dense</td><td>65.8K</td><td>0.5%</td><td>4.2K</td><td>0.1%</td></tr><tr><td><i class="c3"></i>electra.embeddings_project</td><td>33.0K</td><td>0.2%</td><td>8.3K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>discriminator_predictions.dense_prediction</td><td>257</td><td>0.0%</td><td>65</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>electra</td><td>electra</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Electra omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>75</td><td>15</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>GELUActivation</td><td>13</td><td>3</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>DiscriminatorPredictions</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForPreTraining</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">3 fields changed, 32 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>256</td><td>64</td></tr><tr><td>intermediate_size</td><td>1024</td><td>128</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/electra-small-discriminator-mic.json b/shrink/electra-small-discriminator-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..8cf91239fd15bdf9ac7b7b32b83091073934b2b6
--- /dev/null
+++ b/shrink/electra-small-discriminator-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/electra-small-discriminator-mic@ac7792ec804c827b79e2d41cc27764deb121a907", "findings": []}
\ No newline at end of file
diff --git a/shrink/embeddinggemma-2-mic.html b/shrink/embeddinggemma-2-mic.html
index 3c6421de6fa2d5ce2706e849f217426dbbb07bef..e65ad15ccd2c95197c18dc2cb4b2c88269864d20 100644
--- a/shrink/embeddinggemma-2-mic.html
+++ b/shrink/embeddinggemma-2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/embeddinggemma-2: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/embeddinggemma-2-mic@154c2d84cda18ee708d45273b3e509a34dc9d5cb ← google/embeddinggemma-2@f9972ac22e696a105c482202646979e33e4b5a1e · EmbeddingGemma2Model · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/embeddinggemma-2</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/embeddinggemma-2-mic/tree/154c2d84cda18ee708d45273b3e509a34dc9d5cb" title="huggingface-mic/embeddinggemma-2-mic@154c2d84cda18ee708d45273b3e509a34dc9d5cb">huggingface-mic/embeddinggemma-2-mic@154c2d84cda1</a> ← <a href="https://huggingface.co/google/embeddinggemma-2/tree/f9972ac22e696a105c482202646979e33e4b5a1e" title="google/embeddinggemma-2@f9972ac22e696a105c482202646979e33e4b5a1e">google/embeddinggemma-2@f9972ac22e69</a> · EmbeddingGemma2Model · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">744.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">19.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 1: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 2: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 3: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 4: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 5: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 6: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 7: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 8: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 9: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 10: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 11: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 12: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 13: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 14: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 15: Gemma4VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Gemma4VisionEncoderLayer"></i><i class="c0" title="layer 1: Gemma4VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Gemma4VisionEncoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i><i class="c0" title="layer 12: EncoderLayer"></i><i class="c0" title="layer 13: EncoderLayer"></i><i class="c0" title="layer 14: EncoderLayer"></i><i class="c0" title="layer 15: EncoderLayer"></i><i class="c0" title="layer 16: EncoderLayer"></i><i class="c0" title="layer 17: EncoderLayer"></i><i class="c0" title="layer 18: EncoderLayer"></i><i class="c0" title="layer 19: EncoderLayer"></i><i class="c0" title="layer 20: EncoderLayer"></i><i class="c0" title="layer 21: EncoderLayer"></i><i class="c0" title="layer 22: EncoderLayer"></i><i class="c0" title="layer 23: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>audio_tower.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Gemma4AudioLayer"></i><i class="c0" title="layer 1: Gemma4AudioLayer"></i><i class="c0" title="layer 2: Gemma4AudioLayer"></i><i class="c0" title="layer 3: Gemma4AudioLayer"></i><i class="c0" title="layer 4: Gemma4AudioLayer"></i><i class="c0" title="layer 5: Gemma4AudioLayer"></i><i class="c0" title="layer 6: Gemma4AudioLayer"></i><i class="c0" title="layer 7: Gemma4AudioLayer"></i><i class="c0" title="layer 8: Gemma4AudioLayer"></i><i class="c0" title="layer 9: Gemma4AudioLayer"></i><i class="c0" title="layer 10: Gemma4AudioLayer"></i><i class="c0" title="layer 11: Gemma4AudioLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Gemma4AudioLayer"></i><i class="c0" title="layer 1: Gemma4AudioLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Gemma4AudioLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.405931" title="audio_tower.layers: 302.2M (40.6%)"></i><i class="c1" style="flex:0.202918" title="vision_tower.encoder.layers: 151.0M (20.3%)"></i><i class="c2" style="flex:0.180310" title="language_model.embed_tokens: 134.2M (18.0%)"></i><i class="c3" style="flex:0.174777" title="language_model.layers: 130.1M (17.5%)"></i><i class="c4" style="flex:0.021922" title="vision_tower.patch_embedder: 16.3M (2.2%)"></i><i class="c5" style="flex:0.008453" title="language_model.ple: 6.3M (0.8%)"></i><i class="c6" style="flex:0.002115" title="audio_tower.output_proj: 1.6M (0.2%)"></i><i class="c7" style="flex:0.003574" title="5 others: 2.7M (0.4%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.010249" title="audio_tower.layers: 198.4K (1.0%)"></i><i class="c1" style="flex:0.004261" title="vision_tower.encoder.layers: 82.5K (0.4%)"></i><i class="c2" style="flex:0.866542" title="language_model.embed_tokens: 16.8M (86.7%)"></i><i class="c3" style="flex:0.026955" title="language_model.layers: 521.9K (2.7%)"></i><i class="c4" style="flex:0.070237" title="vision_tower.patch_embedder: 1.4M (7.0%)"></i><i class="c5" style="flex:0.003411" title="language_model.ple: 66.0K (0.3%)"></i><i class="c6" style="flex:0.005157" title="audio_tower.output_proj: 99.8K (0.5%)"></i><i class="c7" style="flex:0.013188" title="5 others: 255.3K (1.3%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>audio_tower.layers</td><td>302.2M</td><td>40.6%</td><td>198.4K</td><td>1.0%</td></tr><tr><td><i class="c1"></i>vision_tower.encoder.layers</td><td>151.0M</td><td>20.3%</td><td>82.5K</td><td>0.4%</td></tr><tr><td><i class="c2"></i>language_model.embed_tokens</td><td>134.2M</td><td>18.0%</td><td>16.8M</td><td>86.7%</td></tr><tr><td><i class="c3"></i>language_model.layers</td><td>130.1M</td><td>17.5%</td><td>521.9K</td><td>2.7%</td></tr><tr><td><i class="c4"></i>vision_tower.patch_embedder</td><td>16.3M</td><td>2.2%</td><td>1.4M</td><td>7.0%</td></tr><tr><td><i class="c5"></i>language_model.ple</td><td>6.3M</td><td>0.8%</td><td>66.0K</td><td>0.3%</td></tr><tr><td><i class="c6"></i>audio_tower.output_proj</td><td>1.6M</td><td>0.2%</td><td>99.8K</td><td>0.5%</td></tr><tr><td><i class="c7"></i>5 others</td><td>2.7M</td><td>0.4%</td><td>255.3K</td><td>1.3%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>embedding_gemma2</td><td>embedding_gemma2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (4 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>embedding_gemma2_text</td><td>embedding_gemma2_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>attention style</td><td>MHA (12 heads, 12 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>gemma4_vision</td><td>gemma4_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>gemma4_audio</td><td>gemma4_audio</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix EmbeddingGemma2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>467</td><td>61</td></tr><tr class=""><td>Gemma4ClippableLinear</td><td>232</td><td>34</td></tr><tr class=""><td>Gemma4RMSNorm</td><td>220</td><td>32</td></tr><tr class=""><td>RMSNorm</td><td>196</td><td>20</td></tr><tr class=""><td>GELUTanh</td><td>64</td><td>6</td></tr><tr class=""><td>SiLUActivation</td><td>36</td><td>6</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>TextPLEBlock</td><td>24</td><td>2</td></tr><tr class=""><td>Gemma4AudioFeedForward</td><td>24</td><td>4</td></tr><tr class=""><td>Gemma4VisionAttention</td><td>16</td><td>2</td></tr><tr class=""><td>Gemma4VisionEncoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>Gemma4VisionMLP</td><td>16</td><td>2</td></tr><tr class=""><td>Gemma4AudioAttention</td><td>12</td><td>2</td></tr><tr class=""><td>Gemma4AudioCausalConv1d</td><td>12</td><td>2</td></tr><tr class=""><td>Gemma4AudioLayer</td><td>12</td><td>2</td></tr><tr class=""><td>Gemma4AudioLightConv1d</td><td>12</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Conv2d</td><td>2</td><td>2</td></tr><tr class=""><td>MultimodalEmbedder</td><td>2</td><td>2</td></tr><tr class=""><td>Gemma4AudioSubSampleConvProjectionLayer</td><td>2</td><td>2</td></tr><tr class=""><td>LayerNorm</td><td>2</td><td>2</td></tr><tr class=""><td>ReLU</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextPLE</td><td>1</td><td>1</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4AudioRelPositionalEncoding</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4AudioSubSampleConvProjection</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4VisionPatchEmbedder</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4VisionPooler</td><td>1</td><td>1</td></tr><tr class=""><td>Gemma4VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">19 fields changed, 89 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>audio_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>audio_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>512</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>24 entries: 20× sliding_attention, 4× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>text_config.per_layer_config.05</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.1</td><td>—</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td></tr><tr><td>text_config.per_layer_config.11</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.17</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.23</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>vision_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>vision_config.num_key_value_heads</td><td>12</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/embeddinggemma-2-mic.json b/shrink/embeddinggemma-2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..ae892a30bfcb770f2aaf49315ecbfc4e82c55b5e
--- /dev/null
+++ b/shrink/embeddinggemma-2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/embeddinggemma-2-mic@154c2d84cda18ee708d45273b3e509a34dc9d5cb", "findings": []}
\ No newline at end of file
diff --git a/shrink/esm2_t6_8M_UR50D-mic.html b/shrink/esm2_t6_8M_UR50D-mic.html
index ff7449cfd49e2d875c9e6d623d2040eaef4077b6..adcf1ae58ec616e9e721b7cfc8a4f34c93a99741 100644
--- a/shrink/esm2_t6_8M_UR50D-mic.html
+++ b/shrink/esm2_t6_8M_UR50D-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/esm2_t6_8M_UR50D: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/esm2_t6_8M_UR50D-mic@4b96e1d6e983ef0096ac5ca7257ccb03fdd256ca ← facebook/esm2_t6_8M_UR50D@c731040fcd8d73dceaa04b0a8e6329b345b0f5df · EsmForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/esm2_t6_8M_UR50D</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/esm2_t6_8M_UR50D-mic/tree/4b96e1d6e983ef0096ac5ca7257ccb03fdd256ca" title="huggingface-mic/esm2_t6_8M_UR50D-mic@4b96e1d6e983ef0096ac5ca7257ccb03fdd256ca">huggingface-mic/esm2_t6_8M_UR50D-mic@4b96e1d6e983</a> ← <a href="https://huggingface.co/facebook/esm2_t6_8M_UR50D/tree/c731040fcd8d73dceaa04b0a8e6329b345b0f5df" title="facebook/esm2_t6_8M_UR50D@c731040fcd8d73dceaa04b0a8e6329b345b0f5df">facebook/esm2_t6_8M_UR50D@c731040fcd8d</a> · EsmForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">7.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">73.5K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>esm.encoder.layer</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.984730" title="esm.encoder.layer: 7.4M (98.5%)"></i><i class="c1" style="flex:0.013673" title="lm_head.dense: 102.7K (1.4%)"></i><i class="c2" style="flex:0.001406" title="esm.embeddings: 10.6K (0.1%)"></i><i class="c3" style="flex:0.000085" title="lm_head.layer_norm: 640 (0.0%)"></i><i class="c4" style="flex:0.000085" title="esm.encoder: 640 (0.0%)"></i><i class="c5" style="flex:0.000016" title="esm.contact_head: 121 (0.0%)"></i><i class="c6" style="flex:0.000004" title="lm_head: 33 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.910629" title="esm.encoder.layer: 66.9K (91.1%)"></i><i class="c1" style="flex:0.056588" title="lm_head.dense: 4.2K (5.7%)"></i><i class="c2" style="flex:0.028729" title="esm.embeddings: 2.1K (2.9%)"></i><i class="c3" style="flex:0.001741" title="lm_head.layer_norm: 128 (0.2%)"></i><i class="c4" style="flex:0.001741" title="esm.encoder: 128 (0.2%)"></i><i class="c5" style="flex:0.000122" title="esm.contact_head: 9 (0.0%)"></i><i class="c6" style="flex:0.000449" title="lm_head: 33 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>esm.encoder.layer</td><td>7.4M</td><td>98.5%</td><td>66.9K</td><td>91.1%</td></tr><tr><td><i class="c1"></i>lm_head.dense</td><td>102.7K</td><td>1.4%</td><td>4.2K</td><td>5.7%</td></tr><tr><td><i class="c2"></i>esm.embeddings</td><td>10.6K</td><td>0.1%</td><td>2.1K</td><td>2.9%</td></tr><tr><td><i class="c3"></i>lm_head.layer_norm</td><td>640</td><td>0.0%</td><td>128</td><td>0.2%</td></tr><tr><td><i class="c4"></i>esm.encoder</td><td>640</td><td>0.0%</td><td>128</td><td>0.2%</td></tr><tr><td><i class="c5"></i>esm.contact_head</td><td>121</td><td>0.0%</td><td>9</td><td>0.0%</td></tr><tr><td><i class="c6"></i>lm_head</td><td>33</td><td>0.0%</td><td>33</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>esm</td><td>esm</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>33</td><td>33</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>1026</td><td>1026</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>rope_theta</td><td>10000.0</td><td>10000.0</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Esm omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>39</td><td>15</td></tr><tr class=""><td>LayerNorm</td><td>14</td><td>6</td></tr><tr class=""><td>Dropout</td><td>13</td><td>5</td></tr><tr class=""><td>Attention</td><td>6</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>6</td><td>2</td></tr><tr class=""><td>Layer</td><td>6</td><td>2</td></tr><tr class=""><td>Output</td><td>6</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>6</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>6</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ContactPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>LMHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>Sigmoid</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 32 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>320</td><td>64</td></tr><tr><td>intermediate_size</td><td>1280</td><td>128</td></tr><tr><td>num_attention_heads</td><td>20</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>6</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/esm2_t6_8M_UR50D-mic.json b/shrink/esm2_t6_8M_UR50D-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..eb93c5d627e5eb42a94c5282ea8cff0b0d212b89
--- /dev/null
+++ b/shrink/esm2_t6_8M_UR50D-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/esm2_t6_8M_UR50D-mic@4b96e1d6e983ef0096ac5ca7257ccb03fdd256ca", "findings": []}
\ No newline at end of file
diff --git a/shrink/falcon-7b-mic.html b/shrink/falcon-7b-mic.html
index 63d30c5c1cfa99307e9408d346871377ecee1e71..7e1d362810497760f6fb85c9bfcb5aed6a20477d 100644
--- a/shrink/falcon-7b-mic.html
+++ b/shrink/falcon-7b-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>tiiuae/falcon-7b: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/falcon-7b-mic@c2ffcf79ca756a8157331b274fea56a5485a34d0 ← tiiuae/falcon-7b@ec89142b67d748a1865ea4451372db8313ada0d8 · FalconForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>tiiuae/falcon-7b</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/falcon-7b-mic/tree/c2ffcf79ca756a8157331b274fea56a5485a34d0" title="huggingface-mic/falcon-7b-mic@c2ffcf79ca756a8157331b274fea56a5485a34d0">huggingface-mic/falcon-7b-mic@c2ffcf79ca75</a> ← <a href="https://huggingface.co/tiiuae/falcon-7b/tree/ec89142b67d748a1865ea4451372db8313ada0d8" title="tiiuae/falcon-7b@ec89142b67d748a1865ea4451372db8313ada0d8">tiiuae/falcon-7b@ec89142b67d7</a> · FalconForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">6.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">8.8M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>transformer.h</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.957311" title="transformer.h: 6.6B (95.7%)"></i><i class="c1" style="flex:0.042687" title="transformer.word_embeddings: 295.5M (4.3%)"></i><i class="c2" style="flex:0.000001" title="transformer.ln_f: 9.1K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.528981" title="transformer.h: 4.7M (52.9%)"></i><i class="c1" style="flex:0.471004" title="transformer.word_embeddings: 4.2M (47.1%)"></i><i class="c2" style="flex:0.000014" title="transformer.ln_f: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>transformer.h</td><td>6.6B</td><td>95.7%</td><td>4.7M</td><td>52.9%</td></tr><tr><td><i class="c1"></i>transformer.word_embeddings</td><td>295.5M</td><td>4.3%</td><td>4.2M</td><td>47.1%</td></tr><tr><td><i class="c2"></i>transformer.ln_f</td><td>9.1K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>falcon</td><td>falcon</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>65024</td><td>65024</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Falcon omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>128</td><td>8</td></tr><tr class=""><td>LayerNorm</td><td>33</td><td>3</td></tr><tr class=""><td>Dropout</td><td>32</td><td>2</td></tr><tr class=""><td>Attention</td><td>32</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Linear</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">3 fields changed, 32 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>4544</td><td>64</td></tr><tr><td>num_attention_heads</td><td>71</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/falcon-7b-mic.json b/shrink/falcon-7b-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..22bf147b41a6c15da1ec9bf3de401ebab252e0d6
--- /dev/null
+++ b/shrink/falcon-7b-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/falcon-7b-mic@c2ffcf79ca756a8157331b274fea56a5485a34d0", "findings": []}
\ No newline at end of file
diff --git a/shrink/fnet-base-mic.html b/shrink/fnet-base-mic.html
index f33490964dd6ac942fba63d30f4c662f915e939b..e6441d73c44898901887f0bfd3e52d4f19ea49f1 100644
--- a/shrink/fnet-base-mic.html
+++ b/shrink/fnet-base-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/fnet-base: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/fnet-base-mic@0b11088ebc19ded28553b5f9b49c0c66d5d4c346 ← google/fnet-base@d89b6fad3cf5384848b783dc480f9685f49d008c · FNetForPreTraining · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/fnet-base</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/fnet-base-mic/tree/0b11088ebc19ded28553b5f9b49c0c66d5d4c346" title="huggingface-mic/fnet-base-mic@0b11088ebc19ded28553b5f9b49c0c66d5d4c346">huggingface-mic/fnet-base-mic@0b11088ebc19</a> ← <a href="https://huggingface.co/google/fnet-base/tree/d89b6fad3cf5384848b783dc480f9685f49d008c" title="google/fnet-base@d89b6fad3cf5384848b783dc480f9685f49d008c">google/fnet-base@d89b6fad3cf5</a> · FNetForPreTraining · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">83.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>fnet.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.679222" title="fnet.encoder.layer: 56.7M (67.9%)"></i><i class="c1" style="flex:0.306209" title="fnet.embeddings: 25.6M (30.6%)"></i><i class="c2" style="flex:0.007476" title="cls.predictions: 624.1K (0.7%)"></i><i class="c3" style="flex:0.007074" title="fnet.pooler: 590.6K (0.7%)"></i><i class="c4" style="flex:0.000018" title="cls.seq_relationship: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.015588" title="fnet.encoder.layer: 33.7K (1.6%)"></i><i class="c1" style="flex:0.965622" title="fnet.embeddings: 2.1M (96.6%)"></i><i class="c2" style="flex:0.016803" title="cls.predictions: 36.3K (1.7%)"></i><i class="c3" style="flex:0.001926" title="fnet.pooler: 4.2K (0.2%)"></i><i class="c4" style="flex:0.000060" title="cls.seq_relationship: 130 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>fnet.encoder.layer</td><td>56.7M</td><td>67.9%</td><td>33.7K</td><td>1.6%</td></tr><tr><td><i class="c1"></i>fnet.embeddings</td><td>25.6M</td><td>30.6%</td><td>2.1M</td><td>96.6%</td></tr><tr><td><i class="c2"></i>cls.predictions</td><td>624.1K</td><td>0.7%</td><td>36.3K</td><td>1.7%</td></tr><tr><td><i class="c3"></i>fnet.pooler</td><td>590.6K</td><td>0.7%</td><td>4.2K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>cls.seq_relationship</td><td>1.5K</td><td>0.0%</td><td>130</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>fnet</td><td>fnet</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu_new</td><td>gelu_new</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix FNet omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>29</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>Dropout</td><td>13</td><td>3</td></tr><tr class=""><td>NewGELUActivation</td><td>13</td><td>3</td></tr><tr class=""><td>BasicFourierTransform</td><td>12</td><td>2</td></tr><tr class=""><td>BasicOutput</td><td>12</td><td>2</td></tr><tr class=""><td>FourierTransform</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>3</td><td>3</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForPreTraining</td><td>1</td><td>1</td></tr><tr class=""><td>LMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr><tr class=""><td>PreTrainingHeads</td><td>1</td><td>1</td></tr><tr class=""><td>PredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">3 fields changed, 25 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/fnet-base-mic.json b/shrink/fnet-base-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..f77281db5fb7934af0747ee10ab5a3fd4750ece9
--- /dev/null
+++ b/shrink/fnet-base-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/fnet-base-mic@0b11088ebc19ded28553b5f9b49c0c66d5d4c346", "findings": []}
\ No newline at end of file
diff --git a/shrink/gemma-2-2b-it-mic.html b/shrink/gemma-2-2b-it-mic.html
index b1eb4b85463874281129a3e83eda961e67674f9d..47de708cd61d6752e088160661254293065c5f6d 100644
--- a/shrink/gemma-2-2b-it-mic.html
+++ b/shrink/gemma-2-2b-it-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/gemma-2-2b-it: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gemma-2-2b-it-mic@e3af84449295759dc428e54735fc829837f32d81 ← google/gemma-2-2b-it@299a8560bedf22ed1c72a8a11e7dce4a7f9f51f8 · Gemma2ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/gemma-2-2b-it</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gemma-2-2b-it-mic/tree/e3af84449295759dc428e54735fc829837f32d81" title="huggingface-mic/gemma-2-2b-it-mic@e3af84449295759dc428e54735fc829837f32d81">huggingface-mic/gemma-2-2b-it-mic@e3af84449295</a> ← <a href="https://huggingface.co/google/gemma-2-2b-it/tree/299a8560bedf22ed1c72a8a11e7dce4a7f9f51f8" title="google/gemma-2-2b-it@299a8560bedf22ed1c72a8a11e7dce4a7f9f51f8">google/gemma-2-2b-it@299a8560bedf</a> · Gemma2ForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.6B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 26</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c1" title="layer 5: full_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c1" title="layer 7: full_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c1" title="layer 9: full_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c1" title="layer 13: full_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c1" title="layer 15: full_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c1" title="layer 17: full_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c1" title="layer 21: full_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i><i class="c0" title="layer 24: sliding_attention"></i><i class="c1" title="layer 25: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>13</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>13</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.774388" title="model.layers: 2.0B (77.4%)"></i><i class="c1" style="flex:0.225611" title="model.embed_tokens: 589.8M (22.6%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 2.3K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004511" title="model.layers: 74.2K (0.5%)"></i><i class="c1" style="flex:0.995485" title="model.embed_tokens: 16.4M (99.5%)"></i><i class="c2" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>2.0B</td><td>77.4%</td><td>74.2K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>589.8M</td><td>22.6%</td><td>16.4M</td><td>99.5%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>2.3K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (8 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>gemma2</td><td>gemma2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>256000</td><td>256000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Gemma2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>183</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>105</td><td>9</td></tr><tr class=""><td>GELUTanh</td><td>26</td><td>2</td></tr><tr class=""><td>Attention</td><td>26</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>26</td><td>2</td></tr><tr class=""><td>MLP</td><td>26</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 29 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>256</td><td>16</td></tr><tr><td>hidden_size</td><td>2304</td><td>64</td></tr><tr><td>intermediate_size</td><td>9216</td><td>128</td></tr><tr><td>layer_types</td><td>26 entries: 13× sliding_attention, 13× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>26</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gemma-2-2b-it-mic.json b/shrink/gemma-2-2b-it-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..9332541f8cc13dfe7cd7448ad646898de2513d52
--- /dev/null
+++ b/shrink/gemma-2-2b-it-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gemma-2-2b-it-mic@e3af84449295759dc428e54735fc829837f32d81", "findings": []}
\ No newline at end of file
diff --git a/shrink/gemma-2b-mic.html b/shrink/gemma-2b-mic.html
index 1e377fe171b87006277293aa2e277bf5d5e969d2..1b1390099b7edd7eb86dbc7129ca45aaaa4441a7 100644
--- a/shrink/gemma-2b-mic.html
+++ b/shrink/gemma-2b-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/gemma-2b: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gemma-2b-mic@7ce6a0c1274da42dd2eb5c07e703e6198b444783 ← google/gemma-2b@9cf48e52b224239de00d483ec8eb84fb8d0f3a3a · GemmaForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/gemma-2b</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gemma-2b-mic/tree/7ce6a0c1274da42dd2eb5c07e703e6198b444783" title="huggingface-mic/gemma-2b-mic@7ce6a0c1274da42dd2eb5c07e703e6198b444783">huggingface-mic/gemma-2b-mic@7ce6a0c1274d</a> ← <a href="https://huggingface.co/google/gemma-2b/tree/9cf48e52b224239de00d483ec8eb84fb8d0f3a3a" title="google/gemma-2b@9cf48e52b224239de00d483ec8eb84fb8d0f3a3a">google/gemma-2b@9cf48e52b224</a> · GemmaForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.5B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 18</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>18</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.790800" title="model.layers: 2.0B (79.1%)"></i><i class="c1" style="flex:0.209199" title="model.embed_tokens: 524.3M (20.9%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004247" title="model.layers: 69.9K (0.4%)"></i><i class="c1" style="flex:0.995749" title="model.embed_tokens: 16.4M (99.6%)"></i><i class="c2" style="flex:0.000004" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>2.0B</td><td>79.1%</td><td>69.9K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>524.3M</td><td>20.9%</td><td>16.4M</td><td>99.6%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>MQA (8 heads, 1 KV)</td><td>MQA (4 heads, 1 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>gemma</td><td>gemma</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>256000</td><td>256000</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Gemma omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>127</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>37</td><td>5</td></tr><tr class=""><td>GELUTanh</td><td>18</td><td>2</td></tr><tr class=""><td>Attention</td><td>18</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>18</td><td>2</td></tr><tr class=""><td>MLP</td><td>18</td><td>2</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 24 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>256</td><td>16</td></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>16384</td><td>128</td></tr><tr><td>num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>18</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gemma-2b-mic.json b/shrink/gemma-2b-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..06ffb97ca5b443073ad66c7e7ded45cfd2526e65
--- /dev/null
+++ b/shrink/gemma-2b-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gemma-2b-mic@7ce6a0c1274da42dd2eb5c07e703e6198b444783", "findings": []}
\ No newline at end of file
diff --git a/shrink/gemma-3n-E2B-it-mic.html b/shrink/gemma-3n-E2B-it-mic.html
index 2f17c10c11a8b5893413abeea5c0434a2eb56e39..13bcd956aefe597c97a129d37f281815f46499c1 100644
--- a/shrink/gemma-3n-E2B-it-mic.html
+++ b/shrink/gemma-3n-E2B-it-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/gemma-3n-E2B-it: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gemma-3n-E2B-it-mic@bd78bb3f74709093f27ded6bdf0d329ffcf01493 ← google/gemma-3n-E2B-it@5e092ebca197cdcd8d8b195040accf22693501bc · Gemma3nForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/gemma-3n-E2B-it</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gemma-3n-E2B-it-mic/tree/bd78bb3f74709093f27ded6bdf0d329ffcf01493" title="huggingface-mic/gemma-3n-E2B-it-mic@bd78bb3f74709093f27ded6bdf0d329ffcf01493">huggingface-mic/gemma-3n-E2B-it-mic@bd78bb3f7470</a> ← <a href="https://huggingface.co/google/gemma-3n-E2B-it/tree/5e092ebca197cdcd8d8b195040accf22693501bc" title="google/gemma-3n-E2B-it@5e092ebca197cdcd8d8b195040accf22693501bc">google/gemma-3n-E2B-it@5e092ebca197</a> · Gemma3nForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">5.4B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">581.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 30</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c0" title="layer 1: sliding_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c0" title="layer 3: sliding_attention"></i><i class="c1" title="layer 4: full_attention"></i><i class="c0" title="layer 5: sliding_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c0" title="layer 7: sliding_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c1" title="layer 9: full_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c0" title="layer 11: sliding_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c0" title="layer 13: sliding_attention"></i><i class="c1" title="layer 14: full_attention"></i><i class="c0" title="layer 15: sliding_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c0" title="layer 17: sliding_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c0" title="layer 21: sliding_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c0" title="layer 23: sliding_attention"></i><i class="c1" title="layer 24: full_attention"></i><i class="c0" title="layer 25: sliding_attention"></i><i class="c0" title="layer 26: sliding_attention"></i><i class="c0" title="layer 27: sliding_attention"></i><i class="c0" title="layer 28: sliding_attention"></i><i class="c1" title="layer 29: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>24</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.audio_tower.conformer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: AudioConformerBlock"></i><i class="c0" title="layer 1: AudioConformerBlock"></i><i class="c0" title="layer 2: AudioConformerBlock"></i><i class="c0" title="layer 3: AudioConformerBlock"></i><i class="c0" title="layer 4: AudioConformerBlock"></i><i class="c0" title="layer 5: AudioConformerBlock"></i><i class="c0" title="layer 6: AudioConformerBlock"></i><i class="c0" title="layer 7: AudioConformerBlock"></i><i class="c0" title="layer 8: AudioConformerBlock"></i><i class="c0" title="layer 9: AudioConformerBlock"></i><i class="c0" title="layer 10: AudioConformerBlock"></i><i class="c0" title="layer 11: AudioConformerBlock"></i></div></div><div class="row"><span class="lab">Fixture · 12</span><div class="cells"><i class="c0" title="layer 0: AudioConformerBlock"></i><i class="c0" title="layer 1: AudioConformerBlock"></i><i class="c0" title="layer 2: AudioConformerBlock"></i><i class="c0" title="layer 3: AudioConformerBlock"></i><i class="c0" title="layer 4: AudioConformerBlock"></i><i class="c0" title="layer 5: AudioConformerBlock"></i><i class="c0" title="layer 6: AudioConformerBlock"></i><i class="c0" title="layer 7: AudioConformerBlock"></i><i class="c0" title="layer 8: AudioConformerBlock"></i><i class="c0" title="layer 9: AudioConformerBlock"></i><i class="c0" title="layer 10: AudioConformerBlock"></i><i class="c0" title="layer 11: AudioConformerBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioConformerBlock</td><td>12</td><td>12</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.478283" title="model.language_model: 2.6B (47.8%)"></i><i class="c1" style="flex:0.340249" title="model.language_model.layers: 1.8B (34.0%)"></i><i class="c2" style="flex:0.125448" title="model.audio_tower.conformer: 679.7M (12.5%)"></i><i class="c3" style="flex:0.054282" title="model.vision_tower: 294.1M (5.4%)"></i><i class="c4" style="flex:0.000823" title="model.embed_vision: 4.5M (0.1%)"></i><i class="c5" style="flex:0.000617" title="model.embed_audio: 3.3M (0.1%)"></i><i class="c6" style="flex:0.000297" title="model.audio_tower: 1.6M (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.491023" title="model.language_model: 285.3M (49.1%)"></i><i class="c1" style="flex:0.000528" title="model.language_model.layers: 306.6K (0.1%)"></i><i class="c2" style="flex:0.002049" title="model.audio_tower.conformer: 1.2M (0.2%)"></i><i class="c3" style="flex:0.506180" title="model.vision_tower: 294.1M (50.6%)"></i><i class="c4" style="flex:0.000021" title="model.embed_vision: 12.4K (0.0%)"></i><i class="c5" style="flex:0.000021" title="model.embed_audio: 12.4K (0.0%)"></i><i class="c6" style="flex:0.000178" title="model.audio_tower: 103.7K (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model</td><td>2.6B</td><td>47.8%</td><td>285.3M</td><td>49.1%</td></tr><tr><td><i class="c1"></i>model.language_model.layers</td><td>1.8B</td><td>34.0%</td><td>306.6K</td><td>0.1%</td></tr><tr><td><i class="c2"></i>model.audio_tower.conformer</td><td>679.7M</td><td>12.5%</td><td>1.2M</td><td>0.2%</td></tr><tr><td><i class="c3"></i>model.vision_tower</td><td>294.1M</td><td>5.4%</td><td>294.1M</td><td>50.6%</td></tr><tr><td><i class="c4"></i>model.embed_vision</td><td>4.5M</td><td>0.1%</td><td>12.4K</td><td>0.0%</td></tr><tr><td><i class="c5"></i>model.embed_audio</td><td>3.3M</td><td>0.1%</td><td>12.4K</td><td>0.0%</td></tr><tr><td><i class="c6"></i>model.audio_tower</td><td>1.6M</td><td>0.0%</td><td>103.7K</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>gemma3n</td><td>gemma3n</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (8 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>gemma3n_text</td><td>gemma3n_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>262400</td><td>262400</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>gemma3n_vision</td><td>gemma3n_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>vocab_size</td><td>128</td><td>128</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>gemma3n_audio</td><td>gemma3n_audio</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>vocab_size</td><td>128</td><td>128</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Gemma3n omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>543</td><td>195</td></tr><tr class=""><td>Identity</td><td>525</td><td>525</td></tr><tr class=""><td>RMSNorm</td><td>396</td><td>152</td></tr><tr class=""><td>Conv2d</td><td>251</td><td>251</td></tr><tr class=""><td>RmsNormAct2d</td><td>155</td><td>155</td></tr><tr class=""><td>Sequential</td><td>137</td><td>137</td></tr><tr class=""><td>ConvNormAct</td><td>116</td><td>116</td></tr><tr class=""><td>LayerScale2d</td><td>81</td><td>81</td></tr><tr class=""><td>Dropout</td><td>66</td><td>66</td></tr><tr class=""><td>GELUTanh</td><td>60</td><td>8</td></tr><tr class=""><td>GELU</td><td>56</td><td>56</td></tr><tr class=""><td>UniversalInvertedResidual</td><td>49</td><td>49</td></tr><tr class=""><td>Conv2dSame</td><td>33</td><td>33</td></tr><tr class=""><td>MobileAttention</td><td>33</td><td>33</td></tr><tr class=""><td>MultiQueryAttention2d</td><td>33</td><td>33</td></tr><tr class=""><td>TextAltUp</td><td>30</td><td>4</td></tr><tr class=""><td>TextAttention</td><td>30</td><td>4</td></tr><tr class=""><td>TextDecoderLayer</td><td>30</td><td>4</td></tr><tr class=""><td>TextLaurelBlock</td><td>30</td><td>4</td></tr><tr class=""><td>TextMLP</td><td>30</td><td>4</td></tr><tr class=""><td>RmsNorm2d</td><td>29</td><td>29</td></tr><tr class=""><td>AudioConformerFeedForward</td><td>24</td><td>24</td></tr><tr class=""><td>Conv1d</td><td>12</td><td>12</td></tr><tr class=""><td>AudioAttention</td><td>12</td><td>12</td></tr><tr class=""><td>AudioConformerAttention</td><td>12</td><td>12</td></tr><tr class=""><td>AudioConformerBlock</td><td>12</td><td>12</td></tr><tr class=""><td>AudioConformerLightConv1d</td><td>12</td><td>12</td></tr><tr class=""><td>AudioRelativePositionEmbedding</td><td>12</td><td>12</td></tr><tr class=""><td>ModuleList</td><td>4</td><td>4</td></tr><tr class=""><td>EdgeResidual</td><td>3</td><td>3</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>AudioCumulativeGroupNorm</td><td>2</td><td>2</td></tr><tr class=""><td>AudioSSCPConvBlock</td><td>2</td><td>2</td></tr><tr class=""><td>MultimodalEmbedder</td><td>2</td><td>2</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>ReLU</td><td>2</td><td>2</td></tr><tr class=""><td>AudioEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>AudioSubSampleConvProjection</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>MobileNetV5Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>MobileNetV5MultiScaleFusionAdapter</td><td>1</td><td>1</td></tr><tr class=""><td>TimmWrapperModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 96 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>text_config.activation_sparsity_pattern</td><td>30 entries: 20× 0.0, 10× 0.95</td><td>[0.95, 0.95, 0.0, 0.0]</td></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>30 entries: 30× 8192</td><td>[128, 128, 128, 128]</td></tr><tr><td>text_config.layer_types</td><td>30 entries: 24× sliding_attention, 6× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;, &quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>30</td><td>4</td></tr><tr><td>text_config.num_kv_shared_layers</td><td>10</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>2048</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gemma-3n-E2B-it-mic.json b/shrink/gemma-3n-E2B-it-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..ddd2edd6c53f2465841fb42cfc1bff9e3e96c260
--- /dev/null
+++ b/shrink/gemma-3n-E2B-it-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gemma-3n-E2B-it-mic@bd78bb3f74709093f27ded6bdf0d329ffcf01493", "findings": []}
\ No newline at end of file
diff --git a/shrink/gemma-4-12B-it-mic.html b/shrink/gemma-4-12B-it-mic.html
index eef052cef266a8836dd2abda5facc49bad800fa3..7de25c8592315555b3f6e97cac6c83c0f50acd02 100644
--- a/shrink/gemma-4-12B-it-mic.html
+++ b/shrink/gemma-4-12B-it-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/gemma-4-12B-it: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gemma-4-12B-it-mic@a190e22989d3fddef2c51348c8b282c409c20dc3 ← google/gemma-4-12B-it@707f0a3b8a3c7ad586ed01e27eafbad8a27dd0f7 · Gemma4UnifiedForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/gemma-4-12B-it</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gemma-4-12B-it-mic/tree/a190e22989d3fddef2c51348c8b282c409c20dc3" title="huggingface-mic/gemma-4-12B-it-mic@a190e22989d3fddef2c51348c8b282c409c20dc3">huggingface-mic/gemma-4-12B-it-mic@a190e22989d3</a> ← <a href="https://huggingface.co/google/gemma-4-12B-it/tree/707f0a3b8a3c7ad586ed01e27eafbad8a27dd0f7" title="google/gemma-4-12B-it@707f0a3b8a3c7ad586ed01e27eafbad8a27dd0f7">google/gemma-4-12B-it@707f0a3b8a3c</a> · Gemma4UnifiedForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">12.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">52.6M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 48</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c0" title="layer 1: sliding_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c0" title="layer 3: sliding_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c1" title="layer 5: full_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c0" title="layer 7: sliding_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c0" title="layer 9: sliding_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c0" title="layer 13: sliding_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c0" title="layer 15: sliding_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c1" title="layer 17: full_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c0" title="layer 19: sliding_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c0" title="layer 21: sliding_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i><i class="c0" title="layer 24: sliding_attention"></i><i class="c0" title="layer 25: sliding_attention"></i><i class="c0" title="layer 26: sliding_attention"></i><i class="c0" title="layer 27: sliding_attention"></i><i class="c0" title="layer 28: sliding_attention"></i><i class="c1" title="layer 29: full_attention"></i><i class="c0" title="layer 30: sliding_attention"></i><i class="c0" title="layer 31: sliding_attention"></i><i class="c0" title="layer 32: sliding_attention"></i><i class="c0" title="layer 33: sliding_attention"></i><i class="c0" title="layer 34: sliding_attention"></i><i class="c1" title="layer 35: full_attention"></i><i class="c0" title="layer 36: sliding_attention"></i><i class="c0" title="layer 37: sliding_attention"></i><i class="c0" title="layer 38: sliding_attention"></i><i class="c0" title="layer 39: sliding_attention"></i><i class="c0" title="layer 40: sliding_attention"></i><i class="c1" title="layer 41: full_attention"></i><i class="c0" title="layer 42: sliding_attention"></i><i class="c0" title="layer 43: sliding_attention"></i><i class="c0" title="layer 44: sliding_attention"></i><i class="c0" title="layer 45: sliding_attention"></i><i class="c0" title="layer 46: sliding_attention"></i><i class="c1" title="layer 47: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>40</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>8</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.911451" title="model.language_model.layers: 10.9B (91.1%)"></i><i class="c1" style="flex:0.084169" title="model.language_model: 1.0B (8.4%)"></i><i class="c2" style="flex:0.004174" title="model.embed_vision: 49.9M (0.4%)"></i><i class="c3" style="flex:0.000205" title="model.embed_audio: 2.5M (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006805" title="model.language_model.layers: 357.9K (0.7%)"></i><i class="c1" style="flex:0.318968" title="model.language_model: 16.8M (31.9%)"></i><i class="c2" style="flex:0.673448" title="model.embed_vision: 35.4M (67.3%)"></i><i class="c3" style="flex:0.000779" title="model.embed_audio: 41.0K (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>10.9B</td><td>91.1%</td><td>357.9K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>1.0B</td><td>8.4%</td><td>16.8M</td><td>31.9%</td></tr><tr><td><i class="c2"></i>model.embed_vision</td><td>49.9M</td><td>0.4%</td><td>35.4M</td><td>67.3%</td></tr><tr><td><i class="c3"></i>model.embed_audio</td><td>2.5M</td><td>0.0%</td><td>41.0K</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>gemma4_unified</td><td>gemma4_unified</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>gemma4_unified_text</td><td>gemma4_unified_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>gemma4_unified_vision</td><td>gemma4_unified_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>gemma4_unified_audio</td><td>gemma4_unified_audio</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Gemma4Unified omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>RMSNorm</td><td>339</td><td>17</td></tr><tr class=""><td>Linear</td><td>332</td><td>17</td></tr><tr class=""><td>GELUTanh</td><td>48</td><td>2</td></tr><tr class=""><td>TextAttention</td><td>48</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>48</td><td>2</td></tr><tr class=""><td>TextMLP</td><td>48</td><td>2</td></tr><tr class=""><td>LayerNorm</td><td>3</td><td>3</td></tr><tr class=""><td>MultimodalEmbedder</td><td>2</td><td>2</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbedder</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">16 fields changed, 82 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>3840</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>15360</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>48 entries: 40× sliding_attention, 8× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>48</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>text_config.per_layer_config.05</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.1</td><td>—</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td></tr><tr><td>text_config.per_layer_config.11</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.17</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.23</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.29</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.35</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.41</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr><tr><td>text_config.per_layer_config.47</td><td>{&quot;head_dim&quot;: 512, &quot;num_key_value_heads&quot;: 1}</td><td>—</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gemma-4-12B-it-mic.json b/shrink/gemma-4-12B-it-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..265b1f760abe461ad6a2139e081cb05f18264596
--- /dev/null
+++ b/shrink/gemma-4-12B-it-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gemma-4-12B-it-mic@a190e22989d3fddef2c51348c8b282c409c20dc3", "findings": []}
\ No newline at end of file
diff --git a/shrink/gemma-4-E2B-it-mic.html b/shrink/gemma-4-E2B-it-mic.html
index 58bced955280ab819bd00832bf014a7859361fda..da81497ae2e0b5dd2f44ee55dac2ee7d3befa939 100644
--- a/shrink/gemma-4-E2B-it-mic.html
+++ b/shrink/gemma-4-E2B-it-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/gemma-4-E2B-it: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gemma-4-E2B-it-mic@633dbd103d5cfb02c5946a85c614858cabd17053 ← google/gemma-4-E2B-it@3e22461f65e89153144f8adb70e3b8c2cc9845a7 · Gemma4ForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/gemma-4-E2B-it</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gemma-4-E2B-it-mic/tree/633dbd103d5cfb02c5946a85c614858cabd17053" title="huggingface-mic/gemma-4-E2B-it-mic@633dbd103d5cfb02c5946a85c614858cabd17053">huggingface-mic/gemma-4-E2B-it-mic@633dbd103d5c</a> ← <a href="https://huggingface.co/google/gemma-4-E2B-it/tree/3e22461f65e89153144f8adb70e3b8c2cc9845a7" title="google/gemma-4-E2B-it@3e22461f65e89153144f8adb70e3b8c2cc9845a7">google/gemma-4-E2B-it@3e22461f65e8</a> · Gemma4ForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">5.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">288.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i><i class="c0" title="layer 2: VisionEncoderLayer"></i><i class="c0" title="layer 3: VisionEncoderLayer"></i><i class="c0" title="layer 4: VisionEncoderLayer"></i><i class="c0" title="layer 5: VisionEncoderLayer"></i><i class="c0" title="layer 6: VisionEncoderLayer"></i><i class="c0" title="layer 7: VisionEncoderLayer"></i><i class="c0" title="layer 8: VisionEncoderLayer"></i><i class="c0" title="layer 9: VisionEncoderLayer"></i><i class="c0" title="layer 10: VisionEncoderLayer"></i><i class="c0" title="layer 11: VisionEncoderLayer"></i><i class="c0" title="layer 12: VisionEncoderLayer"></i><i class="c0" title="layer 13: VisionEncoderLayer"></i><i class="c0" title="layer 14: VisionEncoderLayer"></i><i class="c0" title="layer 15: VisionEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionEncoderLayer"></i><i class="c0" title="layer 1: VisionEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionEncoderLayer</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 35</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c0" title="layer 1: sliding_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c0" title="layer 3: sliding_attention"></i><i class="c1" title="layer 4: full_attention"></i><i class="c0" title="layer 5: sliding_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c0" title="layer 7: sliding_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c1" title="layer 9: full_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c0" title="layer 11: sliding_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c0" title="layer 13: sliding_attention"></i><i class="c1" title="layer 14: full_attention"></i><i class="c0" title="layer 15: sliding_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c0" title="layer 17: sliding_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c0" title="layer 21: sliding_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c0" title="layer 23: sliding_attention"></i><i class="c1" title="layer 24: full_attention"></i><i class="c0" title="layer 25: sliding_attention"></i><i class="c0" title="layer 26: sliding_attention"></i><i class="c0" title="layer 27: sliding_attention"></i><i class="c0" title="layer 28: sliding_attention"></i><i class="c1" title="layer 29: full_attention"></i><i class="c0" title="layer 30: sliding_attention"></i><i class="c0" title="layer 31: sliding_attention"></i><i class="c0" title="layer 32: sliding_attention"></i><i class="c0" title="layer 33: sliding_attention"></i><i class="c1" title="layer 34: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>28</td><td>2</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>7</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.audio_tower.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: AudioLayer"></i><i class="c0" title="layer 1: AudioLayer"></i><i class="c0" title="layer 2: AudioLayer"></i><i class="c0" title="layer 3: AudioLayer"></i><i class="c0" title="layer 4: AudioLayer"></i><i class="c0" title="layer 5: AudioLayer"></i><i class="c0" title="layer 6: AudioLayer"></i><i class="c0" title="layer 7: AudioLayer"></i><i class="c0" title="layer 8: AudioLayer"></i><i class="c0" title="layer 9: AudioLayer"></i><i class="c0" title="layer 10: AudioLayer"></i><i class="c0" title="layer 11: AudioLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: AudioLayer"></i><i class="c0" title="layer 1: AudioLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>AudioLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.541745" title="model.language_model: 2.8B (54.2%)"></i><i class="c1" style="flex:0.365053" title="model.language_model.layers: 1.9B (36.5%)"></i><i class="c2" style="flex:0.059198" title="model.audio_tower.layers: 302.2M (5.9%)"></i><i class="c3" style="flex:0.029592" title="model.vision_tower.encoder.layers: 151.0M (3.0%)"></i><i class="c4" style="flex:0.003197" title="model.vision_tower: 16.3M (0.3%)"></i><i class="c5" style="flex:0.000521" title="model.audio_tower: 2.7M (0.1%)"></i><i class="c6" style="flex:0.000462" title="model.embed_audio: 2.4M (0.0%)"></i><i class="c7" style="flex:0.000231" title="1 others: 1.2M (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.990155" title="model.language_model: 285.3M (99.0%)"></i><i class="c1" style="flex:0.003088" title="model.language_model.layers: 889.6K (0.3%)"></i><i class="c2" style="flex:0.000689" title="model.audio_tower.layers: 198.4K (0.1%)"></i><i class="c3" style="flex:0.000286" title="model.vision_tower.encoder.layers: 82.5K (0.0%)"></i><i class="c4" style="flex:0.004720" title="model.vision_tower: 1.4M (0.5%)"></i><i class="c5" style="flex:0.000706" title="model.audio_tower: 203.6K (0.1%)"></i><i class="c6" style="flex:0.000341" title="model.embed_audio: 98.3K (0.0%)"></i><i class="c7" style="flex:0.000014" title="1 others: 4.1K (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model</td><td>2.8B</td><td>54.2%</td><td>285.3M</td><td>99.0%</td></tr><tr><td><i class="c1"></i>model.language_model.layers</td><td>1.9B</td><td>36.5%</td><td>889.6K</td><td>0.3%</td></tr><tr><td><i class="c2"></i>model.audio_tower.layers</td><td>302.2M</td><td>5.9%</td><td>198.4K</td><td>0.1%</td></tr><tr><td><i class="c3"></i>model.vision_tower.encoder.layers</td><td>151.0M</td><td>3.0%</td><td>82.5K</td><td>0.0%</td></tr><tr><td><i class="c4"></i>model.vision_tower</td><td>16.3M</td><td>0.3%</td><td>1.4M</td><td>0.5%</td></tr><tr><td><i class="c5"></i>model.audio_tower</td><td>2.7M</td><td>0.1%</td><td>203.6K</td><td>0.1%</td></tr><tr><td><i class="c6"></i>model.embed_audio</td><td>2.4M</td><td>0.0%</td><td>98.3K</td><td>0.0%</td></tr><tr><td><i class="c7"></i>1 others</td><td>1.2M</td><td>0.0%</td><td>4.1K</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>gemma4</td><td>gemma4</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MQA (8 heads, 1 KV)</td><td>MQA (4 heads, 1 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>gemma4_text</td><td>gemma4_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>262144</td><td>262144</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>attention style</td><td>MHA (12 heads, 12 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>RoPE type</td><td>axial</td><td>axial</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>gemma4_vision</td><td>gemma4_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>model_type</td><td>gemma4_audio</td><td>gemma4_audio</td><td>✓ kept</td></tr><tr class=""><td>audio_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Gemma4 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>526</td><td>75</td></tr><tr class=""><td>RMSNorm</td><td>464</td><td>64</td></tr><tr class=""><td>ClippableLinear</td><td>232</td><td>34</td></tr><tr class=""><td>GELUTanh</td><td>86</td><td>10</td></tr><tr class=""><td>SiLUActivation</td><td>36</td><td>6</td></tr><tr class=""><td>TextAttention</td><td>35</td><td>4</td></tr><tr class=""><td>TextDecoderLayer</td><td>35</td><td>4</td></tr><tr class=""><td>TextMLP</td><td>35</td><td>4</td></tr><tr class=""><td>AudioFeedForward</td><td>24</td><td>4</td></tr><tr class=""><td>VisionAttention</td><td>16</td><td>2</td></tr><tr class=""><td>VisionEncoderLayer</td><td>16</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>16</td><td>2</td></tr><tr class=""><td>AudioAttention</td><td>12</td><td>2</td></tr><tr class=""><td>AudioCausalConv1d</td><td>12</td><td>2</td></tr><tr class=""><td>AudioLayer</td><td>12</td><td>2</td></tr><tr class=""><td>AudioLightConv1d</td><td>12</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Conv2d</td><td>2</td><td>2</td></tr><tr class=""><td>AudioSubSampleConvProjectionLayer</td><td>2</td><td>2</td></tr><tr class=""><td>MultimodalEmbedder</td><td>2</td><td>2</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>2</td><td>2</td></tr><tr class=""><td>LayerNorm</td><td>2</td><td>2</td></tr><tr class=""><td>ReLU</td><td>2</td><td>2</td></tr><tr class=""><td>AudioModel</td><td>1</td><td>1</td></tr><tr class=""><td>AudioRelPositionalEncoding</td><td>1</td><td>1</td></tr><tr class=""><td>AudioSubSampleConvProjection</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbedder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPooler</td><td>1</td><td>1</td></tr><tr class=""><td>VisionRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">25 fields changed, 102 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>audio_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>audio_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>audio_config.num_hidden_layers</td><td>12</td><td>2</td></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>1536</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>6144</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>35 entries: 28× sliding_attention, 7× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;, &quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>35</td><td>4</td></tr><tr><td>text_config.num_kv_shared_layers</td><td>20</td><td>2</td></tr><tr><td>text_config.per_layer_config.04</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>text_config.per_layer_config.09</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>text_config.per_layer_config.1</td><td>—</td><td>{&quot;head_dim&quot;: 512}</td></tr><tr><td>text_config.per_layer_config.14</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>text_config.per_layer_config.19</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>text_config.per_layer_config.24</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>text_config.per_layer_config.29</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>text_config.per_layer_config.3</td><td>—</td><td>{&quot;head_dim&quot;: 512}</td></tr><tr><td>text_config.per_layer_config.34</td><td>{&quot;head_dim&quot;: 512}</td><td>—</td></tr><tr><td>vision_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>16</td><td>2</td></tr><tr><td>vision_config.num_key_value_heads</td><td>12</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gemma-4-E2B-it-mic.json b/shrink/gemma-4-E2B-it-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..369f0d1b4e2ab8d14b69f88bada7d102e3a3d743
--- /dev/null
+++ b/shrink/gemma-4-E2B-it-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gemma-4-E2B-it-mic@633dbd103d5cfb02c5946a85c614858cabd17053", "findings": []}
\ No newline at end of file
diff --git a/shrink/gemma3-mic.html b/shrink/gemma3-mic.html
index fb6d4c038ab2baa1bac305c4196fb5c178e236aa..c98aa445291594fa0614171fdb6aa92c08891b4c 100644
--- a/shrink/gemma3-mic.html
+++ b/shrink/gemma3-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>google/gemma-3-4b-it: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gemma3-mic@97b8077b83d60421590191ff1d19535224e77dce ← google/gemma-3-4b-it@093f9f388b31de276ce2de164bdc2081324b9767 · Gemma3ForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>google/gemma-3-4b-it</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gemma3-mic/tree/97b8077b83d60421590191ff1d19535224e77dce" title="huggingface-mic/gemma3-mic@97b8077b83d60421590191ff1d19535224e77dce">huggingface-mic/gemma3-mic@97b8077b83d6</a> ← <a href="https://huggingface.co/google/gemma-3-4b-it/tree/093f9f388b31de276ce2de164bdc2081324b9767" title="google/gemma-3-4b-it@093f9f388b31de276ce2de164bdc2081324b9767">google/gemma-3-4b-it@093f9f388b31</a> · Gemma3ForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">17.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i><i class="c0" title="layer 2: SiglipEncoderLayer"></i><i class="c0" title="layer 3: SiglipEncoderLayer"></i><i class="c0" title="layer 4: SiglipEncoderLayer"></i><i class="c0" title="layer 5: SiglipEncoderLayer"></i><i class="c0" title="layer 6: SiglipEncoderLayer"></i><i class="c0" title="layer 7: SiglipEncoderLayer"></i><i class="c0" title="layer 8: SiglipEncoderLayer"></i><i class="c0" title="layer 9: SiglipEncoderLayer"></i><i class="c0" title="layer 10: SiglipEncoderLayer"></i><i class="c0" title="layer 11: SiglipEncoderLayer"></i><i class="c0" title="layer 12: SiglipEncoderLayer"></i><i class="c0" title="layer 13: SiglipEncoderLayer"></i><i class="c0" title="layer 14: SiglipEncoderLayer"></i><i class="c0" title="layer 15: SiglipEncoderLayer"></i><i class="c0" title="layer 16: SiglipEncoderLayer"></i><i class="c0" title="layer 17: SiglipEncoderLayer"></i><i class="c0" title="layer 18: SiglipEncoderLayer"></i><i class="c0" title="layer 19: SiglipEncoderLayer"></i><i class="c0" title="layer 20: SiglipEncoderLayer"></i><i class="c0" title="layer 21: SiglipEncoderLayer"></i><i class="c0" title="layer 22: SiglipEncoderLayer"></i><i class="c0" title="layer 23: SiglipEncoderLayer"></i><i class="c0" title="layer 24: SiglipEncoderLayer"></i><i class="c0" title="layer 25: SiglipEncoderLayer"></i><i class="c0" title="layer 26: SiglipEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>SiglipEncoderLayer</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 34</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c0" title="layer 1: sliding_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c0" title="layer 3: sliding_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c1" title="layer 5: full_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c0" title="layer 7: sliding_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c0" title="layer 9: sliding_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c0" title="layer 13: sliding_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c0" title="layer 15: sliding_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c1" title="layer 17: full_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c0" title="layer 19: sliding_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c0" title="layer 21: sliding_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i><i class="c0" title="layer 24: sliding_attention"></i><i class="c0" title="layer 25: sliding_attention"></i><i class="c0" title="layer 26: sliding_attention"></i><i class="c0" title="layer 27: sliding_attention"></i><i class="c0" title="layer 28: sliding_attention"></i><i class="c1" title="layer 29: full_attention"></i><i class="c0" title="layer 30: sliding_attention"></i><i class="c0" title="layer 31: sliding_attention"></i><i class="c0" title="layer 32: sliding_attention"></i><i class="c0" title="layer 33: sliding_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>29</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>5</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.746267" title="model.language_model.layers: 3.2B (74.6%)"></i><i class="c1" style="flex:0.156103" title="model.language_model: 671.3M (15.6%)"></i><i class="c2" style="flex:0.095688" title="model.vision_tower.encoder.layers: 411.5M (9.6%)"></i><i class="c3" style="flex:0.001256" title="model.vision_tower: 5.4M (0.1%)"></i><i class="c4" style="flex:0.000686" title="model.multi_modal_projector: 3.0M (0.1%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004313" title="model.language_model.layers: 74.3K (0.4%)"></i><i class="c1" style="flex:0.974146" title="model.language_model: 16.8M (97.4%)"></i><i class="c2" style="flex:0.003886" title="model.vision_tower.encoder.layers: 66.9K (0.4%)"></i><i class="c3" style="flex:0.017413" title="model.vision_tower: 300.0K (1.7%)"></i><i class="c4" style="flex:0.000241" title="model.multi_modal_projector: 4.2K (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>3.2B</td><td>74.6%</td><td>74.3K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>671.3M</td><td>15.6%</td><td>16.8M</td><td>97.4%</td></tr><tr><td><i class="c2"></i>model.vision_tower.encoder.layers</td><td>411.5M</td><td>9.6%</td><td>66.9K</td><td>0.4%</td></tr><tr><td><i class="c3"></i>model.vision_tower</td><td>5.4M</td><td>0.1%</td><td>300.0K</td><td>1.7%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>3.0M</td><td>0.1%</td><td>4.2K</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>gemma3</td><td>gemma3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (8 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>gemma3_text</td><td>gemma3_text</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>262208</td><td>262208</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>siglip_vision_model</td><td>siglip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>896</td><td>896</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Gemma3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>401</td><td>27</td></tr><tr class=""><td>RMSNorm</td><td>206</td><td>14</td></tr><tr class=""><td>GELUTanh</td><td>61</td><td>4</td></tr><tr class=""><td>LayerNorm</td><td>55</td><td>5</td></tr><tr class=""><td>Attention</td><td>34</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>34</td><td>2</td></tr><tr class=""><td>MLP</td><td>34</td><td>2</td></tr><tr class=""><td>SiglipAttention</td><td>27</td><td>2</td></tr><tr class=""><td>SiglipEncoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>SiglipMLP</td><td>27</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>AvgPool2d</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextScaledWordEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 60 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>256</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>2560</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>10240</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>34 entries: 29× sliding_attention, 5× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>8</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>34</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>4</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>27</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gemma3-mic.json b/shrink/gemma3-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..8dfb9f96d08db42a90e2158de5450505219d635f
--- /dev/null
+++ b/shrink/gemma3-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gemma3-mic@97b8077b83d60421590191ff1d19535224e77dce", "findings": []}
\ No newline at end of file
diff --git a/shrink/got_ocr2-mic.html b/shrink/got_ocr2-mic.html
index b6330eaf5661fc0ca6294c75e588a05eca4496d5..a1c04451651af1f8b3f9a3bca33a8405b33938b9 100644
--- a/shrink/got_ocr2-mic.html
+++ b/shrink/got_ocr2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>stepfun-ai/GOT-OCR-2.0-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/got_ocr2-mic@5bfaf853e15a6fdf1c253e1368b81df3ed528ab5 ← stepfun-ai/GOT-OCR-2.0-hf@d3017ef2c2c1395888c8d635c5e0508bcb0ac78d · GotOcr2ForConditionalGeneration · overfit weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>stepfun-ai/GOT-OCR-2.0-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/got_ocr2-mic/tree/5bfaf853e15a6fdf1c253e1368b81df3ed528ab5" title="huggingface-mic/got_ocr2-mic@5bfaf853e15a6fdf1c253e1368b81df3ed528ab5">huggingface-mic/got_ocr2-mic@5bfaf853e15a</a> ← <a href="https://huggingface.co/stepfun-ai/GOT-OCR-2.0-hf/tree/d3017ef2c2c1395888c8d635c5e0508bcb0ac78d" title="stepfun-ai/GOT-OCR-2.0-hf@d3017ef2c2c1395888c8d635c5e0508bcb0ac78d">stepfun-ai/GOT-OCR-2.0-hf@d3017ef2c2c1</a> · GotOcr2ForConditionalGeneration · overfit weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">560.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">13.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: MLPBlock"></i><i class="c0" title="layer 1: MLPBlock"></i><i class="c0" title="layer 2: MLPBlock"></i><i class="c0" title="layer 3: MLPBlock"></i><i class="c0" title="layer 4: MLPBlock"></i><i class="c0" title="layer 5: MLPBlock"></i><i class="c0" title="layer 6: MLPBlock"></i><i class="c0" title="layer 7: MLPBlock"></i><i class="c0" title="layer 8: MLPBlock"></i><i class="c0" title="layer 9: MLPBlock"></i><i class="c0" title="layer 10: MLPBlock"></i><i class="c0" title="layer 11: MLPBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: MLPBlock"></i><i class="c0" title="layer 1: MLPBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>MLPBlock</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i><i class="c0" title="layer 2: Qwen2DecoderLayer"></i><i class="c0" title="layer 3: Qwen2DecoderLayer"></i><i class="c0" title="layer 4: Qwen2DecoderLayer"></i><i class="c0" title="layer 5: Qwen2DecoderLayer"></i><i class="c0" title="layer 6: Qwen2DecoderLayer"></i><i class="c0" title="layer 7: Qwen2DecoderLayer"></i><i class="c0" title="layer 8: Qwen2DecoderLayer"></i><i class="c0" title="layer 9: Qwen2DecoderLayer"></i><i class="c0" title="layer 10: Qwen2DecoderLayer"></i><i class="c0" title="layer 11: Qwen2DecoderLayer"></i><i class="c0" title="layer 12: Qwen2DecoderLayer"></i><i class="c0" title="layer 13: Qwen2DecoderLayer"></i><i class="c0" title="layer 14: Qwen2DecoderLayer"></i><i class="c0" title="layer 15: Qwen2DecoderLayer"></i><i class="c0" title="layer 16: Qwen2DecoderLayer"></i><i class="c0" title="layer 17: Qwen2DecoderLayer"></i><i class="c0" title="layer 18: Qwen2DecoderLayer"></i><i class="c0" title="layer 19: Qwen2DecoderLayer"></i><i class="c0" title="layer 20: Qwen2DecoderLayer"></i><i class="c0" title="layer 21: Qwen2DecoderLayer"></i><i class="c0" title="layer 22: Qwen2DecoderLayer"></i><i class="c0" title="layer 23: Qwen2DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.550202" title="model.language_model.layers: 308.4M (55.0%)"></i><i class="c1" style="flex:0.277427" title="model.language_model: 155.5M (27.7%)"></i><i class="c2" style="flex:0.151905" title="model.vision_tower.layers: 85.1M (15.2%)"></i><i class="c3" style="flex:0.012395" title="model.multi_modal_projector: 6.9M (1.2%)"></i><i class="c4" style="flex:0.008071" title="model.vision_tower: 4.5M (0.8%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006338" title="model.language_model.layers: 82.6K (0.6%)"></i><i class="c1" style="flex:0.746062" title="model.language_model: 9.7M (74.6%)"></i><i class="c2" style="flex:0.063577" title="model.vision_tower.layers: 828.2K (6.4%)"></i><i class="c3" style="flex:0.113510" title="model.multi_modal_projector: 1.5M (11.4%)"></i><i class="c4" style="flex:0.070513" title="model.vision_tower: 918.6K (7.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>308.4M</td><td>55.0%</td><td>82.6K</td><td>0.6%</td></tr><tr><td><i class="c1"></i>model.language_model</td><td>155.5M</td><td>27.7%</td><td>9.7M</td><td>74.6%</td></tr><tr><td><i class="c2"></i>model.vision_tower.layers</td><td>85.1M</td><td>15.2%</td><td>828.2K</td><td>6.4%</td></tr><tr><td><i class="c3"></i>model.multi_modal_projector</td><td>6.9M</td><td>1.2%</td><td>1.5M</td><td>11.4%</td></tr><tr><td><i class="c4"></i>model.vision_tower</td><td>4.5M</td><td>0.8%</td><td>918.6K</td><td>7.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>got_ocr2</td><td>got_ocr2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>got_ocr2_vision</td><td>got_ocr2_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>1024</td><td>1024</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MHA (16 heads, 16 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2</td><td>qwen2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151860</td><td>151860</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GotOcr2 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>218</td><td>24</td></tr><tr class=""><td>Qwen2RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>LayerNorm</td><td>24</td><td>4</td></tr><tr class=""><td>Qwen2Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>MLPBlock</td><td>12</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>12</td><td>2</td></tr><tr class=""><td>VisionLayer</td><td>12</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>5</td><td>5</td></tr><tr class=""><td>LayerNorm</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>PatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionNeck</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 60 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>2816</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>24 entries: 24× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/got_ocr2-mic.json b/shrink/got_ocr2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..737ce0cdb1ec0ab154c98f87cb9e0ea3995801a7
--- /dev/null
+++ b/shrink/got_ocr2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/got_ocr2-mic@5bfaf853e15a6fdf1c253e1368b81df3ed528ab5", "findings": []}
\ No newline at end of file
diff --git a/shrink/gpt2-mic.html b/shrink/gpt2-mic.html
index f224fe762bb8393eb12e210e5d9d47ec3299587c..cbc9c4bf4961a08ebf2d571ed7a218f472b8bcb3 100644
--- a/shrink/gpt2-mic.html
+++ b/shrink/gpt2-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>openai-community/gpt2: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gpt2-mic@0168ed11cf4c9161437d47f144bdf52cdfeddd20 ← openai-community/gpt2@607a30d783dfa663caf39e06633721c8d4cfcd7e · GPT2LMHeadModel · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>openai-community/gpt2</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gpt2-mic/tree/0168ed11cf4c9161437d47f144bdf52cdfeddd20" title="huggingface-mic/gpt2-mic@0168ed11cf4c9161437d47f144bdf52cdfeddd20">huggingface-mic/gpt2-mic@0168ed11cf4c</a> ← <a href="https://huggingface.co/openai-community/gpt2/tree/607a30d783dfa663caf39e06633721c8d4cfcd7e" title="openai-community/gpt2@607a30d783dfa663caf39e06633721c8d4cfcd7e">openai-community/gpt2@607a30d783df</a> · GPT2LMHeadModel · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">124.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>transformer.h</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: GPT2Block"></i><i class="c0" title="layer 1: GPT2Block"></i><i class="c0" title="layer 2: GPT2Block"></i><i class="c0" title="layer 3: GPT2Block"></i><i class="c0" title="layer 4: GPT2Block"></i><i class="c0" title="layer 5: GPT2Block"></i><i class="c0" title="layer 6: GPT2Block"></i><i class="c0" title="layer 7: GPT2Block"></i><i class="c0" title="layer 8: GPT2Block"></i><i class="c0" title="layer 9: GPT2Block"></i><i class="c0" title="layer 10: GPT2Block"></i><i class="c0" title="layer 11: GPT2Block"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: GPT2Block"></i><i class="c0" title="layer 1: GPT2Block"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>GPT2Block</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.683499" title="transformer.h: 85.1M (68.3%)"></i><i class="c1" style="flex:0.310169" title="transformer.wte: 38.6M (31.0%)"></i><i class="c2" style="flex:0.006320" title="transformer.wpe: 786.4K (0.6%)"></i><i class="c3" style="flex:0.000012" title="transformer.ln_f: 1.5K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.029558" title="transformer.h: 100.0K (3.0%)"></i><i class="c1" style="flex:0.951027" title="transformer.wte: 3.2M (95.1%)"></i><i class="c2" style="flex:0.019377" title="transformer.wpe: 65.5K (1.9%)"></i><i class="c3" style="flex:0.000038" title="transformer.ln_f: 128 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>transformer.h</td><td>85.1M</td><td>68.3%</td><td>100.0K</td><td>3.0%</td></tr><tr><td><i class="c1"></i>transformer.wte</td><td>38.6M</td><td>31.0%</td><td>3.2M</td><td>95.1%</td></tr><tr><td><i class="c2"></i>transformer.wpe</td><td>786.4K</td><td>0.6%</td><td>65.5K</td><td>1.9%</td></tr><tr><td><i class="c3"></i>transformer.ln_f</td><td>1.5K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>gpt2</td><td>gpt2</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>50257</td><td>50257</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GPT2LMHead omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Conv1D</td><td>48</td><td>8</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>GPT2Attention</td><td>12</td><td>2</td></tr><tr class=""><td>GPT2Block</td><td>12</td><td>2</td></tr><tr class=""><td>GPT2MLP</td><td>12</td><td>2</td></tr><tr class=""><td>NewGELUActivation</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>GPT2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Linear</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">3 fields changed, 34 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>n_embd</td><td>768</td><td>64</td></tr><tr><td>n_head</td><td>12</td><td>4</td></tr><tr><td>n_layer</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gpt2-mic.json b/shrink/gpt2-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a9260b8921f6bf6ab7828b2dba36d88040579e4f
--- /dev/null
+++ b/shrink/gpt2-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gpt2-mic@0168ed11cf4c9161437d47f144bdf52cdfeddd20", "findings": []}
\ No newline at end of file
diff --git a/shrink/gpt_oss-mic.html b/shrink/gpt_oss-mic.html
index d6c621519e8ecb3e9ca0ad152e11fcf0972f3847..1ba34ee610eb5ffc56c82277e6ec67c418b9280a 100644
--- a/shrink/gpt_oss-mic.html
+++ b/shrink/gpt_oss-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>openai/gpt-oss-20b: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/gpt_oss-mic@8fea014b7848db31a7032015195dc9b6947e63e2 ← openai/gpt-oss-20b@6cee5e81ee83917806bbde320786a8fb61efebee · GptOssForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">20.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">26.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c1" title="layer 5: full_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c1" title="layer 7: full_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c1" title="layer 9: full_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c1" title="layer 13: full_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c1" title="layer 15: full_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c1" title="layer 17: full_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c1" title="layer 21: full_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>12</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>12</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.944619" title="model.layers: 19.8B (94.5%)"></i><i class="c1" style="flex:0.027690" title="model.embed_tokens: 579.1M (2.8%)"></i><i class="c2" style="flex:0.027690" title="lm_head: 579.1M (2.8%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.9K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008662" title="model.layers: 224.9K (0.9%)"></i><i class="c1" style="flex:0.495668" title="model.embed_tokens: 12.9M (49.6%)"></i><i class="c2" style="flex:0.495668" title="lm_head: 12.9M (49.6%)"></i><i class="c3" style="flex:0.000002" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>19.8B</td><td>94.5%</td><td>224.9K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>579.1M</td><td>2.8%</td><td>12.9M</td><td>49.6%</td></tr><tr><td><i class="c2"></i>lm_head</td><td>579.1M</td><td>2.8%</td><td>12.9M</td><td>49.6%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.9K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (64 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-4 of 32 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>gpt_oss</td><td>gpt_oss</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>201088</td><td>201088</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GptOss omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>97</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Experts</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 29 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>2880</td><td>64</td></tr><tr><td>intermediate_size</td><td>2880</td><td>128</td></tr><tr><td>layer_types</td><td>24 entries: 12× sliding_attention, 12× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>4</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>num_local_experts</td><td>32</td><td>4</td></tr><tr><td>quantization_config</td><td>{&quot;modules_to_not_convert&quot;: [&quot;model.layers.*.self_attn&quot;, &quot;model.layers.*.mlp.r…</td><td>null</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>openai/gpt-oss-20b</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/gpt_oss-mic/tree/8fea014b7848db31a7032015195dc9b6947e63e2" title="huggingface-mic/gpt_oss-mic@8fea014b7848db31a7032015195dc9b6947e63e2">huggingface-mic/gpt_oss-mic@8fea014b7848</a> ← <a href="https://huggingface.co/openai/gpt-oss-20b/tree/6cee5e81ee83917806bbde320786a8fb61efebee" title="openai/gpt-oss-20b@6cee5e81ee83917806bbde320786a8fb61efebee">openai/gpt-oss-20b@6cee5e81ee83</a> · GptOssForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">20.9B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">26.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i><i class="c0" title="layer 2: sliding_attention"></i><i class="c1" title="layer 3: full_attention"></i><i class="c0" title="layer 4: sliding_attention"></i><i class="c1" title="layer 5: full_attention"></i><i class="c0" title="layer 6: sliding_attention"></i><i class="c1" title="layer 7: full_attention"></i><i class="c0" title="layer 8: sliding_attention"></i><i class="c1" title="layer 9: full_attention"></i><i class="c0" title="layer 10: sliding_attention"></i><i class="c1" title="layer 11: full_attention"></i><i class="c0" title="layer 12: sliding_attention"></i><i class="c1" title="layer 13: full_attention"></i><i class="c0" title="layer 14: sliding_attention"></i><i class="c1" title="layer 15: full_attention"></i><i class="c0" title="layer 16: sliding_attention"></i><i class="c1" title="layer 17: full_attention"></i><i class="c0" title="layer 18: sliding_attention"></i><i class="c1" title="layer 19: full_attention"></i><i class="c0" title="layer 20: sliding_attention"></i><i class="c1" title="layer 21: full_attention"></i><i class="c0" title="layer 22: sliding_attention"></i><i class="c1" title="layer 23: full_attention"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: sliding_attention"></i><i class="c1" title="layer 1: full_attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>sliding_attention</td><td>12</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>full_attention</td><td>12</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.944619" title="model.layers: 19.8B (94.5%)"></i><i class="c1" style="flex:0.027690" title="lm_head: 579.1M (2.8%)"></i><i class="c2" style="flex:0.027690" title="model.embed_tokens: 579.1M (2.8%)"></i><i class="c3" style="flex:0.000000" title="model.norm: 2.9K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008662" title="model.layers: 224.9K (0.9%)"></i><i class="c1" style="flex:0.495668" title="lm_head: 12.9M (49.6%)"></i><i class="c2" style="flex:0.495668" title="model.embed_tokens: 12.9M (49.6%)"></i><i class="c3" style="flex:0.000002" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>19.8B</td><td>94.5%</td><td>224.9K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>lm_head</td><td>579.1M</td><td>2.8%</td><td>12.9M</td><td>49.6%</td></tr><tr><td><i class="c2"></i>model.embed_tokens</td><td>579.1M</td><td>2.8%</td><td>12.9M</td><td>49.6%</td></tr><tr><td><i class="c3"></i>model.norm</td><td>2.9K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (64 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-4 of 32 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>yarn</td><td>yarn</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>gpt_oss</td><td>gpt_oss</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>201088</td><td>201088</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GptOss omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>97</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Experts</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 29 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>2880</td><td>64</td></tr><tr><td>intermediate_size</td><td>2880</td><td>128</td></tr><tr><td>layer_types</td><td>24 entries: 12× sliding_attention, 12× full_attention</td><td>[&quot;sliding_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>num_attention_heads</td><td>64</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>4</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>num_local_experts</td><td>32</td><td>4</td></tr><tr><td>quantization_config</td><td>{&quot;modules_to_not_convert&quot;: [&quot;model.layers.*.self_attn&quot;, &quot;model.layers.*.mlp.r…</td><td>null</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/gpt_oss-mic.json b/shrink/gpt_oss-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..0bf718db6c495d6142f38ad2c396f050e5f86df3
--- /dev/null
+++ b/shrink/gpt_oss-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/gpt_oss-mic@8fea014b7848db31a7032015195dc9b6947e63e2", "findings": []}
\ No newline at end of file
diff --git a/shrink/granite-3.0-1b-a400m-instruct-mic.html b/shrink/granite-3.0-1b-a400m-instruct-mic.html
index dff7523263bfdcf620b8ab1d9c6ffe86d33cb977..98bfc43ac250110f41ee35be611ab62041702d86 100644
--- a/shrink/granite-3.0-1b-a400m-instruct-mic.html
+++ b/shrink/granite-3.0-1b-a400m-instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ibm-granite/granite-3.0-1b-a400m-instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/granite-3.0-1b-a400m-instruct-mic@94e400b500716487cbda1a86a26a33f232758f04 ← ibm-granite/granite-3.0-1b-a400m-instruct@ffec3c35bdfd97a06f0b4cd5fcc92cd9b1584445 · GraniteMoeForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ibm-granite/granite-3.0-1b-a400m-instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/granite-3.0-1b-a400m-instruct-mic/tree/94e400b500716487cbda1a86a26a33f232758f04" title="huggingface-mic/granite-3.0-1b-a400m-instruct-mic@94e400b500716487cbda1a86a26a33f232758f04">huggingface-mic/granite-3.0-1b-a400m-instruct-mic@94e400b50071</a> ← <a href="https://huggingface.co/ibm-granite/granite-3.0-1b-a400m-instruct/tree/ffec3c35bdfd97a06f0b4cd5fcc92cd9b1584445" title="ibm-granite/granite-3.0-1b-a400m-instruct@ffec3c35bdfd97a06f0b4cd5fcc92cd9b1584445">ibm-granite/granite-3.0-1b-a400m-instruct@ffec3c35bdfd</a> · GraniteMoeForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">1.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.962285" title="model.layers: 1.3B (96.2%)"></i><i class="c1" style="flex:0.037714" title="model.embed_tokens: 50.3M (3.8%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.065901" title="model.layers: 222.0K (6.6%)"></i><i class="c1" style="flex:0.934080" title="model.embed_tokens: 3.1M (93.4%)"></i><i class="c2" style="flex:0.000019" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>1.3B</td><td>96.2%</td><td>222.0K</td><td>6.6%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>50.3M</td><td>3.8%</td><td>3.1M</td><td>93.4%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (16 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>MoE routing</td><td>top-8 of 32 experts</td><td>top-2 of 4 experts</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>granitemoe</td><td>granitemoe</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>49155</td><td>49155</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GraniteMoe omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>97</td><td>9</td></tr><tr class=""><td>RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Experts</td><td>24</td><td>2</td></tr><tr class=""><td>MoE</td><td>24</td><td>2</td></tr><tr class=""><td>TopKRouter</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 28 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>512</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_experts_per_tok</td><td>8</td><td>2</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>num_local_experts</td><td>32</td><td>4</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/granite-3.0-1b-a400m-instruct-mic.json b/shrink/granite-3.0-1b-a400m-instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..358012ed18fe128982c6a7aec78a1a17e00ead59
--- /dev/null
+++ b/shrink/granite-3.0-1b-a400m-instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/granite-3.0-1b-a400m-instruct-mic@94e400b500716487cbda1a86a26a33f232758f04", "findings": []}
\ No newline at end of file
diff --git a/shrink/granite-3.1-2b-instruct-mic.html b/shrink/granite-3.1-2b-instruct-mic.html
index 9c85849bd0f880618b8d046fa4f28c906d8fa093..0b5e44ca3ac16565c22d4f67f62c09fcbab76d62 100644
--- a/shrink/granite-3.1-2b-instruct-mic.html
+++ b/shrink/granite-3.1-2b-instruct-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ibm-granite/granite-3.1-2b-instruct: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/granite-3.1-2b-instruct-mic@fa907c0d3e051d2f91f3a595f80bc9b97a72fccf ← ibm-granite/granite-3.1-2b-instruct@bbc2aed595bd38bd770263dc3ab831db9794441d · GraniteForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ibm-granite/granite-3.1-2b-instruct</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/granite-3.1-2b-instruct-mic/tree/fa907c0d3e051d2f91f3a595f80bc9b97a72fccf" title="huggingface-mic/granite-3.1-2b-instruct-mic@fa907c0d3e051d2f91f3a595f80bc9b97a72fccf">huggingface-mic/granite-3.1-2b-instruct-mic@fa907c0d3e05</a> ← <a href="https://huggingface.co/ibm-granite/granite-3.1-2b-instruct/tree/bbc2aed595bd38bd770263dc3ab831db9794441d" title="ibm-granite/granite-3.1-2b-instruct@bbc2aed595bd38bd770263dc3ab831db9794441d">ibm-granite/granite-3.1-2b-instruct@bbc2aed595bd</a> · GraniteForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.5B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i><i class="c0" title="layer 30: DecoderLayer"></i><i class="c0" title="layer 31: DecoderLayer"></i><i class="c0" title="layer 32: DecoderLayer"></i><i class="c0" title="layer 33: DecoderLayer"></i><i class="c0" title="layer 34: DecoderLayer"></i><i class="c0" title="layer 35: DecoderLayer"></i><i class="c0" title="layer 36: DecoderLayer"></i><i class="c0" title="layer 37: DecoderLayer"></i><i class="c0" title="layer 38: DecoderLayer"></i><i class="c0" title="layer 39: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>40</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.960264" title="model.layers: 2.4B (96.0%)"></i><i class="c1" style="flex:0.039735" title="model.embed_tokens: 100.7M (4.0%)"></i><i class="c2" style="flex:0.000001" title="model.norm: 2.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.022977" title="model.layers: 74.0K (2.3%)"></i><i class="c1" style="flex:0.977003" title="model.embed_tokens: 3.1M (97.7%)"></i><i class="c2" style="flex:0.000020" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>2.4B</td><td>96.0%</td><td>74.0K</td><td>2.3%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>100.7M</td><td>4.0%</td><td>3.1M</td><td>97.7%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>2.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (32 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>granite</td><td>granite</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>49155</td><td>49155</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Granite omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>281</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>81</td><td>5</td></tr><tr class=""><td>Attention</td><td>40</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>MLP</td><td>40</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>40</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">5 fields changed, 27 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>2048</td><td>64</td></tr><tr><td>intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>8</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/granite-3.1-2b-instruct-mic.json b/shrink/granite-3.1-2b-instruct-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..150ca17b0116143a8833e971ee31880a8ca10a45
--- /dev/null
+++ b/shrink/granite-3.1-2b-instruct-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/granite-3.1-2b-instruct-mic@fa907c0d3e051d2f91f3a595f80bc9b97a72fccf", "findings": []}
\ No newline at end of file
diff --git a/shrink/granite-4.0-1b-speech-mic.html b/shrink/granite-4.0-1b-speech-mic.html
index bcabce79aade033a92bf61dc60f370e454af710b..86117a6aead88cd1d6ab8a75a631702e30a1ab7d 100644
--- a/shrink/granite-4.0-1b-speech-mic.html
+++ b/shrink/granite-4.0-1b-speech-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ibm-granite/granite-4.0-1b-speech: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/granite-4.0-1b-speech-mic@15329050e1ffa3cdc15c7448cc320dc19106dff7 ← ibm-granite/granite-4.0-1b-speech@bd87ab862416353633ea431fe49b1614003623c5 · GraniteSpeechForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ibm-granite/granite-4.0-1b-speech</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/granite-4.0-1b-speech-mic/tree/15329050e1ffa3cdc15c7448cc320dc19106dff7" title="huggingface-mic/granite-4.0-1b-speech-mic@15329050e1ffa3cdc15c7448cc320dc19106dff7">huggingface-mic/granite-4.0-1b-speech-mic@15329050e1ff</a> ← <a href="https://huggingface.co/ibm-granite/granite-4.0-1b-speech/tree/bd87ab862416353633ea431fe49b1614003623c5" title="ibm-granite/granite-4.0-1b-speech@bd87ab862416353633ea431fe49b1614003623c5">ibm-granite/granite-4.0-1b-speech@bd87ab862416</a> · GraniteSpeechForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">2.3B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">14.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">3</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 16</span><div class="cells"><i class="c0" title="layer 0: ConformerBlock"></i><i class="c0" title="layer 1: ConformerBlock"></i><i class="c0" title="layer 2: ConformerBlock"></i><i class="c0" title="layer 3: ConformerBlock"></i><i class="c0" title="layer 4: ConformerBlock"></i><i class="c0" title="layer 5: ConformerBlock"></i><i class="c0" title="layer 6: ConformerBlock"></i><i class="c0" title="layer 7: ConformerBlock"></i><i class="c0" title="layer 8: ConformerBlock"></i><i class="c0" title="layer 9: ConformerBlock"></i><i class="c0" title="layer 10: ConformerBlock"></i><i class="c0" title="layer 11: ConformerBlock"></i><i class="c0" title="layer 12: ConformerBlock"></i><i class="c0" title="layer 13: ConformerBlock"></i><i class="c0" title="layer 14: ConformerBlock"></i><i class="c0" title="layer 15: ConformerBlock"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: ConformerBlock"></i><i class="c0" title="layer 1: ConformerBlock"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>ConformerBlock</td><td>16</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.projector.qformer.encoder.layer</h3><div class="row"><span class="lab">Original · 2</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Blip2QFormerLayer</td><td>2</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: GraniteDecoderLayer"></i><i class="c0" title="layer 1: GraniteDecoderLayer"></i><i class="c0" title="layer 2: GraniteDecoderLayer"></i><i class="c0" title="layer 3: GraniteDecoderLayer"></i><i class="c0" title="layer 4: GraniteDecoderLayer"></i><i class="c0" title="layer 5: GraniteDecoderLayer"></i><i class="c0" title="layer 6: GraniteDecoderLayer"></i><i class="c0" title="layer 7: GraniteDecoderLayer"></i><i class="c0" title="layer 8: GraniteDecoderLayer"></i><i class="c0" title="layer 9: GraniteDecoderLayer"></i><i class="c0" title="layer 10: GraniteDecoderLayer"></i><i class="c0" title="layer 11: GraniteDecoderLayer"></i><i class="c0" title="layer 12: GraniteDecoderLayer"></i><i class="c0" title="layer 13: GraniteDecoderLayer"></i><i class="c0" title="layer 14: GraniteDecoderLayer"></i><i class="c0" title="layer 15: GraniteDecoderLayer"></i><i class="c0" title="layer 16: GraniteDecoderLayer"></i><i class="c0" title="layer 17: GraniteDecoderLayer"></i><i class="c0" title="layer 18: GraniteDecoderLayer"></i><i class="c0" title="layer 19: GraniteDecoderLayer"></i><i class="c0" title="layer 20: GraniteDecoderLayer"></i><i class="c0" title="layer 21: GraniteDecoderLayer"></i><i class="c0" title="layer 22: GraniteDecoderLayer"></i><i class="c0" title="layer 23: GraniteDecoderLayer"></i><i class="c0" title="layer 24: GraniteDecoderLayer"></i><i class="c0" title="layer 25: GraniteDecoderLayer"></i><i class="c0" title="layer 26: GraniteDecoderLayer"></i><i class="c0" title="layer 27: GraniteDecoderLayer"></i><i class="c0" title="layer 28: GraniteDecoderLayer"></i><i class="c0" title="layer 29: GraniteDecoderLayer"></i><i class="c0" title="layer 30: GraniteDecoderLayer"></i><i class="c0" title="layer 31: GraniteDecoderLayer"></i><i class="c0" title="layer 32: GraniteDecoderLayer"></i><i class="c0" title="layer 33: GraniteDecoderLayer"></i><i class="c0" title="layer 34: GraniteDecoderLayer"></i><i class="c0" title="layer 35: GraniteDecoderLayer"></i><i class="c0" title="layer 36: GraniteDecoderLayer"></i><i class="c0" title="layer 37: GraniteDecoderLayer"></i><i class="c0" title="layer 38: GraniteDecoderLayer"></i><i class="c0" title="layer 39: GraniteDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: GraniteDecoderLayer"></i><i class="c0" title="layer 1: GraniteDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>GraniteDecoderLayer</td><td>40</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.616576" title="model.language_model.layers: 1.4B (61.7%)"></i><i class="c1" style="flex:0.189910" title="model.encoder.layers: 439.3M (19.0%)"></i><i class="c2" style="flex:0.088851" title="model.language_model: 205.5M (8.9%)"></i><i class="c3" style="flex:0.088850" title="lm_head: 205.5M (8.9%)"></i><i class="c4" style="flex:0.014523" title="model.projector.qformer.encoder.layer: 33.6M (1.5%)"></i><i class="c5" style="flex:0.000910" title="model.projector: 2.1M (0.1%)"></i><i class="c6" style="flex:0.000380" title="model.encoder: 878.9K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.005042" title="model.language_model.layers: 74.0K (0.5%)"></i><i class="c1" style="flex:0.103787" title="model.encoder.layers: 1.5M (10.4%)"></i><i class="c2" style="flex:0.437699" title="model.language_model: 6.4M (43.8%)"></i><i class="c3" style="flex:0.437694" title="lm_head: 6.4M (43.8%)"></i><i class="c4" style="flex:0.007964" title="model.projector.qformer.encoder.layer: 116.9K (0.8%)"></i><i class="c5" style="flex:0.000305" title="model.projector: 4.5K (0.0%)"></i><i class="c6" style="flex:0.007508" title="model.encoder: 110.2K (0.8%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>1.4B</td><td>61.7%</td><td>74.0K</td><td>0.5%</td></tr><tr><td><i class="c1"></i>model.encoder.layers</td><td>439.3M</td><td>19.0%</td><td>1.5M</td><td>10.4%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>205.5M</td><td>8.9%</td><td>6.4M</td><td>43.8%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>205.5M</td><td>8.9%</td><td>6.4M</td><td>43.8%</td></tr><tr><td><i class="c4"></i>model.projector.qformer.encoder.layer</td><td>33.6M</td><td>1.5%</td><td>116.9K</td><td>0.8%</td></tr><tr><td><i class="c5"></i>model.projector</td><td>2.1M</td><td>0.1%</td><td>4.5K</td><td>0.0%</td></tr><tr><td><i class="c6"></i>model.encoder</td><td>878.9K</td><td>0.0%</td><td>110.2K</td><td>0.8%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>granite_speech</td><td>granite_speech</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (16 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>granite</td><td>granite</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>100353</td><td>100353</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr><tr class=""><td>encoder_config</td><td>model_type</td><td>granite_speech_encoder</td><td>granite_speech_encoder</td><td>✓ kept</td></tr><tr class=""><td>projector_config</td><td>model_type</td><td>blip_2_qformer</td><td>blip_2_qformer</td><td>✓ kept</td></tr><tr class=""><td>projector_config</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>projector_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>projector_config</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GraniteSpeech omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>417</td><td>53</td></tr><tr class=""><td>LayerNorm</td><td>87</td><td>17</td></tr><tr class=""><td>GraniteRMSNorm</td><td>81</td><td>5</td></tr><tr class=""><td>Dropout</td><td>71</td><td>15</td></tr><tr class=""><td>Conv1d</td><td>48</td><td>6</td></tr><tr class=""><td>SiLU</td><td>48</td><td>6</td></tr><tr class=""><td>GraniteAttention</td><td>40</td><td>2</td></tr><tr class=""><td>GraniteDecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>GraniteMLP</td><td>40</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>40</td><td>2</td></tr><tr class=""><td>ConformerFeedForward</td><td>32</td><td>4</td></tr><tr class=""><td>Embedding</td><td>17</td><td>3</td></tr><tr class=""><td>BatchNorm1d</td><td>16</td><td>2</td></tr><tr class=""><td>GLU</td><td>16</td><td>2</td></tr><tr class=""><td>ConformerAttention</td><td>16</td><td>2</td></tr><tr class=""><td>ConformerBlock</td><td>16</td><td>2</td></tr><tr class=""><td>ConformerConvModule</td><td>16</td><td>2</td></tr><tr class=""><td>ConformerDepthWiseConv1d</td><td>16</td><td>2</td></tr><tr class=""><td>Blip2QFormerAttention</td><td>4</td><td>4</td></tr><tr class=""><td>Blip2QFormerMultiHeadAttention</td><td>4</td><td>4</td></tr><tr class=""><td>Blip2QFormerSelfOutput</td><td>4</td><td>4</td></tr><tr class=""><td>ModuleList</td><td>3</td><td>3</td></tr><tr class=""><td>Blip2QFormerIntermediate</td><td>2</td><td>2</td></tr><tr class=""><td>Blip2QFormerLayer</td><td>2</td><td>2</td></tr><tr class=""><td>Blip2QFormerOutput</td><td>2</td><td>2</td></tr><tr class=""><td>GELUActivation</td><td>2</td><td>2</td></tr><tr class=""><td>Blip2QFormerEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>Blip2QFormerModel</td><td>1</td><td>1</td></tr><tr class=""><td>GraniteModel</td><td>1</td><td>1</td></tr><tr class=""><td>GraniteRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>CTCEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>EncoderProjector</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">12 fields changed, 82 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>encoder_config.hidden_dim</td><td>1024</td><td>128</td></tr><tr><td>encoder_config.num_heads</td><td>8</td><td>4</td></tr><tr><td>encoder_config.num_layers</td><td>16</td><td>2</td></tr><tr><td>projector_config.encoder_hidden_size</td><td>1024</td><td>128</td></tr><tr><td>projector_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>projector_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>projector_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.hidden_size</td><td>2048</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>4</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/granite-4.0-1b-speech-mic.json b/shrink/granite-4.0-1b-speech-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..9810b8687cbd27166d3a70d4afee4e371192373f
--- /dev/null
+++ b/shrink/granite-4.0-1b-speech-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/granite-4.0-1b-speech-mic@15329050e1ffa3cdc15c7448cc320dc19106dff7", "findings": []}
\ No newline at end of file
diff --git a/shrink/granite-4.0-h-350m-mic.html b/shrink/granite-4.0-h-350m-mic.html
index 41c6bdece42772da3827d55a35af75582e549044..9450fc4f3fe9ee6ed51d98ed904510c859af35b6 100644
--- a/shrink/granite-4.0-h-350m-mic.html
+++ b/shrink/granite-4.0-h-350m-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ibm-granite/granite-4.0-h-350m: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/granite-4.0-h-350m-mic@83dded3d062dd73ed2432e033e86c43d784a5b92 ← ibm-granite/granite-4.0-h-350m@3b17b717b8f2f5d305b0a92c1491e239aeda19c8 · GraniteMoeHybridForCausalLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ibm-granite/granite-4.0-h-350m</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/granite-4.0-h-350m-mic/tree/83dded3d062dd73ed2432e033e86c43d784a5b92" title="huggingface-mic/granite-4.0-h-350m-mic@83dded3d062dd73ed2432e033e86c43d784a5b92">huggingface-mic/granite-4.0-h-350m-mic@83dded3d062d</a> ← <a href="https://huggingface.co/ibm-granite/granite-4.0-h-350m/tree/3b17b717b8f2f5d305b0a92c1491e239aeda19c8" title="ibm-granite/granite-4.0-h-350m@3b17b717b8f2f5d305b0a92c1491e239aeda19c8">ibm-granite/granite-4.0-h-350m@3b17b717b8f2</a> · GraniteMoeHybridForCausalLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">340.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">6.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 1: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 2: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 3: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 4: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 5: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 6: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 7: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 8: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 9: linear_attention, Conv1d, MambaLayer …"></i><i class="c1" title="layer 10: is_causal=True, Attention"></i><i class="c0" title="layer 11: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 12: linear_attention, Conv1d, MambaLayer …"></i><i class="c1" title="layer 13: is_causal=True, Attention"></i><i class="c0" title="layer 14: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 15: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 16: linear_attention, Conv1d, MambaLayer …"></i><i class="c1" title="layer 17: is_causal=True, Attention"></i><i class="c0" title="layer 18: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 19: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 20: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 21: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 22: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 23: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 24: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 25: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 26: linear_attention, Conv1d, MambaLayer …"></i><i class="c1" title="layer 27: is_causal=True, Attention"></i><i class="c0" title="layer 28: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 29: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 30: linear_attention, Conv1d, MambaLayer …"></i><i class="c0" title="layer 31: linear_attention, Conv1d, MambaLayer …"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: linear_attention, Conv1d, MambaLayer …"></i><i class="c1" title="layer 1: is_causal=True, Attention"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>linear_attention, Conv1d, MambaLayer …</td><td>28</td><td>1</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>is_causal=True, Attention</td><td>4</td><td>1</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.773542" title="model.layers: 263.3M (77.4%)"></i><i class="c1" style="flex:0.226456" title="model.embed_tokens: 77.1M (22.6%)"></i><i class="c2" style="flex:0.000002" title="model.norm: 768 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.013787" title="model.layers: 89.8K (1.4%)"></i><i class="c1" style="flex:0.986203" title="model.embed_tokens: 6.4M (98.6%)"></i><i class="c2" style="flex:0.000010" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>263.3M</td><td>77.4%</td><td>89.8K</td><td>1.4%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>77.1M</td><td>22.6%</td><td>6.4M</td><td>98.6%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>768</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (12 heads, 4 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>granitemoehybrid</td><td>granitemoehybrid</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GraniteMoeHybrid omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>137</td><td>11</td></tr><tr class=""><td>RMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>SiLUActivation</td><td>60</td><td>3</td></tr><tr class=""><td>DecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>MLP</td><td>32</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>28</td><td>1</td></tr><tr class=""><td>MambaLayer</td><td>28</td><td>1</td></tr><tr class=""><td>RMSNormGated</td><td>28</td><td>1</td></tr><tr class=""><td>Attention</td><td>4</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 39 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>2048</td><td>128</td></tr><tr><td>layer_types</td><td>32 entries: 28× linear_attention, 4× full_attention</td><td>[&quot;linear_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>mamba_chunk_size</td><td>256</td><td>8</td></tr><tr><td>mamba_d_head</td><td>32</td><td>16</td></tr><tr><td>mamba_d_state</td><td>128</td><td>16</td></tr><tr><td>mamba_n_heads</td><td>48</td><td>8</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>4</td><td>2</td></tr><tr><td>shared_intermediate_size</td><td>2048</td><td>128</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/granite-4.0-h-350m-mic.json b/shrink/granite-4.0-h-350m-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..206de292c2f010978905249f091039f9255d9a68
--- /dev/null
+++ b/shrink/granite-4.0-h-350m-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/granite-4.0-h-350m-mic@83dded3d062dd73ed2432e033e86c43d784a5b92", "findings": []}
\ No newline at end of file
diff --git a/shrink/granite-docling-258M-mic.html b/shrink/granite-docling-258M-mic.html
index 580e5b01d8744daf3e50a4a982c28f717f09f151..d81e63247821820f5e23e05c204cf2da7ad1b900 100644
--- a/shrink/granite-docling-258M-mic.html
+++ b/shrink/granite-docling-258M-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ibm-granite/granite-docling-258M: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/granite-docling-258M-mic@100175103449d7ed8e24b4d0cbd79e15671b781c ← ibm-granite/granite-docling-258M@982fe3b40f2fa73c365bdb1bcacf6c81b7184bfe · Idefics3ForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ibm-granite/granite-docling-258M</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/granite-docling-258M-mic/tree/100175103449d7ed8e24b4d0cbd79e15671b781c" title="huggingface-mic/granite-docling-258M-mic@100175103449d7ed8e24b4d0cbd79e15671b781c">huggingface-mic/granite-docling-258M-mic@100175103449</a> ← <a href="https://huggingface.co/ibm-granite/granite-docling-258M/tree/982fe3b40f2fa73c365bdb1bcacf6c81b7184bfe" title="ibm-granite/granite-docling-258M@982fe3b40f2fa73c365bdb1bcacf6c81b7184bfe">ibm-granite/granite-docling-258M@982fe3b40f2f</a> · Idefics3ForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">257.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">6.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_model.encoder.layers</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i><i class="c0" title="layer 2: EncoderLayer"></i><i class="c0" title="layer 3: EncoderLayer"></i><i class="c0" title="layer 4: EncoderLayer"></i><i class="c0" title="layer 5: EncoderLayer"></i><i class="c0" title="layer 6: EncoderLayer"></i><i class="c0" title="layer 7: EncoderLayer"></i><i class="c0" title="layer 8: EncoderLayer"></i><i class="c0" title="layer 9: EncoderLayer"></i><i class="c0" title="layer 10: EncoderLayer"></i><i class="c0" title="layer 11: EncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: EncoderLayer"></i><i class="c0" title="layer 1: EncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>EncoderLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.text_model.layers</h3><div class="row"><span class="lab">Original · 30</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i><i class="c0" title="layer 2: LlamaDecoderLayer"></i><i class="c0" title="layer 3: LlamaDecoderLayer"></i><i class="c0" title="layer 4: LlamaDecoderLayer"></i><i class="c0" title="layer 5: LlamaDecoderLayer"></i><i class="c0" title="layer 6: LlamaDecoderLayer"></i><i class="c0" title="layer 7: LlamaDecoderLayer"></i><i class="c0" title="layer 8: LlamaDecoderLayer"></i><i class="c0" title="layer 9: LlamaDecoderLayer"></i><i class="c0" title="layer 10: LlamaDecoderLayer"></i><i class="c0" title="layer 11: LlamaDecoderLayer"></i><i class="c0" title="layer 12: LlamaDecoderLayer"></i><i class="c0" title="layer 13: LlamaDecoderLayer"></i><i class="c0" title="layer 14: LlamaDecoderLayer"></i><i class="c0" title="layer 15: LlamaDecoderLayer"></i><i class="c0" title="layer 16: LlamaDecoderLayer"></i><i class="c0" title="layer 17: LlamaDecoderLayer"></i><i class="c0" title="layer 18: LlamaDecoderLayer"></i><i class="c0" title="layer 19: LlamaDecoderLayer"></i><i class="c0" title="layer 20: LlamaDecoderLayer"></i><i class="c0" title="layer 21: LlamaDecoderLayer"></i><i class="c0" title="layer 22: LlamaDecoderLayer"></i><i class="c0" title="layer 23: LlamaDecoderLayer"></i><i class="c0" title="layer 24: LlamaDecoderLayer"></i><i class="c0" title="layer 25: LlamaDecoderLayer"></i><i class="c0" title="layer 26: LlamaDecoderLayer"></i><i class="c0" title="layer 27: LlamaDecoderLayer"></i><i class="c0" title="layer 28: LlamaDecoderLayer"></i><i class="c0" title="layer 29: LlamaDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>LlamaDecoderLayer</td><td>30</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.412411" title="model.text_model.layers: 106.2M (41.2%)"></i><i class="c1" style="flex:0.330287" title="model.vision_model.encoder.layers: 85.1M (33.0%)"></i><i class="c2" style="flex:0.224464" title="model.text_model: 57.8M (22.4%)"></i><i class="c3" style="flex:0.027485" title="model.connector: 7.1M (2.7%)"></i><i class="c4" style="flex:0.005353" title="model.vision_model: 1.4M (0.5%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.010970" title="model.text_model.layers: 74.0K (1.1%)"></i><i class="c1" style="flex:0.009927" title="model.vision_model.encoder.layers: 66.9K (1.0%)"></i><i class="c2" style="flex:0.952351" title="model.text_model: 6.4M (95.2%)"></i><i class="c3" style="flex:0.009718" title="model.connector: 65.5K (1.0%)"></i><i class="c4" style="flex:0.017035" title="model.vision_model: 114.9K (1.7%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.text_model.layers</td><td>106.2M</td><td>41.2%</td><td>74.0K</td><td>1.1%</td></tr><tr><td><i class="c1"></i>model.vision_model.encoder.layers</td><td>85.1M</td><td>33.0%</td><td>66.9K</td><td>1.0%</td></tr><tr><td><i class="c2"></i>model.text_model</td><td>57.8M</td><td>22.4%</td><td>6.4M</td><td>95.2%</td></tr><tr><td><i class="c3"></i>model.connector</td><td>7.1M</td><td>2.7%</td><td>65.5K</td><td>1.0%</td></tr><tr><td><i class="c4"></i>model.vision_model</td><td>1.4M</td><td>0.5%</td><td>114.9K</td><td>1.7%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>idefics3</td><td>idefics3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>idefics3_vision</td><td>idefics3_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>512</td><td>512</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (9 heads, 3 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>llama</td><td>llama</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Idefics3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>284</td><td>28</td></tr><tr class=""><td>LlamaRMSNorm</td><td>61</td><td>5</td></tr><tr class=""><td>LlamaAttention</td><td>30</td><td>2</td></tr><tr class=""><td>LlamaDecoderLayer</td><td>30</td><td>2</td></tr><tr class=""><td>LlamaMLP</td><td>30</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>30</td><td>2</td></tr><tr class=""><td>LayerNorm</td><td>25</td><td>5</td></tr><tr class=""><td>GELUTanh</td><td>12</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>12</td><td>2</td></tr><tr class=""><td>VisionAttention</td><td>12</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Connector</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SimpleMLP</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionTransformer</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaModel</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaRotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">10 fields changed, 59 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>64</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>576</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>1536</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>9</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>30</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>3</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/granite-docling-258M-mic.json b/shrink/granite-docling-258M-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..da82dd2431429d59acb02f834c5fba0bbdadd3b2
--- /dev/null
+++ b/shrink/granite-docling-258M-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/granite-docling-258M-mic@100175103449d7ed8e24b4d0cbd79e15671b781c", "findings": []}
\ No newline at end of file
diff --git a/shrink/granite-vision-4.1-4b-mic.html b/shrink/granite-vision-4.1-4b-mic.html
index b8c0dabce39859153d47819d3c29ca64957ac993..be89419507577a46c44c183635dd793abc6b422c 100644
--- a/shrink/granite-vision-4.1-4b-mic.html
+++ b/shrink/granite-vision-4.1-4b-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>ibm-granite/granite-vision-4.1-4b: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/granite-vision-4.1-4b-mic@a6d32301c1b54aa5ef889d46d64998c1fb01c7df ← ibm-granite/granite-vision-4.1-4b@37d591f06319e8f1638b5adcf58bdf50e0f84f7a · Granite4VisionForConditionalGeneration · random weights</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">7.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">4</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i><i class="c0" title="layer 2: SiglipEncoderLayer"></i><i class="c0" title="layer 3: SiglipEncoderLayer"></i><i class="c0" title="layer 4: SiglipEncoderLayer"></i><i class="c0" title="layer 5: SiglipEncoderLayer"></i><i class="c0" title="layer 6: SiglipEncoderLayer"></i><i class="c0" title="layer 7: SiglipEncoderLayer"></i><i class="c0" title="layer 8: SiglipEncoderLayer"></i><i class="c0" title="layer 9: SiglipEncoderLayer"></i><i class="c0" title="layer 10: SiglipEncoderLayer"></i><i class="c0" title="layer 11: SiglipEncoderLayer"></i><i class="c0" title="layer 12: SiglipEncoderLayer"></i><i class="c0" title="layer 13: SiglipEncoderLayer"></i><i class="c0" title="layer 14: SiglipEncoderLayer"></i><i class="c0" title="layer 15: SiglipEncoderLayer"></i><i class="c0" title="layer 16: SiglipEncoderLayer"></i><i class="c0" title="layer 17: SiglipEncoderLayer"></i><i class="c0" title="layer 18: SiglipEncoderLayer"></i><i class="c0" title="layer 19: SiglipEncoderLayer"></i><i class="c0" title="layer 20: SiglipEncoderLayer"></i><i class="c0" title="layer 21: SiglipEncoderLayer"></i><i class="c0" title="layer 22: SiglipEncoderLayer"></i><i class="c0" title="layer 23: SiglipEncoderLayer"></i><i class="c0" title="layer 24: SiglipEncoderLayer"></i><i class="c0" title="layer 25: SiglipEncoderLayer"></i><i class="c0" title="layer 26: SiglipEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>SiglipEncoderLayer</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i><i class="c0" title="layer 2: TextDecoderLayer"></i><i class="c0" title="layer 3: TextDecoderLayer"></i><i class="c0" title="layer 4: TextDecoderLayer"></i><i class="c0" title="layer 5: TextDecoderLayer"></i><i class="c0" title="layer 6: TextDecoderLayer"></i><i class="c0" title="layer 7: TextDecoderLayer"></i><i class="c0" title="layer 8: TextDecoderLayer"></i><i class="c0" title="layer 9: TextDecoderLayer"></i><i class="c0" title="layer 10: TextDecoderLayer"></i><i class="c0" title="layer 11: TextDecoderLayer"></i><i class="c0" title="layer 12: TextDecoderLayer"></i><i class="c0" title="layer 13: TextDecoderLayer"></i><i class="c0" title="layer 14: TextDecoderLayer"></i><i class="c0" title="layer 15: TextDecoderLayer"></i><i class="c0" title="layer 16: TextDecoderLayer"></i><i class="c0" title="layer 17: TextDecoderLayer"></i><i class="c0" title="layer 18: TextDecoderLayer"></i><i class="c0" title="layer 19: TextDecoderLayer"></i><i class="c0" title="layer 20: TextDecoderLayer"></i><i class="c0" title="layer 21: TextDecoderLayer"></i><i class="c0" title="layer 22: TextDecoderLayer"></i><i class="c0" title="layer 23: TextDecoderLayer"></i><i class="c0" title="layer 24: TextDecoderLayer"></i><i class="c0" title="layer 25: TextDecoderLayer"></i><i class="c0" title="layer 26: TextDecoderLayer"></i><i class="c0" title="layer 27: TextDecoderLayer"></i><i class="c0" title="layer 28: TextDecoderLayer"></i><i class="c0" title="layer 29: TextDecoderLayer"></i><i class="c0" title="layer 30: TextDecoderLayer"></i><i class="c0" title="layer 31: TextDecoderLayer"></i><i class="c0" title="layer 32: TextDecoderLayer"></i><i class="c0" title="layer 33: TextDecoderLayer"></i><i class="c0" title="layer 34: TextDecoderLayer"></i><i class="c0" title="layer 35: TextDecoderLayer"></i><i class="c0" title="layer 36: TextDecoderLayer"></i><i class="c0" title="layer 37: TextDecoderLayer"></i><i class="c0" title="layer 38: TextDecoderLayer"></i><i class="c0" title="layer 39: TextDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextDecoderLayer</td><td>40</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.layerwise_projectors</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Blip2QFormerLayer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.spatial_projectors</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Blip2QFormerLayer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.787033" title="model.language_model.layers: 3.1B (78.7%)"></i><i class="c1" style="flex:0.102939" title="model.vision_tower.encoder.layers: 411.5M (10.3%)"></i><i class="c2" style="flex:0.064271" title="model.language_model: 256.9M (6.4%)"></i><i class="c3" style="flex:0.020778" title="model.spatial_projectors: 83.1M (2.1%)"></i><i class="c4" style="flex:0.020778" title="model.layerwise_projectors: 83.1M (2.1%)"></i><i class="c5" style="flex:0.004200" title="model.vision_tower: 16.8M (0.4%)"></i><i class="c6" style="flex:0.000001" title="model: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.010331" title="model.language_model.layers: 74.0K (1.0%)"></i><i class="c1" style="flex:0.009348" title="model.vision_tower.encoder.layers: 66.9K (0.9%)"></i><i class="c2" style="flex:0.896834" title="model.language_model: 6.4M (89.7%)"></i><i class="c3" style="flex:0.033388" title="model.spatial_projectors: 239.1K (3.3%)"></i><i class="c4" style="flex:0.033388" title="model.layerwise_projectors: 239.1K (3.3%)"></i><i class="c5" style="flex:0.016703" title="model.vision_tower: 119.6K (1.7%)"></i><i class="c6" style="flex:0.000009" title="model: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>3.1B</td><td>78.7%</td><td>74.0K</td><td>1.0%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layers</td><td>411.5M</td><td>10.3%</td><td>66.9K</td><td>0.9%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>256.9M</td><td>6.4%</td><td>6.4M</td><td>89.7%</td></tr><tr><td><i class="c3"></i>model.spatial_projectors</td><td>83.1M</td><td>2.1%</td><td>239.1K</td><td>3.3%</td></tr><tr><td><i class="c4"></i>model.layerwise_projectors</td><td>83.1M</td><td>2.1%</td><td>239.1K</td><td>3.3%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>16.8M</td><td>0.4%</td><td>119.6K</td><td>1.7%</td></tr><tr><td><i class="c6"></i>model</td><td>2.6K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>granite4_vision</td><td>granite4_vision</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>siglip_vision_model</td><td>siglip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>384</td><td>384</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (40 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>granite</td><td>granite</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>model_type</td><td>blip_2_qformer</td><td>blip_2_qformer</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Granite4Vision omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>533</td><td>117</td></tr><tr class=""><td>LayerNorm</td><td>96</td><td>46</td></tr><tr class=""><td>TextRMSNorm</td><td>81</td><td>5</td></tr><tr class=""><td>Dropout</td><td>40</td><td>40</td></tr><tr class=""><td>TextAttention</td><td>40</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>TextMLP</td><td>40</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>40</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>28</td><td>3</td></tr><tr class=""><td>SiglipMLP</td><td>28</td><td>3</td></tr><tr class=""><td>SiglipAttention</td><td>27</td><td>2</td></tr><tr class=""><td>SiglipEncoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>Blip2QFormerAttention</td><td>16</td><td>16</td></tr><tr class=""><td>Blip2QFormerMultiHeadAttention</td><td>16</td><td>16</td></tr><tr class=""><td>Blip2QFormerSelfOutput</td><td>16</td><td>16</td></tr><tr class=""><td>ModuleList</td><td>12</td><td>12</td></tr><tr class=""><td>Blip2QFormerEncoder</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerIntermediate</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerLayer</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerModel</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerOutput</td><td>8</td><td>8</td></tr><tr class=""><td>GELUActivation</td><td>8</td><td>8</td></tr><tr class=""><td>WindowQFormerDownsampler</td><td>8</td><td>8</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>MultiheadAttention</td><td>1</td><td>1</td></tr><tr class=""><td>NonDynamicallyQuantizableLinear</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipMultiheadAttentionPoolingHead</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">15 fields changed, 86 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>deepstack_layer_map</td><td>[[-19, 9], [-13, 6], [-7, 3], [-1, 0]]</td><td>[[-3, 0], [-2, 0], [-2, 0], [-1, 0]]</td></tr><tr><td>qformer_config.encoder_hidden_size</td><td>1152</td><td>64</td></tr><tr><td>qformer_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>qformer_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>qformer_config.num_attention_heads</td><td>18</td><td>1</td></tr><tr><td>spatial_target_layers</td><td>[12, 15, 18, 21]</td><td>[0, 1]</td></tr><tr><td>text_config.hidden_size</td><td>2560</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>40</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>27</td><td>2</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>ibm-granite/granite-vision-4.1-4b</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/granite-vision-4.1-4b-mic/tree/a6d32301c1b54aa5ef889d46d64998c1fb01c7df" title="huggingface-mic/granite-vision-4.1-4b-mic@a6d32301c1b54aa5ef889d46d64998c1fb01c7df">huggingface-mic/granite-vision-4.1-4b-mic@a6d32301c1b5</a> ← <a href="https://huggingface.co/ibm-granite/granite-vision-4.1-4b/tree/37d591f06319e8f1638b5adcf58bdf50e0f84f7a" title="ibm-granite/granite-vision-4.1-4b@37d591f06319e8f1638b5adcf58bdf50e0f84f7a">ibm-granite/granite-vision-4.1-4b@37d591f06319</a> · Granite4VisionForConditionalGeneration · random weights</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">4.0B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">7.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">4</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 27</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i><i class="c0" title="layer 2: SiglipEncoderLayer"></i><i class="c0" title="layer 3: SiglipEncoderLayer"></i><i class="c0" title="layer 4: SiglipEncoderLayer"></i><i class="c0" title="layer 5: SiglipEncoderLayer"></i><i class="c0" title="layer 6: SiglipEncoderLayer"></i><i class="c0" title="layer 7: SiglipEncoderLayer"></i><i class="c0" title="layer 8: SiglipEncoderLayer"></i><i class="c0" title="layer 9: SiglipEncoderLayer"></i><i class="c0" title="layer 10: SiglipEncoderLayer"></i><i class="c0" title="layer 11: SiglipEncoderLayer"></i><i class="c0" title="layer 12: SiglipEncoderLayer"></i><i class="c0" title="layer 13: SiglipEncoderLayer"></i><i class="c0" title="layer 14: SiglipEncoderLayer"></i><i class="c0" title="layer 15: SiglipEncoderLayer"></i><i class="c0" title="layer 16: SiglipEncoderLayer"></i><i class="c0" title="layer 17: SiglipEncoderLayer"></i><i class="c0" title="layer 18: SiglipEncoderLayer"></i><i class="c0" title="layer 19: SiglipEncoderLayer"></i><i class="c0" title="layer 20: SiglipEncoderLayer"></i><i class="c0" title="layer 21: SiglipEncoderLayer"></i><i class="c0" title="layer 22: SiglipEncoderLayer"></i><i class="c0" title="layer 23: SiglipEncoderLayer"></i><i class="c0" title="layer 24: SiglipEncoderLayer"></i><i class="c0" title="layer 25: SiglipEncoderLayer"></i><i class="c0" title="layer 26: SiglipEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>SiglipEncoderLayer</td><td>27</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 40</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i><i class="c0" title="layer 2: TextDecoderLayer"></i><i class="c0" title="layer 3: TextDecoderLayer"></i><i class="c0" title="layer 4: TextDecoderLayer"></i><i class="c0" title="layer 5: TextDecoderLayer"></i><i class="c0" title="layer 6: TextDecoderLayer"></i><i class="c0" title="layer 7: TextDecoderLayer"></i><i class="c0" title="layer 8: TextDecoderLayer"></i><i class="c0" title="layer 9: TextDecoderLayer"></i><i class="c0" title="layer 10: TextDecoderLayer"></i><i class="c0" title="layer 11: TextDecoderLayer"></i><i class="c0" title="layer 12: TextDecoderLayer"></i><i class="c0" title="layer 13: TextDecoderLayer"></i><i class="c0" title="layer 14: TextDecoderLayer"></i><i class="c0" title="layer 15: TextDecoderLayer"></i><i class="c0" title="layer 16: TextDecoderLayer"></i><i class="c0" title="layer 17: TextDecoderLayer"></i><i class="c0" title="layer 18: TextDecoderLayer"></i><i class="c0" title="layer 19: TextDecoderLayer"></i><i class="c0" title="layer 20: TextDecoderLayer"></i><i class="c0" title="layer 21: TextDecoderLayer"></i><i class="c0" title="layer 22: TextDecoderLayer"></i><i class="c0" title="layer 23: TextDecoderLayer"></i><i class="c0" title="layer 24: TextDecoderLayer"></i><i class="c0" title="layer 25: TextDecoderLayer"></i><i class="c0" title="layer 26: TextDecoderLayer"></i><i class="c0" title="layer 27: TextDecoderLayer"></i><i class="c0" title="layer 28: TextDecoderLayer"></i><i class="c0" title="layer 29: TextDecoderLayer"></i><i class="c0" title="layer 30: TextDecoderLayer"></i><i class="c0" title="layer 31: TextDecoderLayer"></i><i class="c0" title="layer 32: TextDecoderLayer"></i><i class="c0" title="layer 33: TextDecoderLayer"></i><i class="c0" title="layer 34: TextDecoderLayer"></i><i class="c0" title="layer 35: TextDecoderLayer"></i><i class="c0" title="layer 36: TextDecoderLayer"></i><i class="c0" title="layer 37: TextDecoderLayer"></i><i class="c0" title="layer 38: TextDecoderLayer"></i><i class="c0" title="layer 39: TextDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: TextDecoderLayer"></i><i class="c0" title="layer 1: TextDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>TextDecoderLayer</td><td>40</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.layerwise_projectors</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Blip2QFormerLayer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.spatial_projectors</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Blip2QFormerLayer"></i><i class="c0" title="layer 1: Blip2QFormerLayer"></i><i class="c0" title="layer 2: Blip2QFormerLayer"></i><i class="c0" title="layer 3: Blip2QFormerLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Blip2QFormerLayer</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.787033" title="model.language_model.layers: 3.1B (78.7%)"></i><i class="c1" style="flex:0.102939" title="model.vision_tower.encoder.layers: 411.5M (10.3%)"></i><i class="c2" style="flex:0.064271" title="model.language_model: 256.9M (6.4%)"></i><i class="c3" style="flex:0.020778" title="model.layerwise_projectors: 83.1M (2.1%)"></i><i class="c4" style="flex:0.020778" title="model.spatial_projectors: 83.1M (2.1%)"></i><i class="c5" style="flex:0.004200" title="model.vision_tower: 16.8M (0.4%)"></i><i class="c6" style="flex:0.000001" title="model: 2.6K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.010331" title="model.language_model.layers: 74.0K (1.0%)"></i><i class="c1" style="flex:0.009348" title="model.vision_tower.encoder.layers: 66.9K (0.9%)"></i><i class="c2" style="flex:0.896834" title="model.language_model: 6.4M (89.7%)"></i><i class="c3" style="flex:0.033388" title="model.layerwise_projectors: 239.1K (3.3%)"></i><i class="c4" style="flex:0.033388" title="model.spatial_projectors: 239.1K (3.3%)"></i><i class="c5" style="flex:0.016703" title="model.vision_tower: 119.6K (1.7%)"></i><i class="c6" style="flex:0.000009" title="model: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>3.1B</td><td>78.7%</td><td>74.0K</td><td>1.0%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layers</td><td>411.5M</td><td>10.3%</td><td>66.9K</td><td>0.9%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>256.9M</td><td>6.4%</td><td>6.4M</td><td>89.7%</td></tr><tr><td><i class="c3"></i>model.layerwise_projectors</td><td>83.1M</td><td>2.1%</td><td>239.1K</td><td>3.3%</td></tr><tr><td><i class="c4"></i>model.spatial_projectors</td><td>83.1M</td><td>2.1%</td><td>239.1K</td><td>3.3%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>16.8M</td><td>0.4%</td><td>119.6K</td><td>1.7%</td></tr><tr><td><i class="c6"></i>model</td><td>2.6K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>granite4_vision</td><td>granite4_vision</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>siglip_vision_model</td><td>siglip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>384</td><td>384</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>16</td><td>16</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (40 heads, 8 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>granite</td><td>granite</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>100352</td><td>100352</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>131072</td><td>131072</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>model_type</td><td>blip_2_qformer</td><td>blip_2_qformer</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>qformer_config</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Granite4Vision omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>533</td><td>117</td></tr><tr class=""><td>LayerNorm</td><td>96</td><td>46</td></tr><tr class=""><td>TextRMSNorm</td><td>81</td><td>5</td></tr><tr class=""><td>Dropout</td><td>40</td><td>40</td></tr><tr class=""><td>TextAttention</td><td>40</td><td>2</td></tr><tr class=""><td>TextDecoderLayer</td><td>40</td><td>2</td></tr><tr class=""><td>TextMLP</td><td>40</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>40</td><td>2</td></tr><tr class=""><td>GELUTanh</td><td>28</td><td>3</td></tr><tr class=""><td>SiglipMLP</td><td>28</td><td>3</td></tr><tr class=""><td>SiglipAttention</td><td>27</td><td>2</td></tr><tr class=""><td>SiglipEncoderLayer</td><td>27</td><td>2</td></tr><tr class=""><td>Blip2QFormerAttention</td><td>16</td><td>16</td></tr><tr class=""><td>Blip2QFormerMultiHeadAttention</td><td>16</td><td>16</td></tr><tr class=""><td>Blip2QFormerSelfOutput</td><td>16</td><td>16</td></tr><tr class=""><td>ModuleList</td><td>12</td><td>12</td></tr><tr class=""><td>Blip2QFormerEncoder</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerIntermediate</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerLayer</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerModel</td><td>8</td><td>8</td></tr><tr class=""><td>Blip2QFormerOutput</td><td>8</td><td>8</td></tr><tr class=""><td>GELUActivation</td><td>8</td><td>8</td></tr><tr class=""><td>WindowQFormerDownsampler</td><td>8</td><td>8</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextModel</td><td>1</td><td>1</td></tr><tr class=""><td>TextRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>MultiheadAttention</td><td>1</td><td>1</td></tr><tr class=""><td>NonDynamicallyQuantizableLinear</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipMultiheadAttentionPoolingHead</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">15 fields changed, 86 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>deepstack_layer_map</td><td>[[-19, 9], [-13, 6], [-7, 3], [-1, 0]]</td><td>[[-3, 0], [-2, 0], [-2, 0], [-1, 0]]</td></tr><tr><td>qformer_config.encoder_hidden_size</td><td>1152</td><td>64</td></tr><tr><td>qformer_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>qformer_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>qformer_config.num_attention_heads</td><td>18</td><td>1</td></tr><tr><td>spatial_target_layers</td><td>[12, 15, 18, 21]</td><td>[0, 1]</td></tr><tr><td>text_config.hidden_size</td><td>2560</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>8192</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>40</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>40</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>8</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>27</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/granite-vision-4.1-4b-mic.json b/shrink/granite-vision-4.1-4b-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..a7c0641f7183f310649a568abecc5acf9d8e3698
--- /dev/null
+++ b/shrink/granite-vision-4.1-4b-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/granite-vision-4.1-4b-mic@a6d32301c1b54aa5ef889d46d64998c1fb01c7df", "findings": []}
\ No newline at end of file
diff --git a/shrink/grounding-dino-tiny-mic.html b/shrink/grounding-dino-tiny-mic.html
index ec55da07e3f7ceaecfc57368f75bdd7863e314f6..075a841a04d5f3654d3a9b50c9e39e321e2c8308 100644
--- a/shrink/grounding-dino-tiny-mic.html
+++ b/shrink/grounding-dino-tiny-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>IDEA-Research/grounding-dino-tiny: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/grounding-dino-tiny-mic@fb35353454540cd677db5cdcf3df93a1b60cca4a ← IDEA-Research/grounding-dino-tiny@a2bb814dd30d776dcf7e30523b00659f4f141c71 · GroundingDinoForObjectDetection · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>IDEA-Research/grounding-dino-tiny</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/grounding-dino-tiny-mic/tree/fb35353454540cd677db5cdcf3df93a1b60cca4a" title="huggingface-mic/grounding-dino-tiny-mic@fb35353454540cd677db5cdcf3df93a1b60cca4a">huggingface-mic/grounding-dino-tiny-mic@fb3535345454</a> ← <a href="https://huggingface.co/IDEA-Research/grounding-dino-tiny/tree/a2bb814dd30d776dcf7e30523b00659f4f141c71" title="IDEA-Research/grounding-dino-tiny@a2bb814dd30d776dcf7e30523b00659f4f141c71">IDEA-Research/grounding-dino-tiny@a2bb814dd30d</a> · GroundingDinoForObjectDetection · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">172.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">7.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">7</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.backbone.conv_encoder.model.swin.encoder.layers</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: SwinPatchMerging"></i><i class="c0" title="layer 1: SwinPatchMerging"></i><i class="c0" title="layer 2: SwinPatchMerging"></i><i class="c1" title="layer 3: no SwinPatchMerging"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: SwinPatchMerging"></i><i class="c0" title="layer 1: SwinPatchMerging"></i><i class="c0" title="layer 2: SwinPatchMerging"></i><i class="c1" title="layer 3: no SwinPatchMerging"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>SwinPatchMerging</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr><td><i class="c1"></i>no SwinPatchMerging</td><td>1</td><td>1</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.input_proj_vision</h3><div class="row"><span class="lab">Original · 4</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i><i class="c0" title="layer 3: Conv2d"></i></div></div><div class="row"><span class="lab">Fixture · 4</span><div class="cells"><i class="c0" title="layer 0: Conv2d"></i><i class="c0" title="layer 1: Conv2d"></i><i class="c0" title="layer 2: Conv2d"></i><i class="c0" title="layer 3: Conv2d"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Conv2d</td><td>4</td><td>4</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.text_backbone.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i><i class="c0" title="layer 2: BertLayer"></i><i class="c0" title="layer 3: BertLayer"></i><i class="c0" title="layer 4: BertLayer"></i><i class="c0" title="layer 5: BertLayer"></i><i class="c0" title="layer 6: BertLayer"></i><i class="c0" title="layer 7: BertLayer"></i><i class="c0" title="layer 8: BertLayer"></i><i class="c0" title="layer 9: BertLayer"></i><i class="c0" title="layer 10: BertLayer"></i><i class="c0" title="layer 11: BertLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: BertLayer"></i><i class="c0" title="layer 1: BertLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>BertLayer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.encoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: FusionLayer"></i><i class="c0" title="layer 1: FusionLayer"></i><i class="c0" title="layer 2: FusionLayer"></i><i class="c0" title="layer 3: FusionLayer"></i><i class="c0" title="layer 4: FusionLayer"></i><i class="c0" title="layer 5: FusionLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: FusionLayer"></i><i class="c0" title="layer 1: FusionLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>FusionLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.layers</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.decoder.bbox_embed</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>bbox_embed</h3><div class="row"><span class="lab">Original · 6</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>6</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.493783" title="model.text_backbone.encoder.layer: 85.1M (49.4%)"></i><i class="c1" style="flex:0.159726" title="model.backbone.conv_encoder.model.swin.encoder.layers: 27.5M (16.0%)"></i><i class="c2" style="flex:0.138387" title="model.text_backbone: 23.8M (13.8%)"></i><i class="c3" style="flex:0.127173" title="model.encoder.layers: 21.9M (12.7%)"></i><i class="c4" style="flex:0.063032" title="model.decoder.layers: 10.9M (6.3%)"></i><i class="c5" style="flex:0.012288" title="model.input_proj_vision: 2.1M (1.2%)"></i><i class="c6" style="flex:0.001338" title="model.query_position_embeddings: 230.4K (0.1%)"></i><i class="c7" style="flex:0.004274" title="8 others: 736.2K (0.4%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.008678" title="model.text_backbone.encoder.layer: 66.9K (0.9%)"></i><i class="c1" style="flex:0.633515" title="model.backbone.conv_encoder.model.swin.encoder.layers: 4.9M (63.4%)"></i><i class="c2" style="flex:0.257497" title="model.text_backbone: 2.0M (25.7%)"></i><i class="c3" style="flex:0.022864" title="model.encoder.layers: 176.4K (2.3%)"></i><i class="c4" style="flex:0.018451" title="model.decoder.layers: 142.3K (1.8%)"></i><i class="c5" style="flex:0.045762" title="model.input_proj_vision: 353.0K (4.6%)"></i><i class="c6" style="flex:0.007467" title="model.query_position_embeddings: 57.6K (0.7%)"></i><i class="c7" style="flex:0.005767" title="8 others: 44.5K (0.6%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.text_backbone.encoder.layer</td><td>85.1M</td><td>49.4%</td><td>66.9K</td><td>0.9%</td></tr><tr><td><i class="c1"></i>model.backbone.conv_encoder.model.swin.encoder.layers</td><td>27.5M</td><td>16.0%</td><td>4.9M</td><td>63.4%</td></tr><tr><td><i class="c2"></i>model.text_backbone</td><td>23.8M</td><td>13.8%</td><td>2.0M</td><td>25.7%</td></tr><tr><td><i class="c3"></i>model.encoder.layers</td><td>21.9M</td><td>12.7%</td><td>176.4K</td><td>2.3%</td></tr><tr><td><i class="c4"></i>model.decoder.layers</td><td>10.9M</td><td>6.3%</td><td>142.3K</td><td>1.8%</td></tr><tr><td><i class="c5"></i>model.input_proj_vision</td><td>2.1M</td><td>1.2%</td><td>353.0K</td><td>4.6%</td></tr><tr><td><i class="c6"></i>model.query_position_embeddings</td><td>230.4K</td><td>0.1%</td><td>57.6K</td><td>0.7%</td></tr><tr><td><i class="c7"></i>8 others</td><td>736.2K</td><td>0.4%</td><td>44.5K</td><td>0.6%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>grounding-dino</td><td>grounding-dino</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>model_type</td><td>swin</td><td>swin</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>image_size</td><td>224</td><td>224</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>patch_size</td><td>4</td><td>4</td><td>✓ kept</td></tr><tr class=""><td>backbone_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>bert</td><td>bert</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix GroundingDino omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>382</td><td>122</td></tr><tr class=""><td>LayerNorm</td><td>119</td><td>43</td></tr><tr class=""><td>Dropout</td><td>68</td><td>18</td></tr><tr class=""><td>ModuleList</td><td>26</td><td>18</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>6</td></tr><tr class=""><td>MultiheadAttention</td><td>18</td><td>6</td></tr><tr class=""><td>ReLU</td><td>18</td><td>6</td></tr><tr class=""><td>MLPPredictionHead</td><td>14</td><td>6</td></tr><tr class=""><td>BertAttention</td><td>12</td><td>2</td></tr><tr class=""><td>BertIntermediate</td><td>12</td><td>2</td></tr><tr class=""><td>BertLayer</td><td>12</td><td>2</td></tr><tr class=""><td>BertOutput</td><td>12</td><td>2</td></tr><tr class=""><td>BertSelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>BertSelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>MultiscaleDeformableAttention</td><td>12</td><td>4</td></tr><tr class=""><td>MultiScaleDeformableAttention</td><td>12</td><td>4</td></tr><tr class=""><td>SwinAttention</td><td>12</td><td>4</td></tr><tr class=""><td>SwinLayer</td><td>12</td><td>4</td></tr><tr class=""><td>SwinMLP</td><td>12</td><td>4</td></tr><tr class=""><td>SwinRelativePositionBias</td><td>12</td><td>4</td></tr><tr class=""><td>SwinDropPath</td><td>11</td><td>3</td></tr><tr class=""><td>ContrastiveEmbedding</td><td>7</td><td>3</td></tr><tr class=""><td>BiMultiHeadAttention</td><td>6</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>DeformableLayer</td><td>6</td><td>2</td></tr><tr class=""><td>DropPath</td><td>6</td><td>2</td></tr><tr class=""><td>EncoderLayer</td><td>6</td><td>2</td></tr><tr class=""><td>FusionLayer</td><td>6</td><td>2</td></tr><tr class=""><td>TextEnhancerLayer</td><td>6</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>5</td><td>5</td></tr><tr class=""><td>Embedding</td><td>4</td><td>4</td></tr><tr class=""><td>GroupNorm</td><td>4</td><td>4</td></tr><tr class=""><td>Sequential</td><td>4</td><td>4</td></tr><tr class=""><td>SwinStage</td><td>4</td><td>4</td></tr><tr class=""><td>SwinPatchMerging</td><td>3</td><td>3</td></tr><tr class=""><td>BertEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>BertEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>BertModel</td><td>1</td><td>1</td></tr><tr class=""><td>ConvEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>ConvModel</td><td>1</td><td>1</td></tr><tr class=""><td>Decoder</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForObjectDetection</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>SinePositionEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>Identity</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleDict</td><td>1</td><td>1</td></tr><tr class=""><td>SwinBackbone</td><td>1</td><td>1</td></tr><tr class=""><td>SwinEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>SwinEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>SwinModel</td><td>1</td><td>1</td></tr><tr class=""><td>SwinPatchEmbeddings</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">16 fields changed, 91 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>backbone_config.depths</td><td>[2, 2, 6, 2]</td><td>[1, 1, 1, 1]</td></tr><tr><td>backbone_config.embed_dim</td><td>96</td><td>64</td></tr><tr><td>backbone_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>backbone_config.num_heads</td><td>[3, 6, 12, 24]</td><td>[2, 4, 8, 16]</td></tr><tr><td>backbone_config.num_layers</td><td>4</td><td>2</td></tr><tr><td>d_model</td><td>256</td><td>64</td></tr><tr><td>decoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>decoder_ffn_dim</td><td>2048</td><td>128</td></tr><tr><td>decoder_layers</td><td>6</td><td>2</td></tr><tr><td>encoder_attention_heads</td><td>8</td><td>4</td></tr><tr><td>encoder_ffn_dim</td><td>2048</td><td>128</td></tr><tr><td>encoder_layers</td><td>6</td><td>2</td></tr><tr><td>text_config.hidden_size</td><td>768</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/grounding-dino-tiny-mic.json b/shrink/grounding-dino-tiny-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..dab864e375b8b8c09498ebf61d5ceec7d4052f65
--- /dev/null
+++ b/shrink/grounding-dino-tiny-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/grounding-dino-tiny-mic@fb35353454540cd677db5cdcf3df93a1b60cca4a", "findings": []}
\ No newline at end of file
diff --git a/shrink/hubert-large-ls960-ft-mic.html b/shrink/hubert-large-ls960-ft-mic.html
index f4934d24388da3bfbab107a08beefd7c43f5e80b..8ddbb8ea217aebfc335fa84cd36b114166c548eb 100644
--- a/shrink/hubert-large-ls960-ft-mic.html
+++ b/shrink/hubert-large-ls960-ft-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>facebook/hubert-large-ls960-ft: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/hubert-large-ls960-ft-mic@5f4ea3f40f370b2a93c05c17a1ea550648a1d8b7 ← facebook/hubert-large-ls960-ft@ece5fabbf034c1073acae96d5401b25be96709d8 · HubertForCTC · overfit weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>facebook/hubert-large-ls960-ft</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/hubert-large-ls960-ft-mic/tree/5f4ea3f40f370b2a93c05c17a1ea550648a1d8b7" title="huggingface-mic/hubert-large-ls960-ft-mic@5f4ea3f40f370b2a93c05c17a1ea550648a1d8b7">huggingface-mic/hubert-large-ls960-ft-mic@5f4ea3f40f37</a> ← <a href="https://huggingface.co/facebook/hubert-large-ls960-ft/tree/ece5fabbf034c1073acae96d5401b25be96709d8" title="facebook/hubert-large-ls960-ft@ece5fabbf034c1073acae96d5401b25be96709d8">facebook/hubert-large-ls960-ft@ece5fabbf034</a> · HubertForCTC · overfit weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">315.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">45.0K</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>hubert.feature_extractor.conv_layers</h3><div class="row"><span class="lab">Original · 7</span><div class="cells"><i class="c0" title="layer 0: LayerNormConvLayer"></i><i class="c0" title="layer 1: LayerNormConvLayer"></i><i class="c0" title="layer 2: LayerNormConvLayer"></i><i class="c0" title="layer 3: LayerNormConvLayer"></i><i class="c0" title="layer 4: LayerNormConvLayer"></i><i class="c0" title="layer 5: LayerNormConvLayer"></i><i class="c0" title="layer 6: LayerNormConvLayer"></i></div></div><div class="row"><span class="lab">Fixture · 7</span><div class="cells"><i class="c0" title="layer 0: LayerNormConvLayer"></i><i class="c0" title="layer 1: LayerNormConvLayer"></i><i class="c0" title="layer 2: LayerNormConvLayer"></i><i class="c0" title="layer 3: LayerNormConvLayer"></i><i class="c0" title="layer 4: LayerNormConvLayer"></i><i class="c0" title="layer 5: LayerNormConvLayer"></i><i class="c0" title="layer 6: LayerNormConvLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>LayerNormConvLayer</td><td>7</td><td>7</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>hubert.encoder.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i><i class="c0" title="layer 2: Linear"></i><i class="c0" title="layer 3: Linear"></i><i class="c0" title="layer 4: Linear"></i><i class="c0" title="layer 5: Linear"></i><i class="c0" title="layer 6: Linear"></i><i class="c0" title="layer 7: Linear"></i><i class="c0" title="layer 8: Linear"></i><i class="c0" title="layer 9: Linear"></i><i class="c0" title="layer 10: Linear"></i><i class="c0" title="layer 11: Linear"></i><i class="c0" title="layer 12: Linear"></i><i class="c0" title="layer 13: Linear"></i><i class="c0" title="layer 14: Linear"></i><i class="c0" title="layer 15: Linear"></i><i class="c0" title="layer 16: Linear"></i><i class="c0" title="layer 17: Linear"></i><i class="c0" title="layer 18: Linear"></i><i class="c0" title="layer 19: Linear"></i><i class="c0" title="layer 20: Linear"></i><i class="c0" title="layer 21: Linear"></i><i class="c0" title="layer 22: Linear"></i><i class="c0" title="layer 23: Linear"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Linear"></i><i class="c0" title="layer 1: Linear"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Linear</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.958278" title="hubert.encoder.layers: 302.3M (95.8%)"></i><i class="c1" style="flex:0.026601" title="hubert.encoder: 8.4M (2.7%)"></i><i class="c2" style="flex:0.013346" title="hubert.feature_extractor.conv_layers: 4.2M (1.3%)"></i><i class="c3" style="flex:0.001668" title="hubert.feature_projection: 526.3K (0.2%)"></i><i class="c4" style="flex:0.000104" title="lm_head: 32.8K (0.0%)"></i><i class="c5" style="flex:0.000003" title="hubert: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.379801" title="hubert.encoder.layers: 17.1K (38.0%)"></i><i class="c1" style="flex:0.184922" title="hubert.encoder: 8.3K (18.5%)"></i><i class="c2" style="flex:0.386202" title="hubert.feature_extractor.conv_layers: 17.4K (38.6%)"></i><i class="c3" style="flex:0.024893" title="hubert.feature_projection: 1.1K (2.5%)"></i><i class="c4" style="flex:0.023471" title="lm_head: 1.1K (2.3%)"></i><i class="c5" style="flex:0.000711" title="hubert: 32 (0.1%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>hubert.encoder.layers</td><td>302.3M</td><td>95.8%</td><td>17.1K</td><td>38.0%</td></tr><tr><td><i class="c1"></i>hubert.encoder</td><td>8.4M</td><td>2.7%</td><td>8.3K</td><td>18.5%</td></tr><tr><td><i class="c2"></i>hubert.feature_extractor.conv_layers</td><td>4.2M</td><td>1.3%</td><td>17.4K</td><td>38.6%</td></tr><tr><td><i class="c3"></i>hubert.feature_projection</td><td>526.3K</td><td>0.2%</td><td>1.1K</td><td>2.5%</td></tr><tr><td><i class="c4"></i>lm_head</td><td>32.8K</td><td>0.0%</td><td>1.1K</td><td>2.3%</td></tr><tr><td><i class="c5"></i>hubert</td><td>1.0K</td><td>0.0%</td><td>32</td><td>0.1%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>hubert</td><td>hubert</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32</td><td>32</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Hubert omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>146</td><td>14</td></tr><tr class=""><td>Dropout</td><td>75</td><td>9</td></tr><tr class=""><td>LayerNorm</td><td>57</td><td>13</td></tr><tr class=""><td>GELUActivation</td><td>32</td><td>10</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>EncoderLayerStableLayerNorm</td><td>24</td><td>2</td></tr><tr class=""><td>FeedForward</td><td>24</td><td>2</td></tr><tr class=""><td>Conv1d</td><td>7</td><td>7</td></tr><tr class=""><td>LayerNormConvLayer</td><td>7</td><td>7</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>EncoderStableLayerNorm</td><td>1</td><td>1</td></tr><tr class=""><td>FeatureEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>FeatureProjection</td><td>1</td><td>1</td></tr><tr class=""><td>ForCTC</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>PositionalConvEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>SamePadLayer</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleDict</td><td>1</td><td>1</td></tr><tr class=""><td>ParametrizationList</td><td>1</td><td>1</td></tr><tr class=""><td>ParametrizedConv1d</td><td>1</td><td>1</td></tr><tr class=""><td>_WeightNorm</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">7 fields changed, 46 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>conv_dim</td><td>7 entries: 7× 512</td><td>7 entries: 7× 32</td></tr><tr><td>hidden_size</td><td>1024</td><td>32</td></tr><tr><td>intermediate_size</td><td>4096</td><td>64</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_conv_pos_embedding_groups</td><td>16</td><td>4</td></tr><tr><td>num_conv_pos_embeddings</td><td>128</td><td>32</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/hubert-large-ls960-ft-mic.json b/shrink/hubert-large-ls960-ft-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..ec3650b4c2096fc8797e26023cec2bef1c20363b
--- /dev/null
+++ b/shrink/hubert-large-ls960-ft-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/hubert-large-ls960-ft-mic@5f4ea3f40f370b2a93c05c17a1ea550648a1d8b7", "findings": []}
\ No newline at end of file
diff --git a/shrink/internvl-mic.html b/shrink/internvl-mic.html
index d866be01dd0aeed19a19b4dc773611206e5747f1..ffd96dd1da654f1fba4f28f975d2590ebbd05593 100644
--- a/shrink/internvl-mic.html
+++ b/shrink/internvl-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>OpenGVLab/InternVL3-1B-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/internvl-mic@9bf394da9450b552cff318878bcdee131e669a91 ← OpenGVLab/InternVL3-1B-hf@014c0583a0d4bedf29fbe2dbff4f865eb998e171 · InternVLForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>OpenGVLab/InternVL3-1B-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/internvl-mic/tree/9bf394da9450b552cff318878bcdee131e669a91" title="huggingface-mic/internvl-mic@9bf394da9450b552cff318878bcdee131e669a91">huggingface-mic/internvl-mic@9bf394da9450</a> ← <a href="https://huggingface.co/OpenGVLab/InternVL3-1B-hf/tree/014c0583a0d4bedf29fbe2dbff4f865eb998e171" title="OpenGVLab/InternVL3-1B-hf@014c0583a0d4bedf29fbe2dbff4f865eb998e171">OpenGVLab/InternVL3-1B-hf@014c0583a0d4</a> · InternVLForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">802.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">10.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layer</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: VisionLayer"></i><i class="c0" title="layer 1: VisionLayer"></i><i class="c0" title="layer 2: VisionLayer"></i><i class="c0" title="layer 3: VisionLayer"></i><i class="c0" title="layer 4: VisionLayer"></i><i class="c0" title="layer 5: VisionLayer"></i><i class="c0" title="layer 6: VisionLayer"></i><i class="c0" title="layer 7: VisionLayer"></i><i class="c0" title="layer 8: VisionLayer"></i><i class="c0" title="layer 9: VisionLayer"></i><i class="c0" title="layer 10: VisionLayer"></i><i class="c0" title="layer 11: VisionLayer"></i><i class="c0" title="layer 12: VisionLayer"></i><i class="c0" title="layer 13: VisionLayer"></i><i class="c0" title="layer 14: VisionLayer"></i><i class="c0" title="layer 15: VisionLayer"></i><i class="c0" title="layer 16: VisionLayer"></i><i class="c0" title="layer 17: VisionLayer"></i><i class="c0" title="layer 18: VisionLayer"></i><i class="c0" title="layer 19: VisionLayer"></i><i class="c0" title="layer 20: VisionLayer"></i><i class="c0" title="layer 21: VisionLayer"></i><i class="c0" title="layer 22: VisionLayer"></i><i class="c0" title="layer 23: VisionLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: VisionLayer"></i><i class="c0" title="layer 1: VisionLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>VisionLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i><i class="c0" title="layer 2: Qwen2DecoderLayer"></i><i class="c0" title="layer 3: Qwen2DecoderLayer"></i><i class="c0" title="layer 4: Qwen2DecoderLayer"></i><i class="c0" title="layer 5: Qwen2DecoderLayer"></i><i class="c0" title="layer 6: Qwen2DecoderLayer"></i><i class="c0" title="layer 7: Qwen2DecoderLayer"></i><i class="c0" title="layer 8: Qwen2DecoderLayer"></i><i class="c0" title="layer 9: Qwen2DecoderLayer"></i><i class="c0" title="layer 10: Qwen2DecoderLayer"></i><i class="c0" title="layer 11: Qwen2DecoderLayer"></i><i class="c0" title="layer 12: Qwen2DecoderLayer"></i><i class="c0" title="layer 13: Qwen2DecoderLayer"></i><i class="c0" title="layer 14: Qwen2DecoderLayer"></i><i class="c0" title="layer 15: Qwen2DecoderLayer"></i><i class="c0" title="layer 16: Qwen2DecoderLayer"></i><i class="c0" title="layer 17: Qwen2DecoderLayer"></i><i class="c0" title="layer 18: Qwen2DecoderLayer"></i><i class="c0" title="layer 19: Qwen2DecoderLayer"></i><i class="c0" title="layer 20: Qwen2DecoderLayer"></i><i class="c0" title="layer 21: Qwen2DecoderLayer"></i><i class="c0" title="layer 22: Qwen2DecoderLayer"></i><i class="c0" title="layer 23: Qwen2DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.446093" title="model.language_model.layers: 357.9M (44.6%)"></i><i class="c1" style="flex:0.376868" title="model.vision_tower.encoder.layer: 302.4M (37.7%)"></i><i class="c2" style="flex:0.169390" title="model.language_model: 135.9M (16.9%)"></i><i class="c3" style="flex:0.005588" title="model.multi_modal_projector: 4.5M (0.6%)"></i><i class="c4" style="flex:0.002061" title="model.vision_tower: 1.7M (0.2%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.007444" title="model.language_model.layers: 74.2K (0.7%)"></i><i class="c1" style="flex:0.006738" title="model.vision_tower.encoder.layer: 67.2K (0.7%)"></i><i class="c2" style="flex:0.973336" title="model.language_model: 9.7M (97.3%)"></i><i class="c3" style="flex:0.002118" title="model.multi_modal_projector: 21.1K (0.2%)"></i><i class="c4" style="flex:0.010364" title="model.vision_tower: 103.4K (1.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>357.9M</td><td>44.6%</td><td>74.2K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layer</td><td>302.4M</td><td>37.7%</td><td>67.2K</td><td>0.7%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>135.9M</td><td>16.9%</td><td>9.7M</td><td>97.3%</td></tr><tr><td><i class="c3"></i>model.multi_modal_projector</td><td>4.5M</td><td>0.6%</td><td>21.1K</td><td>0.2%</td></tr><tr><td><i class="c4"></i>model.vision_tower</td><td>1.7M</td><td>0.2%</td><td>103.4K</td><td>1.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>internvl</td><td>internvl</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>internvl_vision</td><td>internvl_vision</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>[448, 448]</td><td>[448, 448]</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>[14, 14]</td><td>[14, 14]</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (14 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>dynamic</td><td>dynamic</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2</td><td>qwen2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>151674</td><td>151674</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix InternVL omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>315</td><td>29</td></tr><tr class=""><td>Identity</td><td>73</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Qwen2RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>Dropout</td><td>25</td><td>3</td></tr><tr class=""><td>GELUActivation</td><td>25</td><td>3</td></tr><tr class=""><td>VisionAttention</td><td>24</td><td>2</td></tr><tr class=""><td>VisionLayer</td><td>24</td><td>2</td></tr><tr class=""><td>VisionMLP</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>VisionEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>VisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>VisionPatchEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2RotaryEmbedding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 68 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.hidden_size</td><td>896</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>4864</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>24 entries: 24× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>14</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>24</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/internvl-mic.json b/shrink/internvl-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..24c6ecdb342a6fae5b6acb6ad5210f80fd475dde
--- /dev/null
+++ b/shrink/internvl-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/internvl-mic@9bf394da9450b552cff318878bcdee131e669a91", "findings": []}
\ No newline at end of file
diff --git a/shrink/jina-embeddings-v3-hf-mic.html b/shrink/jina-embeddings-v3-hf-mic.html
index 8915da9eb00119a472135c93e0e9e87be198ef9e..25a495da7762d5bb9574be6ee16d4f3b369a4eca 100644
--- a/shrink/jina-embeddings-v3-hf-mic.html
+++ b/shrink/jina-embeddings-v3-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>jinaai/jina-embeddings-v3-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/jina-embeddings-v3-hf-mic@7288a7adb5335fb7805f7c28ffa3efd588b8be67 ← jinaai/jina-embeddings-v3-hf@d18862d9a48706220815554fac3ebb4dfa46fc28 · JinaEmbeddingsV3Model · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>jinaai/jina-embeddings-v3-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/jina-embeddings-v3-hf-mic/tree/7288a7adb5335fb7805f7c28ffa3efd588b8be67" title="huggingface-mic/jina-embeddings-v3-hf-mic@7288a7adb5335fb7805f7c28ffa3efd588b8be67">huggingface-mic/jina-embeddings-v3-hf-mic@7288a7adb533</a> ← <a href="https://huggingface.co/jinaai/jina-embeddings-v3-hf/tree/d18862d9a48706220815554fac3ebb4dfa46fc28" title="jinaai/jina-embeddings-v3-hf@d18862d9a48706220815554fac3ebb4dfa46fc28">jinaai/jina-embeddings-v3-hf@d18862d9a487</a> · JinaEmbeddingsV3Model · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">559.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">16.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.540452" title="layers: 302.3M (54.0%)"></i><i class="c1" style="flex:0.457666" title="embeddings.word_embeddings: 256.0M (45.8%)"></i><i class="c2" style="flex:0.001876" title="pooler.dense: 1.0M (0.2%)"></i><i class="c3" style="flex:0.000004" title="embeddings.LayerNorm: 2.0K (0.0%)"></i><i class="c4" style="flex:0.000002" title="embeddings.token_type_embeddings: 1.0K (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.004165" title="layers: 66.9K (0.4%)"></i><i class="c1" style="flex:0.995564" title="embeddings.word_embeddings: 16.0M (99.6%)"></i><i class="c2" style="flex:0.000259" title="pooler.dense: 4.2K (0.0%)"></i><i class="c3" style="flex:0.000008" title="embeddings.LayerNorm: 128 (0.0%)"></i><i class="c4" style="flex:0.000004" title="embeddings.token_type_embeddings: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>layers</td><td>302.3M</td><td>54.0%</td><td>66.9K</td><td>0.4%</td></tr><tr><td><i class="c1"></i>embeddings.word_embeddings</td><td>256.0M</td><td>45.8%</td><td>16.0M</td><td>99.6%</td></tr><tr><td><i class="c2"></i>pooler.dense</td><td>1.0M</td><td>0.2%</td><td>4.2K</td><td>0.0%</td></tr><tr><td><i class="c3"></i>embeddings.LayerNorm</td><td>2.0K</td><td>0.0%</td><td>128</td><td>0.0%</td></tr><tr><td><i class="c4"></i>embeddings.token_type_embeddings</td><td>1.0K</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>jina_embeddings_v3</td><td>jina_embeddings_v3</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>250002</td><td>250002</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8194</td><td>8194</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix JinaEmbeddingsV3 omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>145</td><td>13</td></tr><tr class=""><td>Dropout</td><td>49</td><td>5</td></tr><tr class=""><td>LayerNorm</td><td>49</td><td>5</td></tr><tr class=""><td>GELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Layer</td><td>24</td><td>2</td></tr><tr class=""><td>MLP</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 38 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>1024</td><td>64</td></tr><tr><td>intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>24</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/jina-embeddings-v3-hf-mic.json b/shrink/jina-embeddings-v3-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..e966761672c5987d4e771f3e4aab0723b9a82015
--- /dev/null
+++ b/shrink/jina-embeddings-v3-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/jina-embeddings-v3-hf-mic@7288a7adb5335fb7805f7c28ffa3efd588b8be67", "findings": []}
\ No newline at end of file
diff --git a/shrink/language-perceiver-mic.html b/shrink/language-perceiver-mic.html
index 4c15f4e047e34c75869f55529ecdf898579c4be5..8b3636b875504e632d77880ac685f6526eba92fc 100644
--- a/shrink/language-perceiver-mic.html
+++ b/shrink/language-perceiver-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>deepmind/language-perceiver: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/language-perceiver-mic@6c70b7e085e8e6fd4544e720e75ea0e0cbdd438a ← deepmind/language-perceiver@47a1ba43d8198ea10cf9a158fce114953e9441a1 · PerceiverForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
-<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">201.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">194.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>perceiver.encoder.self_attends</h3><div class="row"><span class="lab">Original · 26</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i><i class="c0" title="layer 24: Layer"></i><i class="c0" title="layer 25: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 26</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i><i class="c0" title="layer 24: Layer"></i><i class="c0" title="layer 25: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>26</td><td>26</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.933391" title="perceiver.encoder.self_attends: 187.7M (93.3%)"></i><i class="c1" style="flex:0.031997" title="perceiver.encoder: 6.4M (3.2%)"></i><i class="c2" style="flex:0.024161" title="perceiver.decoder: 4.9M (2.4%)"></i><i class="c3" style="flex:0.008822" title="perceiver.input_preprocessor: 1.8M (0.9%)"></i><i class="c4" style="flex:0.001629" title="perceiver.embeddings: 327.7K (0.2%)"></i><i class="c5" style="flex:0.000001" title="embedding_decoder: 262 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.967026" title="perceiver.encoder.self_attends: 187.7M (96.7%)"></i><i class="c1" style="flex:0.027572" title="perceiver.encoder: 5.4M (2.8%)"></i><i class="c2" style="flex:0.002952" title="perceiver.decoder: 572.9K (0.3%)"></i><i class="c3" style="flex:0.000762" title="perceiver.input_preprocessor: 147.8K (0.1%)"></i><i class="c4" style="flex:0.001688" title="perceiver.embeddings: 327.7K (0.2%)"></i><i class="c5" style="flex:0.000001" title="embedding_decoder: 262 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>perceiver.encoder.self_attends</td><td>187.7M</td><td>93.3%</td><td>187.7M</td><td>96.7%</td></tr><tr><td><i class="c1"></i>perceiver.encoder</td><td>6.4M</td><td>3.2%</td><td>5.4M</td><td>2.8%</td></tr><tr><td><i class="c2"></i>perceiver.decoder</td><td>4.9M</td><td>2.4%</td><td>572.9K</td><td>0.3%</td></tr><tr><td><i class="c3"></i>perceiver.input_preprocessor</td><td>1.8M</td><td>0.9%</td><td>147.8K</td><td>0.1%</td></tr><tr><td><i class="c4"></i>perceiver.embeddings</td><td>327.7K</td><td>0.2%</td><td>327.7K</td><td>0.2%</td></tr><tr><td><i class="c5"></i>embedding_decoder</td><td>262</td><td>0.0%</td><td>262</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>perceiver</td><td>perceiver</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>262</td><td>262</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>56</td><td>56</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Perceiver omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>168</td><td>168</td></tr><tr class=""><td>LayerNorm</td><td>58</td><td>58</td></tr><tr class=""><td>Dropout</td><td>28</td><td>28</td></tr><tr class=""><td>GELUActivation</td><td>28</td><td>28</td></tr><tr class=""><td>Identity</td><td>28</td><td>28</td></tr><tr class=""><td>Attention</td><td>28</td><td>28</td></tr><tr class=""><td>Layer</td><td>28</td><td>28</td></tr><tr class=""><td>MLP</td><td>28</td><td>28</td></tr><tr class=""><td>SelfAttention</td><td>28</td><td>28</td></tr><tr class=""><td>SelfOutput</td><td>28</td><td>28</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>BasicDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>EmbeddingDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextPreprocessor</td><td>1</td><td>1</td></tr><tr class=""><td>TrainablePositionEncoding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">1 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>d_model</td><td>768</td><td>64</td></tr></table>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>deepmind/language-perceiver</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/language-perceiver-mic/tree/6c70b7e085e8e6fd4544e720e75ea0e0cbdd438a" title="huggingface-mic/language-perceiver-mic@6c70b7e085e8e6fd4544e720e75ea0e0cbdd438a">huggingface-mic/language-perceiver-mic@6c70b7e085e8</a> ← <a href="https://huggingface.co/deepmind/language-perceiver/tree/47a1ba43d8198ea10cf9a158fce114953e9441a1" title="deepmind/language-perceiver@47a1ba43d8198ea10cf9a158fce114953e9441a1">deepmind/language-perceiver@47a1ba43d819</a> · PerceiverForMaskedLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<div class="stats"><div class="stat"><span class="k">Original</span><span class="v">201.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">194.1M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">1</span><span class="k">to review</span></div></div><section class="findings warn"><h2>⚠ To review</h2><ul><li>the fixture keeps 97% of the original&#x27;s 201.1M parameters</li></ul></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>perceiver.encoder.self_attends</h3><div class="row"><span class="lab">Original · 26</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i><i class="c0" title="layer 24: Layer"></i><i class="c0" title="layer 25: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 26</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i><i class="c0" title="layer 12: Layer"></i><i class="c0" title="layer 13: Layer"></i><i class="c0" title="layer 14: Layer"></i><i class="c0" title="layer 15: Layer"></i><i class="c0" title="layer 16: Layer"></i><i class="c0" title="layer 17: Layer"></i><i class="c0" title="layer 18: Layer"></i><i class="c0" title="layer 19: Layer"></i><i class="c0" title="layer 20: Layer"></i><i class="c0" title="layer 21: Layer"></i><i class="c0" title="layer 22: Layer"></i><i class="c0" title="layer 23: Layer"></i><i class="c0" title="layer 24: Layer"></i><i class="c0" title="layer 25: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>26</td><td>26</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.933391" title="perceiver.encoder.self_attends: 187.7M (93.3%)"></i><i class="c1" style="flex:0.031997" title="perceiver.encoder: 6.4M (3.2%)"></i><i class="c2" style="flex:0.024161" title="perceiver.decoder: 4.9M (2.4%)"></i><i class="c3" style="flex:0.008822" title="perceiver.input_preprocessor: 1.8M (0.9%)"></i><i class="c4" style="flex:0.001629" title="perceiver.embeddings: 327.7K (0.2%)"></i><i class="c5" style="flex:0.000001" title="embedding_decoder: 262 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.967026" title="perceiver.encoder.self_attends: 187.7M (96.7%)"></i><i class="c1" style="flex:0.027572" title="perceiver.encoder: 5.4M (2.8%)"></i><i class="c2" style="flex:0.002952" title="perceiver.decoder: 572.9K (0.3%)"></i><i class="c3" style="flex:0.000762" title="perceiver.input_preprocessor: 147.8K (0.1%)"></i><i class="c4" style="flex:0.001688" title="perceiver.embeddings: 327.7K (0.2%)"></i><i class="c5" style="flex:0.000001" title="embedding_decoder: 262 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>perceiver.encoder.self_attends</td><td>187.7M</td><td>93.3%</td><td>187.7M</td><td>96.7%</td></tr><tr><td><i class="c1"></i>perceiver.encoder</td><td>6.4M</td><td>3.2%</td><td>5.4M</td><td>2.8%</td></tr><tr><td><i class="c2"></i>perceiver.decoder</td><td>4.9M</td><td>2.4%</td><td>572.9K</td><td>0.3%</td></tr><tr><td><i class="c3"></i>perceiver.input_preprocessor</td><td>1.8M</td><td>0.9%</td><td>147.8K</td><td>0.1%</td></tr><tr><td><i class="c4"></i>perceiver.embeddings</td><td>327.7K</td><td>0.2%</td><td>327.7K</td><td>0.2%</td></tr><tr><td><i class="c5"></i>embedding_decoder</td><td>262</td><td>0.0%</td><td>262</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>perceiver</td><td>perceiver</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>262</td><td>262</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>2048</td><td>2048</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>image_size</td><td>56</td><td>56</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Perceiver omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>168</td><td>168</td></tr><tr class=""><td>LayerNorm</td><td>58</td><td>58</td></tr><tr class=""><td>Dropout</td><td>28</td><td>28</td></tr><tr class=""><td>GELUActivation</td><td>28</td><td>28</td></tr><tr class=""><td>Identity</td><td>28</td><td>28</td></tr><tr class=""><td>Attention</td><td>28</td><td>28</td></tr><tr class=""><td>Layer</td><td>28</td><td>28</td></tr><tr class=""><td>MLP</td><td>28</td><td>28</td></tr><tr class=""><td>SelfAttention</td><td>28</td><td>28</td></tr><tr class=""><td>SelfOutput</td><td>28</td><td>28</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>BasicDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>EmbeddingDecoder</td><td>1</td><td>1</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>TextPreprocessor</td><td>1</td><td>1</td></tr><tr class=""><td>TrainablePositionEncoding</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">1 fields changed, 35 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>d_model</td><td>768</td><td>64</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/language-perceiver-mic.json b/shrink/language-perceiver-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..bfff2d51048d82b19de38a888d9e40166836912c
--- /dev/null
+++ b/shrink/language-perceiver-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/language-perceiver-mic@6c70b7e085e8e6fd4544e720e75ea0e0cbdd438a", "findings": ["the fixture keeps 97% of the original's 201.1M parameters"]}
\ No newline at end of file
diff --git a/shrink/layoutlm-base-uncased-mic.html b/shrink/layoutlm-base-uncased-mic.html
index 518865d49dfd45901791914153268d30b67f0b31..976ec11e5d5759e41c965704735d5b748814aa9d 100644
--- a/shrink/layoutlm-base-uncased-mic.html
+++ b/shrink/layoutlm-base-uncased-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>microsoft/layoutlm-base-uncased: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/layoutlm-base-uncased-mic@7aa363dd7f97db6248afc704386db78edea79ae6 ← microsoft/layoutlm-base-uncased@30e3cdd39c11f09757b0fcf7598533d05052acef · LayoutLMForMaskedLM · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>microsoft/layoutlm-base-uncased</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/layoutlm-base-uncased-mic/tree/7aa363dd7f97db6248afc704386db78edea79ae6" title="huggingface-mic/layoutlm-base-uncased-mic@7aa363dd7f97db6248afc704386db78edea79ae6">huggingface-mic/layoutlm-base-uncased-mic@7aa363dd7f97</a> ← <a href="https://huggingface.co/microsoft/layoutlm-base-uncased/tree/30e3cdd39c11f09757b0fcf7598533d05052acef" title="microsoft/layoutlm-base-uncased@30e3cdd39c11f09757b0fcf7598533d05052acef">microsoft/layoutlm-base-uncased@30e3cdd39c11</a> · LayoutLMForMaskedLM · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">113.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">2.4M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>layoutlm.encoder.layer</h3><div class="row"><span class="lab">Original · 12</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i><i class="c0" title="layer 2: Layer"></i><i class="c0" title="layer 3: Layer"></i><i class="c0" title="layer 4: Layer"></i><i class="c0" title="layer 5: Layer"></i><i class="c0" title="layer 6: Layer"></i><i class="c0" title="layer 7: Layer"></i><i class="c0" title="layer 8: Layer"></i><i class="c0" title="layer 9: Layer"></i><i class="c0" title="layer 10: Layer"></i><i class="c0" title="layer 11: Layer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Layer"></i><i class="c0" title="layer 1: Layer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Layer</td><td>12</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.751029" title="layoutlm.encoder.layer: 85.1M (75.1%)"></i><i class="c1" style="flex:0.238258" title="layoutlm.embeddings: 27.0M (23.8%)"></i><i class="c2" style="flex:0.005498" title="cls.predictions: 622.6K (0.5%)"></i><i class="c3" style="flex:0.005215" title="layoutlm.pooler: 590.6K (0.5%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.028432" title="layoutlm.encoder.layer: 66.9K (2.8%)"></i><i class="c1" style="flex:0.955016" title="layoutlm.embeddings: 2.2M (95.5%)"></i><i class="c2" style="flex:0.014785" title="cls.predictions: 34.8K (1.5%)"></i><i class="c3" style="flex:0.001767" title="layoutlm.pooler: 4.2K (0.2%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>layoutlm.encoder.layer</td><td>85.1M</td><td>75.1%</td><td>66.9K</td><td>2.8%</td></tr><tr><td><i class="c1"></i>layoutlm.embeddings</td><td>27.0M</td><td>23.8%</td><td>2.2M</td><td>95.5%</td></tr><tr><td><i class="c2"></i>cls.predictions</td><td>622.6K</td><td>0.5%</td><td>34.8K</td><td>1.5%</td></tr><tr><td><i class="c3"></i>layoutlm.pooler</td><td>590.6K</td><td>0.5%</td><td>4.2K</td><td>0.2%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>layoutlm</td><td>layoutlm</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>30522</td><td>30522</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>gelu</td><td>gelu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>512</td><td>512</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix LayoutLM omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>75</td><td>15</td></tr><tr class=""><td>Dropout</td><td>37</td><td>7</td></tr><tr class=""><td>LayerNorm</td><td>26</td><td>6</td></tr><tr class=""><td>GELUActivation</td><td>13</td><td>3</td></tr><tr class=""><td>Attention</td><td>12</td><td>2</td></tr><tr class=""><td>Intermediate</td><td>12</td><td>2</td></tr><tr class=""><td>Layer</td><td>12</td><td>2</td></tr><tr class=""><td>Output</td><td>12</td><td>2</td></tr><tr class=""><td>SelfAttention</td><td>12</td><td>2</td></tr><tr class=""><td>SelfOutput</td><td>12</td><td>2</td></tr><tr class=""><td>Embedding</td><td>7</td><td>7</td></tr><tr class=""><td>Embeddings</td><td>1</td><td>1</td></tr><tr class=""><td>Encoder</td><td>1</td><td>1</td></tr><tr class=""><td>ForMaskedLM</td><td>1</td><td>1</td></tr><tr class=""><td>LMPredictionHead</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>OnlyMLMHead</td><td>1</td><td>1</td></tr><tr class=""><td>Pooler</td><td>1</td><td>1</td></tr><tr class=""><td>PredictionHeadTransform</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr><tr class=""><td>Tanh</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">4 fields changed, 25 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>hidden_size</td><td>768</td><td>64</td></tr><tr><td>intermediate_size</td><td>3072</td><td>128</td></tr><tr><td>num_attention_heads</td><td>12</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>12</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/layoutlm-base-uncased-mic.json b/shrink/layoutlm-base-uncased-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..032de0df3ea32a3b3239924c198eac5759f4b94e
--- /dev/null
+++ b/shrink/layoutlm-base-uncased-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/layoutlm-base-uncased-mic@7aa363dd7f97db6248afc704386db78edea79ae6", "findings": []}
\ No newline at end of file
diff --git a/shrink/llama-mic.html b/shrink/llama-mic.html
index 58c7cd961f5f9fd4cc6dec3a59f3351ddf316ff7..4bb4d3186dbe3b1906ce7e9c76b71ba58bc1be17 100644
--- a/shrink/llama-mic.html
+++ b/shrink/llama-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>HuggingFaceTB/SmolLM2-135M: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/llama-mic@8231ab98d1dd670f702753a80097b854fb5e91b5 ← HuggingFaceTB/SmolLM2-135M@93efa2f097d58c2a74874c7e644dbc9b0cee75a2 · LlamaForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>HuggingFaceTB/SmolLM2-135M</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/llama-mic/tree/8231ab98d1dd670f702753a80097b854fb5e91b5" title="huggingface-mic/llama-mic@8231ab98d1dd670f702753a80097b854fb5e91b5">huggingface-mic/llama-mic@8231ab98d1dd</a> ← <a href="https://huggingface.co/HuggingFaceTB/SmolLM2-135M/tree/93efa2f097d58c2a74874c7e644dbc9b0cee75a2" title="HuggingFaceTB/SmolLM2-135M@93efa2f097d58c2a74874c7e644dbc9b0cee75a2">HuggingFaceTB/SmolLM2-135M@93efa2f097d5</a> · LlamaForCausalLM · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">134.5M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">3.2M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">1</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.layers</h3><div class="row"><span class="lab">Original · 30</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i><i class="c0" title="layer 2: DecoderLayer"></i><i class="c0" title="layer 3: DecoderLayer"></i><i class="c0" title="layer 4: DecoderLayer"></i><i class="c0" title="layer 5: DecoderLayer"></i><i class="c0" title="layer 6: DecoderLayer"></i><i class="c0" title="layer 7: DecoderLayer"></i><i class="c0" title="layer 8: DecoderLayer"></i><i class="c0" title="layer 9: DecoderLayer"></i><i class="c0" title="layer 10: DecoderLayer"></i><i class="c0" title="layer 11: DecoderLayer"></i><i class="c0" title="layer 12: DecoderLayer"></i><i class="c0" title="layer 13: DecoderLayer"></i><i class="c0" title="layer 14: DecoderLayer"></i><i class="c0" title="layer 15: DecoderLayer"></i><i class="c0" title="layer 16: DecoderLayer"></i><i class="c0" title="layer 17: DecoderLayer"></i><i class="c0" title="layer 18: DecoderLayer"></i><i class="c0" title="layer 19: DecoderLayer"></i><i class="c0" title="layer 20: DecoderLayer"></i><i class="c0" title="layer 21: DecoderLayer"></i><i class="c0" title="layer 22: DecoderLayer"></i><i class="c0" title="layer 23: DecoderLayer"></i><i class="c0" title="layer 24: DecoderLayer"></i><i class="c0" title="layer 25: DecoderLayer"></i><i class="c0" title="layer 26: DecoderLayer"></i><i class="c0" title="layer 27: DecoderLayer"></i><i class="c0" title="layer 28: DecoderLayer"></i><i class="c0" title="layer 29: DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: DecoderLayer"></i><i class="c0" title="layer 1: DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>DecoderLayer</td><td>30</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.789524" title="model.layers: 106.2M (79.0%)"></i><i class="c1" style="flex:0.210471" title="model.embed_tokens: 28.3M (21.0%)"></i><i class="c2" style="flex:0.000004" title="model.norm: 576 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.022978" title="model.layers: 74.0K (2.3%)"></i><i class="c1" style="flex:0.977002" title="model.embed_tokens: 3.1M (97.7%)"></i><i class="c2" style="flex:0.000020" title="model.norm: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.layers</td><td>106.2M</td><td>79.0%</td><td>74.0K</td><td>2.3%</td></tr><tr><td><i class="c1"></i>model.embed_tokens</td><td>28.3M</td><td>21.0%</td><td>3.1M</td><td>97.7%</td></tr><tr><td><i class="c2"></i>model.norm</td><td>576</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>attention style</td><td>GQA (9 heads, 3 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>model_type</td><td>llama</td><td>llama</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>49152</td><td>49152</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>max_position_embeddings</td><td>8192</td><td>8192</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Llama omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>211</td><td>15</td></tr><tr class=""><td>RMSNorm</td><td>61</td><td>5</td></tr><tr class=""><td>Attention</td><td>30</td><td>2</td></tr><tr class=""><td>DecoderLayer</td><td>30</td><td>2</td></tr><tr class=""><td>MLP</td><td>30</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>30</td><td>2</td></tr><tr class=""><td>Embedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForCausalLM</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ModuleList</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">6 fields changed, 26 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>head_dim</td><td>64</td><td>16</td></tr><tr><td>hidden_size</td><td>576</td><td>64</td></tr><tr><td>intermediate_size</td><td>1536</td><td>128</td></tr><tr><td>num_attention_heads</td><td>9</td><td>4</td></tr><tr><td>num_hidden_layers</td><td>30</td><td>2</td></tr><tr><td>num_key_value_heads</td><td>3</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/llama-mic.json b/shrink/llama-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..77b7841235a4a901cb680c333f247a1c9c1f260f
--- /dev/null
+++ b/shrink/llama-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/llama-mic@8231ab98d1dd670f702753a80097b854fb5e91b5", "findings": []}
\ No newline at end of file
diff --git a/shrink/llava-mic.html b/shrink/llava-mic.html
index 0040252a3a4a2726e009eb78841d19093e2ce5bb..b4c73dc6dfa74bf087c72bbf3c238a7d719d1932 100644
--- a/shrink/llava-mic.html
+++ b/shrink/llava-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>llava-hf/llava-1.5-7b-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/llava-mic@dab442b1a1f407ad1083d0a1b5cf11dc99c02294 ← llava-hf/llava-1.5-7b-hf@b234b804b114d9e37bb655e11cbbb5f5e971b7a9 · LlavaForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>llava-hf/llava-1.5-7b-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/llava-mic/tree/dab442b1a1f407ad1083d0a1b5cf11dc99c02294" title="huggingface-mic/llava-mic@dab442b1a1f407ad1083d0a1b5cf11dc99c02294">huggingface-mic/llava-mic@dab442b1a1f4</a> ← <a href="https://huggingface.co/llava-hf/llava-1.5-7b-hf/tree/b234b804b114d9e37bb655e11cbbb5f5e971b7a9" title="llava-hf/llava-1.5-7b-hf@b234b804b114d9e37bb655e11cbbb5f5e971b7a9">llava-hf/llava-1.5-7b-hf@b234b804b114</a> · LlavaForConditionalGeneration · random weights · built before the shrink rule (no recorded recipe)</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">7.1B</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">4.3M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i><i class="c0" title="layer 2: CLIPEncoderLayer"></i><i class="c0" title="layer 3: CLIPEncoderLayer"></i><i class="c0" title="layer 4: CLIPEncoderLayer"></i><i class="c0" title="layer 5: CLIPEncoderLayer"></i><i class="c0" title="layer 6: CLIPEncoderLayer"></i><i class="c0" title="layer 7: CLIPEncoderLayer"></i><i class="c0" title="layer 8: CLIPEncoderLayer"></i><i class="c0" title="layer 9: CLIPEncoderLayer"></i><i class="c0" title="layer 10: CLIPEncoderLayer"></i><i class="c0" title="layer 11: CLIPEncoderLayer"></i><i class="c0" title="layer 12: CLIPEncoderLayer"></i><i class="c0" title="layer 13: CLIPEncoderLayer"></i><i class="c0" title="layer 14: CLIPEncoderLayer"></i><i class="c0" title="layer 15: CLIPEncoderLayer"></i><i class="c0" title="layer 16: CLIPEncoderLayer"></i><i class="c0" title="layer 17: CLIPEncoderLayer"></i><i class="c0" title="layer 18: CLIPEncoderLayer"></i><i class="c0" title="layer 19: CLIPEncoderLayer"></i><i class="c0" title="layer 20: CLIPEncoderLayer"></i><i class="c0" title="layer 21: CLIPEncoderLayer"></i><i class="c0" title="layer 22: CLIPEncoderLayer"></i><i class="c0" title="layer 23: CLIPEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: CLIPEncoderLayer"></i><i class="c0" title="layer 1: CLIPEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>CLIPEncoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 32</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i><i class="c0" title="layer 2: LlamaDecoderLayer"></i><i class="c0" title="layer 3: LlamaDecoderLayer"></i><i class="c0" title="layer 4: LlamaDecoderLayer"></i><i class="c0" title="layer 5: LlamaDecoderLayer"></i><i class="c0" title="layer 6: LlamaDecoderLayer"></i><i class="c0" title="layer 7: LlamaDecoderLayer"></i><i class="c0" title="layer 8: LlamaDecoderLayer"></i><i class="c0" title="layer 9: LlamaDecoderLayer"></i><i class="c0" title="layer 10: LlamaDecoderLayer"></i><i class="c0" title="layer 11: LlamaDecoderLayer"></i><i class="c0" title="layer 12: LlamaDecoderLayer"></i><i class="c0" title="layer 13: LlamaDecoderLayer"></i><i class="c0" title="layer 14: LlamaDecoderLayer"></i><i class="c0" title="layer 15: LlamaDecoderLayer"></i><i class="c0" title="layer 16: LlamaDecoderLayer"></i><i class="c0" title="layer 17: LlamaDecoderLayer"></i><i class="c0" title="layer 18: LlamaDecoderLayer"></i><i class="c0" title="layer 19: LlamaDecoderLayer"></i><i class="c0" title="layer 20: LlamaDecoderLayer"></i><i class="c0" title="layer 21: LlamaDecoderLayer"></i><i class="c0" title="layer 22: LlamaDecoderLayer"></i><i class="c0" title="layer 23: LlamaDecoderLayer"></i><i class="c0" title="layer 24: LlamaDecoderLayer"></i><i class="c0" title="layer 25: LlamaDecoderLayer"></i><i class="c0" title="layer 26: LlamaDecoderLayer"></i><i class="c0" title="layer 27: LlamaDecoderLayer"></i><i class="c0" title="layer 28: LlamaDecoderLayer"></i><i class="c0" title="layer 29: LlamaDecoderLayer"></i><i class="c0" title="layer 30: LlamaDecoderLayer"></i><i class="c0" title="layer 31: LlamaDecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: LlamaDecoderLayer"></i><i class="c0" title="layer 1: LlamaDecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>LlamaDecoderLayer</td><td>32</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.916873" title="model.language_model.layers: 6.5B (91.7%)"></i><i class="c1" style="flex:0.042799" title="model.vision_tower.encoder.layers: 302.3M (4.3%)"></i><i class="c2" style="flex:0.018594" title="model.language_model: 131.3M (1.9%)"></i><i class="c3" style="flex:0.018594" title="lm_head: 131.3M (1.9%)"></i><i class="c4" style="flex:0.002970" title="model.multi_modal_projector: 21.0M (0.3%)"></i><i class="c5" style="flex:0.000170" title="model.vision_tower: 1.2M (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.018950" title="model.language_model.layers: 82.2K (1.9%)"></i><i class="c1" style="flex:0.015437" title="model.vision_tower.encoder.layers: 66.9K (1.5%)"></i><i class="c2" style="flex:0.473221" title="model.language_model: 2.1M (47.3%)"></i><i class="c3" style="flex:0.473207" title="lm_head: 2.1M (47.3%)"></i><i class="c4" style="flex:0.001919" title="model.multi_modal_projector: 8.3K (0.2%)"></i><i class="c5" style="flex:0.017267" title="model.vision_tower: 74.9K (1.7%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.language_model.layers</td><td>6.5B</td><td>91.7%</td><td>82.2K</td><td>1.9%</td></tr><tr><td><i class="c1"></i>model.vision_tower.encoder.layers</td><td>302.3M</td><td>4.3%</td><td>66.9K</td><td>1.5%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>131.3M</td><td>1.9%</td><td>2.1M</td><td>47.3%</td></tr><tr><td><i class="c3"></i>lm_head</td><td>131.3M</td><td>1.9%</td><td>2.1M</td><td>47.3%</td></tr><tr><td><i class="c4"></i>model.multi_modal_projector</td><td>21.0M</td><td>0.3%</td><td>8.3K</td><td>0.2%</td></tr><tr><td><i class="c5"></i>model.vision_tower</td><td>1.2M</td><td>0.0%</td><td>74.9K</td><td>1.7%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>llava</td><td>llava</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>vocab_size</td><td>32064</td><td>32064</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>clip_vision_model</td><td>clip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>vocab_size</td><td>32000</td><td>32000</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>quick_gelu</td><td>quick_gelu</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>336</td><td>336</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>MHA (32 heads, 32 KV)</td><td>MHA (4 heads, 4 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>llama</td><td>llama</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>32064</td><td>32064</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>False</td><td>False</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>4096</td><td>4096</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix Llava omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>371</td><td>29</td></tr><tr class=""><td>LlamaRMSNorm</td><td>65</td><td>5</td></tr><tr class=""><td>LayerNorm</td><td>50</td><td>6</td></tr><tr class=""><td>LlamaAttention</td><td>32</td><td>2</td></tr><tr class=""><td>LlamaDecoderLayer</td><td>32</td><td>2</td></tr><tr class=""><td>LlamaMLP</td><td>32</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>32</td><td>2</td></tr><tr class=""><td>CLIPAttention</td><td>24</td><td>2</td></tr><tr class=""><td>CLIPEncoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>CLIPMLP</td><td>24</td><td>2</td></tr><tr class=""><td>QuickGELUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>CLIPEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>CLIPVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>CLIPVisionModel</td><td>1</td><td>1</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaModel</td><td>1</td><td>1</td></tr><tr class=""><td>LlamaRotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">11 fields changed, 61 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.head_dim</td><td>128</td><td>16</td></tr><tr><td>text_config.hidden_size</td><td>4096</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>11008</td><td>128</td></tr><tr><td>text_config.num_attention_heads</td><td>32</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>32</td><td>2</td></tr><tr><td>text_config.num_key_value_heads</td><td>32</td><td>4</td></tr><tr><td>vision_config.hidden_size</td><td>1024</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4096</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.projection_dim</td><td>768</td><td>32</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/llava-mic.json b/shrink/llava-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..4584615c09ee3fed6d4c9ce26f18370ca0f8d2c6
--- /dev/null
+++ b/shrink/llava-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/llava-mic@dab442b1a1f407ad1083d0a1b5cf11dc99c02294", "findings": []}
\ No newline at end of file
diff --git a/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.html b/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.html
index 711d257983e23ba0cd7525dc8085c66a7144e069..13545079852f93ac6920b08b6434a9ca6c70c6ca 100644
--- a/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.html
+++ b/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.html
@@ -1,6 +1,6 @@
 <!doctype html>
 <html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Shrunk architecture</title>
+<title>llava-hf/llava-onevision-qwen2-0.5b-ov-hf: shrunk architecture</title>
 <style>
 :root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#52514e;--line:#e4e3df;--ok:#1d7a43;--warn:#b35c00;--s0:#2a78d6;--s1:#eb6834;--s2:#1baf7a;--s3:#eda100;--s4:#e87ba4;--s5:#008300;--s6:#6250d6;--s7:#e34948;--s8:#8a8984;}
 @media (prefers-color-scheme:dark){:root:not([data-theme="light"]){--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3936;--ok:#2f9e5a;--warn:#d9822b;--s0:#3987e5;--s1:#d95926;--s2:#199e70;--s3:#c98500;--s4:#d55181;--s5:#008300;--s6:#9085e9;--s7:#e66767;--s8:#6f6e69;}}
@@ -9,6 +9,7 @@ body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.5 system-ui,sans-
 main{max-width:1040px;margin:0 auto;padding:24px 16px 64px}
 h1{font-size:24px;margin:0 0 4px}h2{font-size:18px;margin:36px 0 8px}h3{font-size:14px;margin:0 0 8px;font-family:ui-monospace,monospace}
 .meta,.note{color:var(--muted);margin:0 0 12px;overflow-wrap:anywhere}
+a{color:inherit}
 .stats{display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));gap:12px;margin:20px 0}
 .stat{border:1px solid var(--line);border-radius:8px;padding:12px;display:flex;flex-direction:column}
 .stat .v{font-size:26px;font-weight:600}.stat .k{color:var(--muted);font-size:13px}
@@ -29,6 +30,7 @@ td i{display:inline-block;width:10px;height:10px;border-radius:2px;margin-right:
 tr.bad td{color:var(--warn)}
 .c0{background:var(--s0)}.c1{background:var(--s1)}.c2{background:var(--s2)}.c3{background:var(--s3)}.c4{background:var(--s4)}.c5{background:var(--s5)}.c6{background:var(--s6)}.c7{background:var(--s7)}.c8{background:var(--s8)}
 </style></head><body><main>
-<h1>Shrunk architecture</h1><p class="meta">huggingface-mic/llava-onevision-qwen2-0.5b-ov-hf-mic@c8d8d3c952d7ba613a1c76b3eb2d1641b8fd053f ← llava-hf/llava-onevision-qwen2-0.5b-ov-hf@74dd0bf867a4cda7950c17663794267c60cf4b40 · LlavaOnevisionForConditionalGeneration · random weights</p>
+<p class="meta"><a href="../index.html">← MIC coverage of Transformers model types</a></p>
+<h1>llava-hf/llava-onevision-qwen2-0.5b-ov-hf</h1><p class="meta"><a href="https://huggingface.co/huggingface-mic/llava-onevision-qwen2-0.5b-ov-hf-mic/tree/c8d8d3c952d7ba613a1c76b3eb2d1641b8fd053f" title="huggingface-mic/llava-onevision-qwen2-0.5b-ov-hf-mic@c8d8d3c952d7ba613a1c76b3eb2d1641b8fd053f">huggingface-mic/llava-onevision-qwen2-0.5b-ov-hf-mic@c8d8d3c952d7</a> ← <a href="https://huggingface.co/llava-hf/llava-onevision-qwen2-0.5b-ov-hf/tree/74dd0bf867a4cda7950c17663794267c60cf4b40" title="llava-hf/llava-onevision-qwen2-0.5b-ov-hf@74dd0bf867a4cda7950c17663794267c60cf4b40">llava-hf/llava-onevision-qwen2-0.5b-ov-hf@74dd0bf867a4</a> · LlavaOnevisionForConditionalGeneration · random weights</p>
 <div class="stats"><div class="stat"><span class="k">Original</span><span class="v">893.7M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Fixture</span><span class="v">10.0M</span><span class="k">parameters</span></div><div class="stat"><span class="k">Layer stacks</span><span class="v">2</span><span class="k">compared</span></div><div class="stat"><span class="k">Findings</span><span class="v">0</span><span class="k">to review</span></div></div><section class="findings ok"><h2>✓ Same architecture, smaller</h2><p>Every module class, layer kind and checked property of the original is in the fixture.</p></section><h2>Layer stacks</h2><p class="note">One square per layer, colored by kind: the layer's module classes and type attributes. Hover a square for its index and kind.</p><div class="stack"><h3>model.vision_tower.encoder.layers</h3><div class="row"><span class="lab">Original · 26</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i><i class="c0" title="layer 2: SiglipEncoderLayer"></i><i class="c0" title="layer 3: SiglipEncoderLayer"></i><i class="c0" title="layer 4: SiglipEncoderLayer"></i><i class="c0" title="layer 5: SiglipEncoderLayer"></i><i class="c0" title="layer 6: SiglipEncoderLayer"></i><i class="c0" title="layer 7: SiglipEncoderLayer"></i><i class="c0" title="layer 8: SiglipEncoderLayer"></i><i class="c0" title="layer 9: SiglipEncoderLayer"></i><i class="c0" title="layer 10: SiglipEncoderLayer"></i><i class="c0" title="layer 11: SiglipEncoderLayer"></i><i class="c0" title="layer 12: SiglipEncoderLayer"></i><i class="c0" title="layer 13: SiglipEncoderLayer"></i><i class="c0" title="layer 14: SiglipEncoderLayer"></i><i class="c0" title="layer 15: SiglipEncoderLayer"></i><i class="c0" title="layer 16: SiglipEncoderLayer"></i><i class="c0" title="layer 17: SiglipEncoderLayer"></i><i class="c0" title="layer 18: SiglipEncoderLayer"></i><i class="c0" title="layer 19: SiglipEncoderLayer"></i><i class="c0" title="layer 20: SiglipEncoderLayer"></i><i class="c0" title="layer 21: SiglipEncoderLayer"></i><i class="c0" title="layer 22: SiglipEncoderLayer"></i><i class="c0" title="layer 23: SiglipEncoderLayer"></i><i class="c0" title="layer 24: SiglipEncoderLayer"></i><i class="c0" title="layer 25: SiglipEncoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: SiglipEncoderLayer"></i><i class="c0" title="layer 1: SiglipEncoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>SiglipEncoderLayer</td><td>26</td><td>2</td><td>✓ kept</td></tr></table></div><div class="stack"><h3>model.language_model.layers</h3><div class="row"><span class="lab">Original · 24</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i><i class="c0" title="layer 2: Qwen2DecoderLayer"></i><i class="c0" title="layer 3: Qwen2DecoderLayer"></i><i class="c0" title="layer 4: Qwen2DecoderLayer"></i><i class="c0" title="layer 5: Qwen2DecoderLayer"></i><i class="c0" title="layer 6: Qwen2DecoderLayer"></i><i class="c0" title="layer 7: Qwen2DecoderLayer"></i><i class="c0" title="layer 8: Qwen2DecoderLayer"></i><i class="c0" title="layer 9: Qwen2DecoderLayer"></i><i class="c0" title="layer 10: Qwen2DecoderLayer"></i><i class="c0" title="layer 11: Qwen2DecoderLayer"></i><i class="c0" title="layer 12: Qwen2DecoderLayer"></i><i class="c0" title="layer 13: Qwen2DecoderLayer"></i><i class="c0" title="layer 14: Qwen2DecoderLayer"></i><i class="c0" title="layer 15: Qwen2DecoderLayer"></i><i class="c0" title="layer 16: Qwen2DecoderLayer"></i><i class="c0" title="layer 17: Qwen2DecoderLayer"></i><i class="c0" title="layer 18: Qwen2DecoderLayer"></i><i class="c0" title="layer 19: Qwen2DecoderLayer"></i><i class="c0" title="layer 20: Qwen2DecoderLayer"></i><i class="c0" title="layer 21: Qwen2DecoderLayer"></i><i class="c0" title="layer 22: Qwen2DecoderLayer"></i><i class="c0" title="layer 23: Qwen2DecoderLayer"></i></div></div><div class="row"><span class="lab">Fixture · 2</span><div class="cells"><i class="c0" title="layer 0: Qwen2DecoderLayer"></i><i class="c0" title="layer 1: Qwen2DecoderLayer"></i></div></div><table class="legend"><tr><th>Kind</th><th>Original</th><th>Fixture</th><th></th></tr><tr><td><i class="c0"></i>Qwen2DecoderLayer</td><td>24</td><td>2</td><td>✓ kept</td></tr></table></div><h2>Where the parameters are</h2><p class="note">Share of each model's parameters per component. Vocabulary and patch sizes are kept, so embeddings usually take a larger share of the fixture.</p><div class="row"><span class="lab">Original</span><div class="bar"><i class="c0" style="flex:0.443368" title="model.vision_tower.encoder.layers: 396.2M (44.3%)"></i><i class="c1" style="flex:0.400478" title="model.language_model.layers: 357.9M (40.0%)"></i><i class="c2" style="flex:0.152396" title="model.language_model: 136.2M (15.2%)"></i><i class="c3" style="flex:0.002055" title="model.multi_modal_projector: 1.8M (0.2%)"></i><i class="c4" style="flex:0.001702" title="model.vision_tower: 1.5M (0.2%)"></i><i class="c5" style="flex:0.000001" title="model: 896 (0.0%)"></i></div></div><div class="row"><span class="lab">Fixture</span><div class="bar"><i class="c0" style="flex:0.006720" title="model.vision_tower.encoder.layers: 66.9K (0.7%)"></i><i class="c1" style="flex:0.007452" title="model.language_model.layers: 74.2K (0.7%)"></i><i class="c2" style="flex:0.976506" title="model.language_model: 9.7M (97.7%)"></i><i class="c3" style="flex:0.000835" title="model.multi_modal_projector: 8.3K (0.1%)"></i><i class="c4" style="flex:0.008480" title="model.vision_tower: 84.5K (0.8%)"></i><i class="c5" style="flex:0.000006" title="model: 64 (0.0%)"></i></div></div><table class="legend"><tr><th>Component</th><th>Original</th><th>Share</th><th>Fixture</th><th>Share</th></tr><tr><td><i class="c0"></i>model.vision_tower.encoder.layers</td><td>396.2M</td><td>44.3%</td><td>66.9K</td><td>0.7%</td></tr><tr><td><i class="c1"></i>model.language_model.layers</td><td>357.9M</td><td>40.0%</td><td>74.2K</td><td>0.7%</td></tr><tr><td><i class="c2"></i>model.language_model</td><td>136.2M</td><td>15.2%</td><td>9.7M</td><td>97.7%</td></tr><tr><td><i class="c3"></i>model.multi_modal_projector</td><td>1.8M</td><td>0.2%</td><td>8.3K</td><td>0.1%</td></tr><tr><td><i class="c4"></i>model.vision_tower</td><td>1.5M</td><td>0.2%</td><td>84.5K</td><td>0.8%</td></tr><tr><td><i class="c5"></i>model</td><td>896</td><td>0.0%</td><td>64</td><td>0.0%</td></tr></table><h2>Preserved properties</h2><table><tr><th>Config</th><th>Property</th><th>Original</th><th>Fixture</th><th></th></tr><tr class=""><td>model</td><td>model_type</td><td>llava_onevision</td><td>llava_onevision</td><td>✓ kept</td></tr><tr class=""><td>model</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>model_type</td><td>siglip_vision_model</td><td>siglip_vision_model</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>hidden_act</td><td>gelu_pytorch_tanh</td><td>gelu_pytorch_tanh</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>image_size</td><td>384</td><td>384</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>patch_size</td><td>14</td><td>14</td><td>✓ kept</td></tr><tr class=""><td>vision_config</td><td>num_channels</td><td>3</td><td>3</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>attention style</td><td>GQA (14 heads, 2 KV)</td><td>GQA (4 heads, 2 KV)</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>RoPE type</td><td>default</td><td>default</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>model_type</td><td>qwen2</td><td>qwen2</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>vocab_size</td><td>152000</td><td>152000</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>tie_word_embeddings</td><td>True</td><td>True</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>hidden_act</td><td>silu</td><td>silu</td><td>✓ kept</td></tr><tr class=""><td>text_config</td><td>max_position_embeddings</td><td>32768</td><td>32768</td><td>✓ kept</td></tr></table><h2>Module classes</h2><p class="note">Instances of each class in the built model (prefix LlavaOnevision omitted).</p><table><tr><th>Class</th><th>Original</th><th>Fixture</th></tr><tr class=""><td>Linear</td><td>327</td><td>29</td></tr><tr class=""><td>LayerNorm</td><td>53</td><td>5</td></tr><tr class=""><td>Qwen2RMSNorm</td><td>49</td><td>5</td></tr><tr class=""><td>GELUTanh</td><td>26</td><td>2</td></tr><tr class=""><td>SiglipAttention</td><td>26</td><td>2</td></tr><tr class=""><td>SiglipEncoderLayer</td><td>26</td><td>2</td></tr><tr class=""><td>SiglipMLP</td><td>26</td><td>2</td></tr><tr class=""><td>Qwen2Attention</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2DecoderLayer</td><td>24</td><td>2</td></tr><tr class=""><td>Qwen2MLP</td><td>24</td><td>2</td></tr><tr class=""><td>SiLUActivation</td><td>24</td><td>2</td></tr><tr class=""><td>Embedding</td><td>2</td><td>2</td></tr><tr class=""><td>ModuleList</td><td>2</td><td>2</td></tr><tr class=""><td>Conv2d</td><td>1</td><td>1</td></tr><tr class=""><td>GELUActivation</td><td>1</td><td>1</td></tr><tr class=""><td>ForConditionalGeneration</td><td>1</td><td>1</td></tr><tr class=""><td>Model</td><td>1</td><td>1</td></tr><tr class=""><td>MultiModalProjector</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2Model</td><td>1</td><td>1</td></tr><tr class=""><td>Qwen2RotaryEmbedding</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipEncoder</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionEmbeddings</td><td>1</td><td>1</td></tr><tr class=""><td>SiglipVisionModel</td><td>1</td><td>1</td></tr></table><h2>Config changes</h2><p class="note">9 fields changed, 61 unchanged.</p><table><tr><th>Field</th><th>Original</th><th>Fixture</th></tr><tr><td>text_config.hidden_size</td><td>896</td><td>64</td></tr><tr><td>text_config.intermediate_size</td><td>4864</td><td>128</td></tr><tr><td>text_config.layer_types</td><td>24 entries: 24× full_attention</td><td>[&quot;full_attention&quot;, &quot;full_attention&quot;]</td></tr><tr><td>text_config.num_attention_heads</td><td>14</td><td>4</td></tr><tr><td>text_config.num_hidden_layers</td><td>24</td><td>2</td></tr><tr><td>vision_config.hidden_size</td><td>1152</td><td>64</td></tr><tr><td>vision_config.intermediate_size</td><td>4304</td><td>128</td></tr><tr><td>vision_config.num_attention_heads</td><td>16</td><td>4</td></tr><tr><td>vision_config.num_hidden_layers</td><td>26</td><td>2</td></tr></table>
 </main></body></html>
\ No newline at end of file
diff --git a/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.json b/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.json
new file mode 100644
index 0000000000000000000000000000000000000000..b963342da990d82b424689bb9fc59773fe4ae78a
--- /dev/null
+++ b/shrink/llava-onevision-qwen2-0.5b-ov-hf-mic.json
@@ -0,0 +1 @@
+{"fixture": "huggingface-mic/llava-onevision-qwen2-0.5b-ov-hf-mic@c8d8d3c952d7ba613a1c76b3eb2d1641b8fd053f", "findings": []}
\ No newline at end of file