Add BF16-sourced Trim INT6 and native NVFP4 encoder for Track C

README.md

@@ -0,0 +1,43 @@
+---
+library_name: fastvideo
+license: other
+license_name: source-model-licenses
+license_link: https://huggingface.co/FastVideo/FastVideo-FastH3-Trim-8-Step
+---
+# FastH3 Trim 8-step MLX INT6
+
+Converted from `FastVideo/FastVideo-FastH3-Trim-8-Step`, revision
+`9edcb6511d5ff8dd1579e8eec8afee77e098b8d8`, from the released BF16 transformer.
+The transformer uses MLX affine INT6 weights, retained VSA gates, and a cached
+rank-16 AdaLN ladder. The native NVFP4 text encoder is stored in MLX layout
+without requantization. The 26-layer lightweight video VAE and audio VAE are
+included. `fastvideo_inference.json` supplies the eight-forward schedule.
+
+Use the Spark/Mac code in hao-ai-lab/FastVideo#1920. After downloading this
+snapshot to MODEL_ROOT, run:
+
+```python
+from fastvideo.mlx_runtime.minimax_h3_pipeline import MiniMaxH3MLXPipeline
+from pathlib import Path
+root = Path("MODEL_ROOT")
+pipeline = MiniMaxH3MLXPipeline(
+    model_root=root, mlx_dit_checkpoint=root / "transformer",
+    conditioner_mode="nvfp4", vae_dtype="fp16", resident=False,
+    metal_wired_limit_gib=27,
+)
+try:
+    pipeline.generate("A red fox in snow.", output_path="clip.mp4",
+                      width=832, height=480, num_frames=124, seed=1234,
+                      num_steps=8, vsa=True, vsa_sparsity=0.8,
+                      vsa_tile_size=64, vsa_impl="reference")
+finally:
+    pipeline.close()
+```
+
+Use reference attention for release generation. The optional SIMD Metal
+backend has a known partial-tile correctness issue. Phased placement keeps one
+heavy component loaded at a time on the 36 GiB Mac. Transformer storage is
+15,355,458,984 bytes including its manifest, before the encoder and VAEs.
+
+This is an experimental pruned model. Source-model licensing and acceptable-use
+terms still apply. See the source repository and its parent model licenses.

audio_scheduler/scheduler_config.json

@@ -0,0 +1,5 @@
+{
+  "_class_name": "MiniMaxH3Scheduler",
+  "_diffusers_version": "0.36.0.dev0",
+  "shift": 3.0
+}

audio_vae/config.json

@@ -0,0 +1,125 @@
+{
+  "_class_name": "AutoencoderKLMiniMaxH3Audio",
+  "_diffusers_version": "0.36.0.dev0",
+  "encoder_dim": 64,
+  "encoder_rates": [
+    2,
+    4,
+    4,
+    5,
+    5
+  ],
+  "latent_dim": 2048,
+  "latent_channels": 32,
+  "decoder_dim": 1024,
+  "decoder_rates": [
+    5,
+    5,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "decoder_kernel_sizes": [
+    9,
+    9,
+    4,
+    4,
+    4,
+    4,
+    4
+  ],
+  "num_attention_heads": 8,
+  "resblock_kernel_sizes": [
+    3,
+    7,
+    11
+  ],
+  "resblock_dilation_sizes": [
+    [
+      1,
+      3,
+      5
+    ],
+    [
+      1,
+      3,
+      5
+    ],
+    [
+      1,
+      3,
+      5
+    ]
+  ],
+  "sampling_rate": 32000,
+  "latents_mean": [
+    -0.020211687488382354,
+    0.3876466479950502,
+    -0.04398279799186767,
+    -0.28591514936373,
+    0.08179686214561671,
+    -0.35782641352446604,
+    0.040623809960919084,
+    -0.01552534501956604,
+    -0.223362481667332,
+    0.1821006842509091,
+    0.2941778783780663,
+    -0.07901167601970885,
+    -0.056815072777201,
+    -0.3699028221860095,
+    -0.31616315591624855,
+    0.5905951377425391,
+    -0.052139568068853864,
+    0.013673160263486295,
+    -0.03691647864630577,
+    0.09732660653298163,
+    -0.3394662328788498,
+    -0.30685677538541667,
+    -0.24504598907458763,
+    -0.034698524462007344,
+    0.02868032184767538,
+    -0.21217779266454084,
+    -0.1678263169941987,
+    0.3221287889040614,
+    -0.1223055851554907,
+    0.4356604928128464,
+    -0.0502599202236253,
+    0.3979258376211797
+  ],
+  "latents_std": [
+    1.6895524230479284,
+    2.76263727217653,
+    1.7945344281264435,
+    1.6801681847309828,
+    1.6390226546605453,
+    2.7788298348882177,
+    1.7659090095747236,
+    1.6199757612137327,
+    2.6336525640336896,
+    1.8539356672817833,
+    2.5056497896915633,
+    1.811019237886178,
+    1.9579657790720237,
+    1.6685498243529284,
+    1.4922469314453364,
+    3.298670198067373,
+    1.9491804496832168,
+    1.8720003270431442,
+    1.8334080103291832,
+    1.6488070416529093,
+    1.6176957696319716,
+    1.9131449234774398,
+    1.5695245398428617,
+    1.6943659940415912,
+    1.8318420762504692,
+    1.5540637421583379,
+    1.9344930328968526,
+    1.599198216109855,
+    1.718045989838149,
+    1.6307219190837705,
+    1.8661226051202384,
+    1.5613768203168363
+  ]
+}

audio_vae/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52c59e67ba8de5477c81bfbced0327aabf500f1bfdeefd5ee754529241cb26cb
+size 605429340

conversion_provenance.json

@@ -0,0 +1,114 @@
+{
+  "source_repo": "FastVideo/FastVideo-FastH3-Trim-8-Step",
+  "source_revision": "9edcb6511d5ff8dd1579e8eec8afee77e098b8d8",
+  "source_checks": [
+    {
+      "file": "audio_vae/diffusion_pytorch_model.safetensors",
+      "sha256": "52c59e67ba8de5477c81bfbced0327aabf500f1bfdeefd5ee754529241cb26cb",
+      "released_sha256": "52c59e67ba8de5477c81bfbced0327aabf500f1bfdeefd5ee754529241cb26cb",
+      "equal": true
+    },
+    {
+      "file": "fastvideo_inference.json",
+      "equal": true
+    },
+    {
+      "file": "transformer/config.json",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00001-of-00008.safetensors",
+      "sha256": "5cc09786e36031231b07d9976ec18dd0011e51ecf034ceaea4adb9174f035217",
+      "released_sha256": "5cc09786e36031231b07d9976ec18dd0011e51ecf034ceaea4adb9174f035217",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00002-of-00008.safetensors",
+      "sha256": "133c044a3c8b7341284e952b43db2ab6585d803e0e18520c1b56e0ce8974bde3",
+      "released_sha256": "133c044a3c8b7341284e952b43db2ab6585d803e0e18520c1b56e0ce8974bde3",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00003-of-00008.safetensors",
+      "sha256": "d2d8507e473eb350ac92606fd8202550d871f30e855d728c8efa203d2019d919",
+      "released_sha256": "d2d8507e473eb350ac92606fd8202550d871f30e855d728c8efa203d2019d919",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00004-of-00008.safetensors",
+      "sha256": "6e55b26defc637f6b4271916912e6a7c0adaea260b96372f0ebce7668b11f3ba",
+      "released_sha256": "6e55b26defc637f6b4271916912e6a7c0adaea260b96372f0ebce7668b11f3ba",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00005-of-00008.safetensors",
+      "sha256": "85d47855214ab9273eefced40142fb45832f620395bc6a73ddab8c884db625b2",
+      "released_sha256": "85d47855214ab9273eefced40142fb45832f620395bc6a73ddab8c884db625b2",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00006-of-00008.safetensors",
+      "sha256": "3de5ca9bc4004d8c1db500c624a3ba9ff8d90c61a7cf8b6c4fb8d1bc27bb4d0c",
+      "released_sha256": "3de5ca9bc4004d8c1db500c624a3ba9ff8d90c61a7cf8b6c4fb8d1bc27bb4d0c",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00007-of-00008.safetensors",
+      "sha256": "176f1d7dcb9c2596281a1f07cd419fe28be53901b347519bf5012b8797bc0a2d",
+      "released_sha256": "176f1d7dcb9c2596281a1f07cd419fe28be53901b347519bf5012b8797bc0a2d",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model-00008-of-00008.safetensors",
+      "sha256": "92a0b0400b37b961f40ad68936552e809af92410453be8cf001db7a19b61882d",
+      "released_sha256": "92a0b0400b37b961f40ad68936552e809af92410453be8cf001db7a19b61882d",
+      "equal": true
+    },
+    {
+      "file": "transformer/diffusion_pytorch_model.safetensors.index.json",
+      "equal": true
+    },
+    {
+      "file": "vae/config.json",
+      "equal": true
+    },
+    {
+      "file": "vae/diffusion_pytorch_model.safetensors",
+      "sha256": "73fc5a61caeba438633eb8f8efd3a57f4915adfa46731a1c991312c724a3b5cc",
+      "released_sha256": "73fc5a61caeba438633eb8f8efd3a57f4915adfa46731a1c991312c724a3b5cc",
+      "equal": true
+    }
+  ],
+  "converted_from": "BF16",
+  "quantization": "MLX affine INT6 with VSA gates and AdaLN timestep cache",
+  "code_commit": "6ccdbc761e6b854003f472e39b826c06cb54de60",
+  "placement": "phased on M4 Max 36 GiB",
+  "attention": "reference",
+  "encoder_source_checks": [
+    {
+      "file": "text_encoder/model-00001-of-00004.safetensors",
+      "sha256": "969e7d251a873d16d3f487d9a5e82f8901132cfac9344530aa4e25ba9338e9e7",
+      "released_sha256": "969e7d251a873d16d3f487d9a5e82f8901132cfac9344530aa4e25ba9338e9e7",
+      "equal": true
+    },
+    {
+      "file": "text_encoder/model-00002-of-00004.safetensors",
+      "sha256": "ff092cd64c489ce4609a08ba5d9f8548591bf835cd98fa4154a7eccd0ea16dfd",
+      "released_sha256": "ff092cd64c489ce4609a08ba5d9f8548591bf835cd98fa4154a7eccd0ea16dfd",
+      "equal": true
+    },
+    {
+      "file": "text_encoder/model-00003-of-00004.safetensors",
+      "sha256": "3dab4af0342cfc9b31839458d059327c7cd1e5be1fe779ded253939cabb36a51",
+      "released_sha256": "3dab4af0342cfc9b31839458d059327c7cd1e5be1fe779ded253939cabb36a51",
+      "equal": true
+    },
+    {
+      "file": "text_encoder/model-00004-of-00004.safetensors",
+      "sha256": "628852eb94bd2e3c297d120683f900b49051f236e2cd9d1f26476c30cedc2add",
+      "released_sha256": "628852eb94bd2e3c297d120683f900b49051f236e2cd9d1f26476c30cedc2add",
+      "equal": true
+    }
+  ],
+  "encoder_cache_note": "Native layout conversion without requantization; first50 transformer layers read by H3 retained; exact conditioning compared with raw encoder in prior controls"
+}

fastvideo_inference.json

@@ -0,0 +1,15 @@
+{
+  "attention_backend": "VIDEO_SPARSE_ATTN_H3",
+  "audio_scheduler_shift": 3.0,
+  "checkpoint_step": 300,
+  "dmd_denoising_steps": [999, 874, 749, 624, 500, 375, 250, 125],
+  "guidance_scale": 1.0,
+  "num_inference_steps": 9,
+  "schema_version": "fasth3-inference-contract-v1",
+  "task": "t2av",
+  "training_parent_run_id": "s42-r16-dmd8-vsa80-20260929-v1",
+  "transformer_forwards": 8,
+  "video_scheduler_shift": 10.0,
+  "vsa_sparsity": 0.8,
+  "vsa_tile_size": 64
+}

processor/chat_template.json

@@ -0,0 +1,3 @@
+{
+    "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {%- if messages[0].content is string %}\n            {{- messages[0].content }}\n        {%- else %}\n            {%- for content in messages[0].content %}\n                {%- if 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' }}\n        {%- if messages[0].content is string %}\n            {{- messages[0].content }}\n        {%- else %}\n            {%- for content in messages[0].content %}\n                {%- if 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set image_count = namespace(value=0) %}\n{%- set video_count = namespace(value=0) %}\n{%- for message in messages %}\n    {%- if message.role == \"user\" %}\n        {{- '<|im_start|>' + message.role + '\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content in message.content %}\n                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}\n                    {%- set image_count.value = image_count.value + 1 %}\n                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}\n                    <|vision_start|><|image_pad|><|vision_end|>\n                {%- elif content.type == 'video' or 'video' in content %}\n                    {%- set video_count.value = video_count.value + 1 %}\n                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}\n                    <|vision_start|><|video_pad|><|vision_end|>\n                {%- elif 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role + '\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content_item in message.content %}\n                {%- if 'text' in content_item %}\n                    {{- content_item.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and message.content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content in message.content %}\n                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}\n                    {%- set image_count.value = image_count.value + 1 %}\n                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}\n                    <|vision_start|><|image_pad|><|vision_end|>\n                {%- elif content.type == 'video' or 'video' in content %}\n                    {%- set video_count.value = video_count.value + 1 %}\n                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}\n                    <|vision_start|><|video_pad|><|vision_end|>\n                {%- elif 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n"
+}

processor/preprocessor_config.json

@@ -0,0 +1,21 @@
+{
+    "size": {
+        "longest_edge": 16777216,
+        "shortest_edge": 65536
+    },
+    "patch_size": 16,
+    "temporal_patch_size": 2,
+    "merge_size": 2,
+    "image_mean": [
+        0.5,
+        0.5,
+        0.5
+    ],
+    "image_std": [
+        0.5,
+        0.5,
+        0.5
+    ],
+    "processor_class": "Qwen3VLProcessor",
+    "image_processor_type": "Qwen2VLImageProcessorFast"
+}

processor/tokenizer_config.json

@@ -0,0 +1,246 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>",
+    "<d>",
+    "</d>",
+    "<|cutoff|>",
+    "<|lyrics_start|>",
+    "<|lyrics_end|>",
+    "<|caption_start|>",
+    "<|caption_end|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {%- if messages[0].content is string %}\n            {{- messages[0].content }}\n        {%- else %}\n            {%- for content in messages[0].content %}\n                {%- if 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' }}\n        {%- if messages[0].content is string %}\n            {{- messages[0].content }}\n        {%- else %}\n            {%- for content in messages[0].content %}\n                {%- if 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set image_count = namespace(value=0) %}\n{%- set video_count = namespace(value=0) %}\n{%- for message in messages %}\n    {%- if message.role == \"user\" %}\n        {{- '<|im_start|>' + message.role + '\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content in message.content %}\n                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}\n                    {%- set image_count.value = image_count.value + 1 %}\n                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}\n                    <|vision_start|><|image_pad|><|vision_end|>\n                {%- elif content.type == 'video' or 'video' in content %}\n                    {%- set video_count.value = video_count.value + 1 %}\n                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}\n                    <|vision_start|><|video_pad|><|vision_end|>\n                {%- elif 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role + '\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content_item in message.content %}\n                {%- if 'text' in content_item %}\n                    {{- content_item.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and message.content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content in message.content %}\n                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}\n                    {%- set image_count.value = image_count.value + 1 %}\n                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}\n                    <|vision_start|><|image_pad|><|vision_end|>\n                {%- elif content.type == 'video' or 'video' in content %}\n                    {%- set video_count.value = video_count.value + 1 %}\n                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}\n                    <|vision_start|><|video_pad|><|vision_end|>\n                {%- elif 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

processor/video_preprocessor_config.json

@@ -0,0 +1,21 @@
+{
+    "size": {
+        "longest_edge": 25165824,
+        "shortest_edge": 4096
+    },
+    "patch_size": 16,
+    "temporal_patch_size": 2,
+    "merge_size": 2,
+    "image_mean": [
+        0.5,
+        0.5,
+        0.5
+    ],
+    "image_std": [
+        0.5,
+        0.5,
+        0.5
+    ],
+    "processor_class": "Qwen3VLProcessor",
+    "video_processor_type": "Qwen3VLVideoProcessor"
+}

scheduler/scheduler_config.json

@@ -0,0 +1,5 @@
+{
+  "_class_name": "MiniMaxH3Scheduler",
+  "_diffusers_version": "0.36.0.dev0",
+  "shift": 10.0
+}

text_encoder/config.json

@@ -0,0 +1,80 @@
+{
+  "architectures": [
+    "Qwen3VLForConditionalGeneration"
+  ],
+  "image_token_id": 151655,
+  "model_type": "qwen3_vl",
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "dtype": "bfloat16",
+    "eos_token_id": 151645,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 5120,
+    "initializer_range": 0.02,
+    "intermediate_size": 25600,
+    "max_position_embeddings": 262144,
+    "model_type": "qwen3_vl_text",
+    "num_attention_heads": 64,
+    "num_hidden_layers": 64,
+    "num_key_value_heads": 8,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": {
+      "mrope_interleaved": true,
+      "mrope_section": [
+        24,
+        20,
+        20
+      ],
+      "rope_type": "default"
+    },
+    "rope_theta": 5000000,
+    "use_cache": true,
+    "vocab_size": 151936
+  },
+  "tie_word_embeddings": false,
+  "transformers_version": "4.57.0.dev0",
+  "video_token_id": 151656,
+  "vision_config": {
+    "deepstack_visual_indexes": [
+      8,
+      16,
+      24
+    ],
+    "depth": 27,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 4304,
+    "model_type": "qwen3_vl",
+    "num_heads": 16,
+    "num_position_embeddings": 2304,
+    "out_hidden_size": 5120,
+    "patch_size": 16,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "num_hidden_layers_override": 50,
+  "quantization_config": {
+    "quant_method": "nvfp4",
+    "activation_scheme": "dynamic",
+    "fmt": "e2m1",
+    "group_size": 16,
+    "scale_fmt": "e4m3",
+    "scale_layout": "128x4",
+    "modules_to_not_convert": [
+      "model.visual",
+      "lm_head"
+    ],
+    "producer": {
+      "converter": "convert_minimax_h3_text_encoder_nvfp4.py",
+      "flashinfer": "0.7.1rc2",
+      "kept_language_layers": 50
+    }
+  }
+}

text_encoder/mlx_h3_nvfp4_encoder.json

@@ -0,0 +1,1260 @@
+{
+  "format_version": 1,
+  "language_layers": 50,
+  "matrices": {
+    "model.language_model.layers.0.mlp.down_proj.weight": {
+      "global_scale": 2626.44287109375
+    },
+    "model.language_model.layers.0.mlp.gate_proj.weight": {
+      "global_scale": 7864.31982421875
+    },
+    "model.language_model.layers.0.mlp.up_proj.weight": {
+      "global_scale": 8290.6982421875
+    },
+    "model.language_model.layers.0.self_attn.k_proj.weight": {
+      "global_scale": 8765.9619140625
+    },
+    "model.language_model.layers.0.self_attn.o_proj.weight": {
+      "global_scale": 3780.923095703125
+    },
+    "model.language_model.layers.0.self_attn.q_proj.weight": {
+      "global_scale": 7864.31982421875
+    },
+    "model.language_model.layers.0.self_attn.v_proj.weight": {
+      "global_scale": 19802.244140625
+    },
+    "model.language_model.layers.1.mlp.down_proj.weight": {
+      "global_scale": 2422.98583984375
+    },
+    "model.language_model.layers.1.mlp.gate_proj.weight": {
+      "global_scale": 7955.23681640625
+    },
+    "model.language_model.layers.1.mlp.up_proj.weight": {
+      "global_scale": 5461.33349609375
+    },
+    "model.language_model.layers.1.self_attn.k_proj.weight": {
+      "global_scale": 7094.10302734375
+    },
+    "model.language_model.layers.1.self_attn.o_proj.weight": {
+      "global_scale": 6284.27392578125
+    },
+    "model.language_model.layers.1.self_attn.q_proj.weight": {
+      "global_scale": 10836.6611328125
+    },
+    "model.language_model.layers.1.self_attn.v_proj.weight": {
+      "global_scale": 14411.0576171875
+    },
+    "model.language_model.layers.10.mlp.down_proj.weight": {
+      "global_scale": 1911.4666748046875
+    },
+    "model.language_model.layers.10.mlp.gate_proj.weight": {
+      "global_scale": 5213.0908203125
+    },
+    "model.language_model.layers.10.mlp.up_proj.weight": {
+      "global_scale": 6616.61572265625
+    },
+    "model.language_model.layers.10.self_attn.k_proj.weight": {
+      "global_scale": 3373.176513671875
+    },
+    "model.language_model.layers.10.self_attn.o_proj.weight": {
+      "global_scale": 2688.0
+    },
+    "model.language_model.layers.10.self_attn.q_proj.weight": {
+      "global_scale": 11421.2119140625
+    },
+    "model.language_model.layers.10.self_attn.v_proj.weight": {
+      "global_scale": 9972.8701171875
+    },
+    "model.language_model.layers.11.mlp.down_proj.weight": {
+      "global_scale": 1890.4615478515625
+    },
+    "model.language_model.layers.11.mlp.gate_proj.weight": {
+      "global_scale": 5806.9873046875
+    },
+    "model.language_model.layers.11.mlp.up_proj.weight": {
+      "global_scale": 7561.84619140625
+    },
+    "model.language_model.layers.11.self_attn.k_proj.weight": {
+      "global_scale": 3292.478271484375
+    },
+    "model.language_model.layers.11.self_attn.o_proj.weight": {
+      "global_scale": 5831.59326171875
+    },
+    "model.language_model.layers.11.self_attn.q_proj.weight": {
+      "global_scale": 10668.6513671875
+    },
+    "model.language_model.layers.11.self_attn.v_proj.weight": {
+      "global_scale": 21011.54296875
+    },
+    "model.language_model.layers.12.mlp.down_proj.weight": {
+      "global_scale": 2060.263671875
+    },
+    "model.language_model.layers.12.mlp.gate_proj.weight": {
+      "global_scale": 5571.88671875
+    },
+    "model.language_model.layers.12.mlp.up_proj.weight": {
+      "global_scale": 7021.71435546875
+    },
+    "model.language_model.layers.12.self_attn.k_proj.weight": {
+      "global_scale": 5252.8857421875
+    },
+    "model.language_model.layers.12.self_attn.o_proj.weight": {
+      "global_scale": 3171.0966796875
+    },
+    "model.language_model.layers.12.self_attn.q_proj.weight": {
+      "global_scale": 10347.7900390625
+    },
+    "model.language_model.layers.12.self_attn.v_proj.weight": {
+      "global_scale": 12684.38671875
+    },
+    "model.language_model.layers.13.mlp.down_proj.weight": {
+      "global_scale": 2219.767822265625
+    },
+    "model.language_model.layers.13.mlp.gate_proj.weight": {
+      "global_scale": 4047.811767578125
+    },
+    "model.language_model.layers.13.mlp.up_proj.weight": {
+      "global_scale": 8241.0546875
+    },
+    "model.language_model.layers.13.self_attn.k_proj.weight": {
+      "global_scale": 4247.70361328125
+    },
+    "model.language_model.layers.13.self_attn.o_proj.weight": {
+      "global_scale": 3423.522216796875
+    },
+    "model.language_model.layers.13.self_attn.q_proj.weight": {
+      "global_scale": 9299.02734375
+    },
+    "model.language_model.layers.13.self_attn.v_proj.weight": {
+      "global_scale": 15728.6396484375
+    },
+    "model.language_model.layers.14.mlp.down_proj.weight": {
+      "global_scale": 2234.181884765625
+    },
+    "model.language_model.layers.14.mlp.gate_proj.weight": {
+      "global_scale": 4120.52734375
+    },
+    "model.language_model.layers.14.mlp.up_proj.weight": {
+      "global_scale": 6881.27978515625
+    },
+    "model.language_model.layers.14.self_attn.k_proj.weight": {
+      "global_scale": 5213.0908203125
+    },
+    "model.language_model.layers.14.self_attn.o_proj.weight": {
+      "global_scale": 2797.26806640625
+    },
+    "model.language_model.layers.14.self_attn.q_proj.weight": {
+      "global_scale": 12125.603515625
+    },
+    "model.language_model.layers.14.self_attn.v_proj.weight": {
+      "global_scale": 13045.0810546875
+    },
+    "model.language_model.layers.15.mlp.down_proj.weight": {
+      "global_scale": 2309.154296875
+    },
+    "model.language_model.layers.15.mlp.gate_proj.weight": {
+      "global_scale": 4497.56884765625
+    },
+    "model.language_model.layers.15.mlp.up_proj.weight": {
+      "global_scale": 7399.2255859375
+    },
+    "model.language_model.layers.15.self_attn.k_proj.weight": {
+      "global_scale": 4300.80029296875
+    },
+    "model.language_model.layers.15.self_attn.o_proj.weight": {
+      "global_scale": 5252.8857421875
+    },
+    "model.language_model.layers.15.self_attn.q_proj.weight": {
+      "global_scale": 8655.6982421875
+    },
+    "model.language_model.layers.15.self_attn.v_proj.weight": {
+      "global_scale": 16581.396484375
+    },
+    "model.language_model.layers.16.mlp.down_proj.weight": {
+      "global_scale": 1988.8092041015625
+    },
+    "model.language_model.layers.16.mlp.gate_proj.weight": {
+      "global_scale": 3423.522216796875
+    },
+    "model.language_model.layers.16.mlp.up_proj.weight": {
+      "global_scale": 7603.62451171875
+    },
+    "model.language_model.layers.16.self_attn.k_proj.weight": {
+      "global_scale": 4382.98095703125
+    },
+    "model.language_model.layers.16.self_attn.o_proj.weight": {
+      "global_scale": 4300.80029296875
+    },
+    "model.language_model.layers.16.self_attn.q_proj.weight": {
+      "global_scale": 8765.9619140625
+    },
+    "model.language_model.layers.16.self_attn.v_proj.weight": {
+      "global_scale": 13694.0888671875
+    },
+    "model.language_model.layers.17.mlp.down_proj.weight": {
+      "global_scale": 2191.490478515625
+    },
+    "model.language_model.layers.17.mlp.gate_proj.weight": {
+      "global_scale": 3185.77783203125
+    },
+    "model.language_model.layers.17.mlp.up_proj.weight": {
+      "global_scale": 9830.400390625
+    },
+    "model.language_model.layers.17.self_attn.k_proj.weight": {
+      "global_scale": 4355.24072265625
+    },
+    "model.language_model.layers.17.self_attn.o_proj.weight": {
+      "global_scale": 6171.5517578125
+    },
+    "model.language_model.layers.17.self_attn.q_proj.weight": {
+      "global_scale": 11054.2646484375
+    },
+    "model.language_model.layers.17.self_attn.v_proj.weight": {
+      "global_scale": 19660.80078125
+    },
+    "model.language_model.layers.18.mlp.down_proj.weight": {
+      "global_scale": 2511.416015625
+    },
+    "model.language_model.layers.18.mlp.gate_proj.weight": {
+      "global_scale": 3142.136962890625
+    },
+    "model.language_model.layers.18.mlp.up_proj.weight": {
+      "global_scale": 5549.4189453125
+    },
+    "model.language_model.layers.18.self_attn.k_proj.weight": {
+      "global_scale": 6089.62841796875
+    },
+    "model.language_model.layers.18.self_attn.o_proj.weight": {
+      "global_scale": 6199.3515625
+    },
+    "model.language_model.layers.18.self_attn.q_proj.weight": {
+      "global_scale": 11967.443359375
+    },
+    "model.language_model.layers.18.self_attn.v_proj.weight": {
+      "global_scale": 19521.361328125
+    },
+    "model.language_model.layers.19.mlp.down_proj.weight": {
+      "global_scale": 1764.4307861328125
+    },
+    "model.language_model.layers.19.mlp.gate_proj.weight": {
+      "global_scale": 3200.59521484375
+    },
+    "model.language_model.layers.19.mlp.up_proj.weight": {
+      "global_scale": 3865.8876953125
+    },
+    "model.language_model.layers.19.self_attn.k_proj.weight": {
+      "global_scale": 4024.140380859375
+    },
+    "model.language_model.layers.19.self_attn.o_proj.weight": {
+      "global_scale": 8936.7275390625
+    },
+    "model.language_model.layers.19.self_attn.q_proj.weight": {
+      "global_scale": 10668.6513671875
+    },
+    "model.language_model.layers.19.self_attn.v_proj.weight": {
+      "global_scale": 20695.580078125
+    },
+    "model.language_model.layers.2.mlp.down_proj.weight": {
+      "global_scale": 2340.5712890625
+    },
+    "model.language_model.layers.2.mlp.gate_proj.weight": {
+      "global_scale": 7819.63671875
+    },
+    "model.language_model.layers.2.mlp.up_proj.weight": {
+      "global_scale": 6813.1484375
+    },
+    "model.language_model.layers.2.self_attn.k_proj.weight": {
+      "global_scale": 6144.00048828125
+    },
+    "model.language_model.layers.2.self_attn.o_proj.weight": {
+      "global_scale": 6461.2958984375
+    },
+    "model.language_model.layers.2.self_attn.q_proj.weight": {
+      "global_scale": 7819.63671875
+    },
+    "model.language_model.layers.2.self_attn.v_proj.weight": {
+      "global_scale": 16886.576171875
+    },
+    "model.language_model.layers.20.mlp.down_proj.weight": {
+      "global_scale": 2205.53857421875
+    },
+    "model.language_model.layers.20.mlp.gate_proj.weight": {
+      "global_scale": 3292.478271484375
+    },
+    "model.language_model.layers.20.mlp.up_proj.weight": {
+      "global_scale": 7281.77734375
+    },
+    "model.language_model.layers.20.self_attn.k_proj.weight": {
+      "global_scale": 4327.84912109375
+    },
+    "model.language_model.layers.20.self_attn.o_proj.weight": {
+      "global_scale": 6746.35302734375
+    },
+    "model.language_model.layers.20.self_attn.q_proj.weight": {
+      "global_scale": 10794.1650390625
+    },
+    "model.language_model.layers.20.self_attn.v_proj.weight": {
+      "global_scale": 18350.080078125
+    },
+    "model.language_model.layers.21.mlp.down_proj.weight": {
+      "global_scale": 2475.280517578125
+    },
+    "model.language_model.layers.21.mlp.gate_proj.weight": {
+      "global_scale": 3276.800048828125
+    },
+    "model.language_model.layers.21.mlp.up_proj.weight": {
+      "global_scale": 4986.43505859375
+    },
+    "model.language_model.layers.21.self_attn.k_proj.weight": {
+      "global_scale": 5376.0
+    },
+    "model.language_model.layers.21.self_attn.o_proj.weight": {
+      "global_scale": 4681.142578125
+    },
+    "model.language_model.layers.21.self_attn.q_proj.weight": {
+      "global_scale": 11098.837890625
+    },
+    "model.language_model.layers.21.self_attn.v_proj.weight": {
+      "global_scale": 17531.923828125
+    },
+    "model.language_model.layers.22.mlp.down_proj.weight": {
+      "global_scale": 2060.263671875
+    },
+    "model.language_model.layers.22.mlp.gate_proj.weight": {
+      "global_scale": 3493.035400390625
+    },
+    "model.language_model.layers.22.mlp.up_proj.weight": {
+      "global_scale": 6144.00048828125
+    },
+    "model.language_model.layers.22.self_attn.k_proj.weight": {
+      "global_scale": 6144.00048828125
+    },
+    "model.language_model.layers.22.self_attn.o_proj.weight": {
+      "global_scale": 5527.13232421875
+    },
+    "model.language_model.layers.22.self_attn.q_proj.weight": {
+      "global_scale": 9760.6806640625
+    },
+    "model.language_model.layers.22.self_attn.v_proj.weight": {
+      "global_scale": 15463.55078125
+    },
+    "model.language_model.layers.23.mlp.down_proj.weight": {
+      "global_scale": 2060.263671875
+    },
+    "model.language_model.layers.23.mlp.gate_proj.weight": {
+      "global_scale": 3822.933349609375
+    },
+    "model.language_model.layers.23.mlp.up_proj.weight": {
+      "global_scale": 8143.52685546875
+    },
+    "model.language_model.layers.23.self_attn.k_proj.weight": {
+      "global_scale": 5097.244140625
+    },
+    "model.language_model.layers.23.self_attn.o_proj.weight": {
+      "global_scale": 6584.95654296875
+    },
+    "model.language_model.layers.23.self_attn.q_proj.weight": {
+      "global_scale": 12233.38671875
+    },
+    "model.language_model.layers.23.self_attn.v_proj.weight": {
+      "global_scale": 13901.576171875
+    },
+    "model.language_model.layers.24.mlp.down_proj.weight": {
+      "global_scale": 3602.764404296875
+    },
+    "model.language_model.layers.24.mlp.gate_proj.weight": {
+      "global_scale": 4812.083984375
+    },
+    "model.language_model.layers.24.mlp.up_proj.weight": {
+      "global_scale": 8601.6005859375
+    },
+    "model.language_model.layers.24.self_attn.k_proj.weight": {
+      "global_scale": 5687.0078125
+    },
+    "model.language_model.layers.24.self_attn.o_proj.weight": {
+      "global_scale": 8340.9453125
+    },
+    "model.language_model.layers.24.self_attn.q_proj.weight": {
+      "global_scale": 6713.44384765625
+    },
+    "model.language_model.layers.24.self_attn.v_proj.weight": {
+      "global_scale": 15463.55078125
+    },
+    "model.language_model.layers.25.mlp.down_proj.weight": {
+      "global_scale": 1922.1451416015625
+    },
+    "model.language_model.layers.25.mlp.gate_proj.weight": {
+      "global_scale": 4950.56103515625
+    },
+    "model.language_model.layers.25.mlp.up_proj.weight": {
+      "global_scale": 6431.1025390625
+    },
+    "model.language_model.layers.25.self_attn.k_proj.weight": {
+      "global_scale": 4915.2001953125
+    },
+    "model.language_model.layers.25.self_attn.o_proj.weight": {
+      "global_scale": 5527.13232421875
+    },
+    "model.language_model.layers.25.self_attn.q_proj.weight": {
+      "global_scale": 12179.2568359375
+    },
+    "model.language_model.layers.25.self_attn.v_proj.weight": {
+      "global_scale": 17758.142578125
+    },
+    "model.language_model.layers.26.mlp.down_proj.weight": {
+      "global_scale": 2177.620361328125
+    },
+    "model.language_model.layers.26.mlp.gate_proj.weight": {
+      "global_scale": 4024.140380859375
+    },
+    "model.language_model.layers.26.mlp.up_proj.weight": {
+      "global_scale": 3440.639892578125
+    },
+    "model.language_model.layers.26.self_attn.k_proj.weight": {
+      "global_scale": 5549.4189453125
+    },
+    "model.language_model.layers.26.self_attn.o_proj.weight": {
+      "global_scale": 7320.51025390625
+    },
+    "model.language_model.layers.26.self_attn.q_proj.weight": {
+      "global_scale": 10752.0
+    },
+    "model.language_model.layers.26.self_attn.v_proj.weight": {
+      "global_scale": 18350.080078125
+    },
+    "model.language_model.layers.27.mlp.down_proj.weight": {
+      "global_scale": 1849.806396484375
+    },
+    "model.language_model.layers.27.mlp.gate_proj.weight": {
+      "global_scale": 4300.80029296875
+    },
+    "model.language_model.layers.27.mlp.up_proj.weight": {
+      "global_scale": 4024.140380859375
+    },
+    "model.language_model.layers.27.self_attn.k_proj.weight": {
+      "global_scale": 4497.56884765625
+    },
+    "model.language_model.layers.27.self_attn.o_proj.weight": {
+      "global_scale": 2567.641845703125
+    },
+    "model.language_model.layers.27.self_attn.q_proj.weight": {
+      "global_scale": 8290.6982421875
+    },
+    "model.language_model.layers.27.self_attn.v_proj.weight": {
+      "global_scale": 11613.974609375
+    },
+    "model.language_model.layers.28.mlp.down_proj.weight": {
+      "global_scale": 2493.217529296875
+    },
+    "model.language_model.layers.28.mlp.gate_proj.weight": {
+      "global_scale": 5376.0
+    },
+    "model.language_model.layers.28.mlp.up_proj.weight": {
+      "global_scale": 7094.10302734375
+    },
+    "model.language_model.layers.28.self_attn.k_proj.weight": {
+      "global_scale": 4986.43505859375
+    },
+    "model.language_model.layers.28.self_attn.o_proj.weight": {
+      "global_scale": 9362.28515625
+    },
+    "model.language_model.layers.28.self_attn.q_proj.weight": {
+      "global_scale": 11915.63671875
+    },
+    "model.language_model.layers.28.self_attn.v_proj.weight": {
+      "global_scale": 13626.296875
+    },
+    "model.language_model.layers.29.mlp.down_proj.weight": {
+      "global_scale": 1869.9130859375
+    },
+    "model.language_model.layers.29.mlp.gate_proj.weight": {
+      "global_scale": 4439.53564453125
+    },
+    "model.language_model.layers.29.mlp.up_proj.weight": {
+      "global_scale": 6847.04443359375
+    },
+    "model.language_model.layers.29.self_attn.k_proj.weight": {
+      "global_scale": 5334.32568359375
+    },
+    "model.language_model.layers.29.self_attn.o_proj.weight": {
+      "global_scale": 4327.84912109375
+    },
+    "model.language_model.layers.29.self_attn.q_proj.weight": {
+      "global_scale": 12072.4208984375
+    },
+    "model.language_model.layers.29.self_attn.v_proj.weight": {
+      "global_scale": 19114.666015625
+    },
+    "model.language_model.layers.3.mlp.down_proj.weight": {
+      "global_scale": 2163.924560546875
+    },
+    "model.language_model.layers.3.mlp.gate_proj.weight": {
+      "global_scale": 7775.45751953125
+    },
+    "model.language_model.layers.3.mlp.up_proj.weight": {
+      "global_scale": 5059.7646484375
+    },
+    "model.language_model.layers.3.self_attn.k_proj.weight": {
+      "global_scale": 6847.04443359375
+    },
+    "model.language_model.layers.3.self_attn.o_proj.weight": {
+      "global_scale": 6062.8017578125
+    },
+    "model.language_model.layers.3.self_attn.q_proj.weight": {
+      "global_scale": 10045.6640625
+    },
+    "model.language_model.layers.3.self_attn.v_proj.weight": {
+      "global_scale": 14798.451171875
+    },
+    "model.language_model.layers.30.mlp.down_proj.weight": {
+      "global_scale": 3584.0
+    },
+    "model.language_model.layers.30.mlp.gate_proj.weight": {
+      "global_scale": 5710.60595703125
+    },
+    "model.language_model.layers.30.mlp.up_proj.weight": {
+      "global_scale": 8601.6005859375
+    },
+    "model.language_model.layers.30.self_attn.k_proj.weight": {
+      "global_scale": 5213.0908203125
+    },
+    "model.language_model.layers.30.self_attn.o_proj.weight": {
+      "global_scale": 3602.764404296875
+    },
+    "model.language_model.layers.30.self_attn.q_proj.weight": {
+      "global_scale": 10752.0
+    },
+    "model.language_model.layers.30.self_attn.v_proj.weight": {
+      "global_scale": 11054.2646484375
+    },
+    "model.language_model.layers.31.mlp.down_proj.weight": {
+      "global_scale": 1977.3792724609375
+    },
+    "model.language_model.layers.31.mlp.gate_proj.weight": {
+      "global_scale": 6616.61572265625
+    },
+    "model.language_model.layers.31.mlp.up_proj.weight": {
+      "global_scale": 7205.52880859375
+    },
+    "model.language_model.layers.31.self_attn.k_proj.weight": {
+      "global_scale": 4024.140380859375
+    },
+    "model.language_model.layers.31.self_attn.o_proj.weight": {
+      "global_scale": 3760.26220703125
+    },
+    "model.language_model.layers.31.self_attn.q_proj.weight": {
+      "global_scale": 8095.62353515625
+    },
+    "model.language_model.layers.31.self_attn.v_proj.weight": {
+      "global_scale": 16287.0537109375
+    },
+    "model.language_model.layers.32.mlp.down_proj.weight": {
+      "global_scale": 1988.8092041015625
+    },
+    "model.language_model.layers.32.mlp.gate_proj.weight": {
+      "global_scale": 4880.34033203125
+    },
+    "model.language_model.layers.32.mlp.up_proj.weight": {
+      "global_scale": 6847.04443359375
+    },
+    "model.language_model.layers.32.self_attn.k_proj.weight": {
+      "global_scale": 5376.0
+    },
+    "model.language_model.layers.32.self_attn.o_proj.weight": {
+      "global_scale": 5334.32568359375
+    },
+    "model.language_model.layers.32.self_attn.q_proj.weight": {
+      "global_scale": 5594.5361328125
+    },
+    "model.language_model.layers.32.self_attn.v_proj.weight": {
+      "global_scale": 12862.205078125
+    },
+    "model.language_model.layers.33.mlp.down_proj.weight": {
+      "global_scale": 1746.5177001953125
+    },
+    "model.language_model.layers.33.mlp.gate_proj.weight": {
+      "global_scale": 4527.15771484375
+    },
+    "model.language_model.layers.33.mlp.up_proj.weight": {
+      "global_scale": 5806.9873046875
+    },
+    "model.language_model.layers.33.self_attn.k_proj.weight": {
+      "global_scale": 6062.8017578125
+    },
+    "model.language_model.layers.33.self_attn.o_proj.weight": {
+      "global_scale": 4587.52001953125
+    },
+    "model.language_model.layers.33.self_attn.q_proj.weight": {
+      "global_scale": 7688.58056640625
+    },
+    "model.language_model.layers.33.self_attn.v_proj.weight": {
+      "global_scale": 12568.5478515625
+    },
+    "model.language_model.layers.34.mlp.down_proj.weight": {
+      "global_scale": 2035.8817138671875
+    },
+    "model.language_model.layers.34.mlp.gate_proj.weight": {
+      "global_scale": 5173.89501953125
+    },
+    "model.language_model.layers.34.mlp.up_proj.weight": {
+      "global_scale": 5059.7646484375
+    },
+    "model.language_model.layers.34.self_attn.k_proj.weight": {
+      "global_scale": 4145.34912109375
+    },
+    "model.language_model.layers.34.self_attn.o_proj.weight": {
+      "global_scale": 2293.760009765625
+    },
+    "model.language_model.layers.34.self_attn.q_proj.weight": {
+      "global_scale": 9426.4111328125
+    },
+    "model.language_model.layers.34.self_attn.v_proj.weight": {
+      "global_scale": 10794.1650390625
+    },
+    "model.language_model.layers.35.mlp.down_proj.weight": {
+      "global_scale": 2023.9058837890625
+    },
+    "model.language_model.layers.35.mlp.gate_proj.weight": {
+      "global_scale": 4649.513671875
+    },
+    "model.language_model.layers.35.mlp.up_proj.weight": {
+      "global_scale": 6062.8017578125
+    },
+    "model.language_model.layers.35.self_attn.k_proj.weight": {
+      "global_scale": 4745.71044921875
+    },
+    "model.language_model.layers.35.self_attn.o_proj.weight": {
+      "global_scale": 2234.181884765625
+    },
+    "model.language_model.layers.35.self_attn.q_proj.weight": {
+      "global_scale": 7130.85986328125
+    },
+    "model.language_model.layers.35.self_attn.v_proj.weight": {
+      "global_scale": 14959.3046875
+    },
+    "model.language_model.layers.36.mlp.down_proj.weight": {
+      "global_scale": 1782.7149658203125
+    },
+    "model.language_model.layers.36.mlp.gate_proj.weight": {
+      "global_scale": 5640.39306640625
+    },
+    "model.language_model.layers.36.mlp.up_proj.weight": {
+      "global_scale": 6648.57958984375
+    },
+    "model.language_model.layers.36.self_attn.k_proj.weight": {
+      "global_scale": 5334.32568359375
+    },
+    "model.language_model.layers.36.self_attn.o_proj.weight": {
+      "global_scale": 2309.154296875
+    },
+    "model.language_model.layers.36.self_attn.q_proj.weight": {
+      "global_scale": 8241.0546875
+    },
+    "model.language_model.layers.36.self_attn.v_proj.weight": {
+      "global_scale": 17420.962890625
+    },
+    "model.language_model.layers.37.mlp.down_proj.weight": {
+      "global_scale": 1670.213623046875
+    },
+    "model.language_model.layers.37.mlp.gate_proj.weight": {
+      "global_scale": 4170.47265625
+    },
+    "model.language_model.layers.37.mlp.up_proj.weight": {
+      "global_scale": 2843.50390625
+    },
+    "model.language_model.layers.37.self_attn.k_proj.weight": {
+      "global_scale": 6813.1484375
+    },
+    "model.language_model.layers.37.self_attn.o_proj.weight": {
+      "global_scale": 4382.98095703125
+    },
+    "model.language_model.layers.37.self_attn.q_proj.weight": {
+      "global_scale": 5881.43603515625
+    },
+    "model.language_model.layers.37.self_attn.v_proj.weight": {
+      "global_scale": 14188.2060546875
+    },
+    "model.language_model.layers.38.mlp.down_proj.weight": {
+      "global_scale": 1801.3822021484375
+    },
+    "model.language_model.layers.38.mlp.gate_proj.weight": {
+      "global_scale": 4439.53564453125
+    },
+    "model.language_model.layers.38.mlp.up_proj.weight": {
+      "global_scale": 4618.30859375
+    },
+    "model.language_model.layers.38.self_attn.k_proj.weight": {
+      "global_scale": 6461.2958984375
+    },
+    "model.language_model.layers.38.self_attn.o_proj.weight": {
+      "global_scale": 1515.700439453125
+    },
+    "model.language_model.layers.38.self_attn.q_proj.weight": {
+      "global_scale": 9901.1220703125
+    },
+    "model.language_model.layers.38.self_attn.v_proj.weight": {
+      "global_scale": 15728.6396484375
+    },
+    "model.language_model.layers.39.mlp.down_proj.weight": {
+      "global_scale": 1810.86328125
+    },
+    "model.language_model.layers.39.mlp.gate_proj.weight": {
+      "global_scale": 3640.888671875
+    },
+    "model.language_model.layers.39.mlp.up_proj.weight": {
+      "global_scale": 5293.29248046875
+    },
+    "model.language_model.layers.39.self_attn.k_proj.weight": {
+      "global_scale": 6915.859375
+    },
+    "model.language_model.layers.39.self_attn.o_proj.weight": {
+      "global_scale": 2667.162841796875
+    },
+    "model.language_model.layers.39.self_attn.q_proj.weight": {
+      "global_scale": 7205.52880859375
+    },
+    "model.language_model.layers.39.self_attn.v_proj.weight": {
+      "global_scale": 13626.296875
+    },
+    "model.language_model.layers.4.mlp.down_proj.weight": {
+      "global_scale": 2493.217529296875
+    },
+    "model.language_model.layers.4.mlp.gate_proj.weight": {
+      "global_scale": 7057.72314453125
+    },
+    "model.language_model.layers.4.mlp.up_proj.weight": {
+      "global_scale": 5376.0
+    },
+    "model.language_model.layers.4.self_attn.k_proj.weight": {
+      "global_scale": 7094.10302734375
+    },
+    "model.language_model.layers.4.self_attn.o_proj.weight": {
+      "global_scale": 5831.59326171875
+    },
+    "model.language_model.layers.4.self_attn.q_proj.weight": {
+      "global_scale": 10922.6669921875
+    },
+    "model.language_model.layers.4.self_attn.v_proj.weight": {
+      "global_scale": 16681.890625
+    },
+    "model.language_model.layers.40.mlp.down_proj.weight": {
+      "global_scale": 2110.822021484375
+    },
+    "model.language_model.layers.40.mlp.gate_proj.weight": {
+      "global_scale": 4915.2001953125
+    },
+    "model.language_model.layers.40.mlp.up_proj.weight": {
+      "global_scale": 6144.00048828125
+    },
+    "model.language_model.layers.40.self_attn.k_proj.weight": {
+      "global_scale": 6847.04443359375
+    },
+    "model.language_model.layers.40.self_attn.o_proj.weight": {
+      "global_scale": 2150.400146484375
+    },
+    "model.language_model.layers.40.self_attn.q_proj.weight": {
+      "global_scale": 7731.775390625
+    },
+    "model.language_model.layers.40.self_attn.v_proj.weight": {
+      "global_scale": 15639.2734375
+    },
+    "model.language_model.layers.41.mlp.down_proj.weight": {
+      "global_scale": 1773.5257568359375
+    },
+    "model.language_model.layers.41.mlp.gate_proj.weight": {
+      "global_scale": 3780.923095703125
+    },
+    "model.language_model.layers.41.mlp.up_proj.weight": {
+      "global_scale": 5173.89501953125
+    },
+    "model.language_model.layers.41.self_attn.k_proj.weight": {
+      "global_scale": 6813.1484375
+    },
+    "model.language_model.layers.41.self_attn.o_proj.weight": {
+      "global_scale": 5687.0078125
+    },
+    "model.language_model.layers.41.self_attn.q_proj.weight": {
+      "global_scale": 7168.0
+    },
+    "model.language_model.layers.41.self_attn.v_proj.weight": {
+      "global_scale": 14188.2060546875
+    },
+    "model.language_model.layers.42.mlp.down_proj.weight": {
+      "global_scale": 2191.490478515625
+    },
+    "model.language_model.layers.42.mlp.gate_proj.weight": {
+      "global_scale": 4221.64404296875
+    },
+    "model.language_model.layers.42.mlp.up_proj.weight": {
+      "global_scale": 5173.89501953125
+    },
+    "model.language_model.layers.42.self_attn.k_proj.weight": {
+      "global_scale": 9426.4111328125
+    },
+    "model.language_model.layers.42.self_attn.o_proj.weight": {
+      "global_scale": 3018.105224609375
+    },
+    "model.language_model.layers.42.self_attn.q_proj.weight": {
+      "global_scale": 6371.5556640625
+    },
+    "model.language_model.layers.42.self_attn.v_proj.weight": {
+      "global_scale": 11864.275390625
+    },
+    "model.language_model.layers.43.mlp.down_proj.weight": {
+      "global_scale": 1162.37841796875
+    },
+    "model.language_model.layers.43.mlp.gate_proj.weight": {
+      "global_scale": 4096.0
+    },
+    "model.language_model.layers.43.mlp.up_proj.weight": {
+      "global_scale": 1439.5982666015625
+    },
+    "model.language_model.layers.43.self_attn.k_proj.weight": {
+      "global_scale": 8340.9453125
+    },
+    "model.language_model.layers.43.self_attn.o_proj.weight": {
+      "global_scale": 3185.77783203125
+    },
+    "model.language_model.layers.43.self_attn.q_proj.weight": {
+      "global_scale": 7688.58056640625
+    },
+    "model.language_model.layers.43.self_attn.v_proj.weight": {
+      "global_scale": 8879.0712890625
+    },
+    "model.language_model.layers.44.mlp.down_proj.weight": {
+      "global_scale": 1943.8643798828125
+    },
+    "model.language_model.layers.44.mlp.gate_proj.weight": {
+      "global_scale": 3292.478271484375
+    },
+    "model.language_model.layers.44.mlp.up_proj.weight": {
+      "global_scale": 6491.7734375
+    },
+    "model.language_model.layers.44.self_attn.k_proj.weight": {
+      "global_scale": 6915.859375
+    },
+    "model.language_model.layers.44.self_attn.o_proj.weight": {
+      "global_scale": 2356.602783203125
+    },
+    "model.language_model.layers.44.self_attn.q_proj.weight": {
+      "global_scale": 6713.44384765625
+    },
+    "model.language_model.layers.44.self_attn.v_proj.weight": {
+      "global_scale": 16581.396484375
+    },
+    "model.language_model.layers.45.mlp.down_proj.weight": {
+      "global_scale": 1943.8643798828125
+    },
+    "model.language_model.layers.45.mlp.gate_proj.weight": {
+      "global_scale": 4071.763427734375
+    },
+    "model.language_model.layers.45.mlp.up_proj.weight": {
+      "global_scale": 3200.59521484375
+    },
+    "model.language_model.layers.45.self_attn.k_proj.weight": {
+      "global_scale": 5376.0
+    },
+    "model.language_model.layers.45.self_attn.o_proj.weight": {
+      "global_scale": 2035.8817138671875
+    },
+    "model.language_model.layers.45.self_attn.q_proj.weight": {
+      "global_scale": 7561.84619140625
+    },
+    "model.language_model.layers.45.self_attn.v_proj.weight": {
+      "global_scale": 13972.1416015625
+    },
+    "model.language_model.layers.46.mlp.down_proj.weight": {
+      "global_scale": 1859.805419921875
+    },
+    "model.language_model.layers.46.mlp.gate_proj.weight": {
+      "global_scale": 3719.61083984375
+    },
+    "model.language_model.layers.46.mlp.up_proj.weight": {
+      "global_scale": 5906.67822265625
+    },
+    "model.language_model.layers.46.self_attn.k_proj.weight": {
+      "global_scale": 10794.1650390625
+    },
+    "model.language_model.layers.46.self_attn.o_proj.weight": {
+      "global_scale": 3440.639892578125
+    },
+    "model.language_model.layers.46.self_attn.q_proj.weight": {
+      "global_scale": 7021.71435546875
+    },
+    "model.language_model.layers.46.self_attn.v_proj.weight": {
+      "global_scale": 12019.703125
+    },
+    "model.language_model.layers.47.mlp.down_proj.weight": {
+      "global_scale": 2048.0
+    },
+    "model.language_model.layers.47.mlp.gate_proj.weight": {
+      "global_scale": 4120.52734375
+    },
+    "model.language_model.layers.47.mlp.up_proj.weight": {
+      "global_scale": 5710.60595703125
+    },
+    "model.language_model.layers.47.self_attn.k_proj.weight": {
+      "global_scale": 9624.16796875
+    },
+    "model.language_model.layers.47.self_attn.o_proj.weight": {
+      "global_scale": 5252.8857421875
+    },
+    "model.language_model.layers.47.self_attn.q_proj.weight": {
+      "global_scale": 7021.71435546875
+    },
+    "model.language_model.layers.47.self_attn.v_proj.weight": {
+      "global_scale": 9624.16796875
+    },
+    "model.language_model.layers.48.mlp.down_proj.weight": {
+      "global_scale": 2763.566162109375
+    },
+    "model.language_model.layers.48.mlp.gate_proj.weight": {
+      "global_scale": 3127.8544921875
+    },
+    "model.language_model.layers.48.mlp.up_proj.weight": {
+      "global_scale": 6553.60009765625
+    },
+    "model.language_model.layers.48.self_attn.k_proj.weight": {
+      "global_scale": 9175.0400390625
+    },
+    "model.language_model.layers.48.self_attn.o_proj.weight": {
+      "global_scale": 2548.6220703125
+    },
+    "model.language_model.layers.48.self_attn.q_proj.weight": {
+      "global_scale": 8048.28076171875
+    },
+    "model.language_model.layers.48.self_attn.v_proj.weight": {
+      "global_scale": 11967.443359375
+    },
+    "model.language_model.layers.49.mlp.down_proj.weight": {
+      "global_scale": 2719.87353515625
+    },
+    "model.language_model.layers.49.mlp.gate_proj.weight": {
+      "global_scale": 3142.136962890625
+    },
+    "model.language_model.layers.49.mlp.up_proj.weight": {
+      "global_scale": 5640.39306640625
+    },
+    "model.language_model.layers.49.self_attn.k_proj.weight": {
+      "global_scale": 9760.6806640625
+    },
+    "model.language_model.layers.49.self_attn.o_proj.weight": {
+      "global_scale": 3865.8876953125
+    },
+    "model.language_model.layers.49.self_attn.q_proj.weight": {
+      "global_scale": 10426.181640625
+    },
+    "model.language_model.layers.49.self_attn.v_proj.weight": {
+      "global_scale": 13297.1591796875
+    },
+    "model.language_model.layers.5.mlp.down_proj.weight": {
+      "global_scale": 3389.79296875
+    },
+    "model.language_model.layers.5.mlp.gate_proj.weight": {
+      "global_scale": 7479.65234375
+    },
+    "model.language_model.layers.5.mlp.up_proj.weight": {
+      "global_scale": 4300.80029296875
+    },
+    "model.language_model.layers.5.self_attn.k_proj.weight": {
+      "global_scale": 6950.7880859375
+    },
+    "model.language_model.layers.5.self_attn.o_proj.weight": {
+      "global_scale": 5549.4189453125
+    },
+    "model.language_model.layers.5.self_attn.q_proj.weight": {
+      "global_scale": 11374.015625
+    },
+    "model.language_model.layers.5.self_attn.v_proj.weight": {
+      "global_scale": 14563.5546875
+    },
+    "model.language_model.layers.6.mlp.down_proj.weight": {
+      "global_scale": 1131.7894287109375
+    },
+    "model.language_model.layers.6.mlp.gate_proj.weight": {
+      "global_scale": 6009.8515625
+    },
+    "model.language_model.layers.6.mlp.up_proj.weight": {
+      "global_scale": 3004.92578125
+    },
+    "model.language_model.layers.6.self_attn.k_proj.weight": {
+      "global_scale": 8340.9453125
+    },
+    "model.language_model.layers.6.self_attn.o_proj.weight": {
+      "global_scale": 3932.159912109375
+    },
+    "model.language_model.layers.6.self_attn.q_proj.weight": {
+      "global_scale": 8548.173828125
+    },
+    "model.language_model.layers.6.self_attn.v_proj.weight": {
+      "global_scale": 16096.5615234375
+    },
+    "model.language_model.layers.7.mlp.down_proj.weight": {
+      "global_scale": 2110.822021484375
+    },
+    "model.language_model.layers.7.mlp.gate_proj.weight": {
+      "global_scale": 4986.43505859375
+    },
+    "model.language_model.layers.7.mlp.up_proj.weight": {
+      "global_scale": 2548.6220703125
+    },
+    "model.language_model.layers.7.self_attn.k_proj.weight": {
+      "global_scale": 6401.1904296875
+    },
+    "model.language_model.layers.7.self_attn.o_proj.weight": {
+      "global_scale": 2567.641845703125
+    },
+    "model.language_model.layers.7.self_attn.q_proj.weight": {
+      "global_scale": 9236.6171875
+    },
+    "model.language_model.layers.7.self_attn.v_proj.weight": {
+      "global_scale": 16681.890625
+    },
+    "model.language_model.layers.8.mlp.down_proj.weight": {
+      "global_scale": 2278.569580078125
+    },
+    "model.language_model.layers.8.mlp.gate_proj.weight": {
+      "global_scale": 5983.7216796875
+    },
+    "model.language_model.layers.8.mlp.up_proj.weight": {
+      "global_scale": 9362.28515625
+    },
+    "model.language_model.layers.8.self_attn.k_proj.weight": {
+      "global_scale": 4845.9716796875
+    },
+    "model.language_model.layers.8.self_attn.o_proj.weight": {
+      "global_scale": 3679.828857421875
+    },
+    "model.language_model.layers.8.self_attn.q_proj.weight": {
+      "global_scale": 6779.5859375
+    },
+    "model.language_model.layers.8.self_attn.v_proj.weight": {
+      "global_scale": 13361.708984375
+    },
+    "model.language_model.layers.9.mlp.down_proj.weight": {
+      "global_scale": 1670.213623046875
+    },
+    "model.language_model.layers.9.mlp.gate_proj.weight": {
+      "global_scale": 4411.0771484375
+    },
+    "model.language_model.layers.9.mlp.up_proj.weight": {
+      "global_scale": 6986.07080078125
+    },
+    "model.language_model.layers.9.self_attn.k_proj.weight": {
+      "global_scale": 3932.159912109375
+    },
+    "model.language_model.layers.9.self_attn.o_proj.weight": {
+      "global_scale": 2831.802490234375
+    },
+    "model.language_model.layers.9.self_attn.q_proj.weight": {
+      "global_scale": 9299.02734375
+    },
+    "model.language_model.layers.9.self_attn.v_proj.weight": {
+      "global_scale": 15291.7333984375
+    }
+  },
+  "dense_keys": [
+    "model.language_model.embed_tokens.weight",
+    "model.language_model.layers.0.input_layernorm.weight",
+    "model.language_model.layers.0.post_attention_layernorm.weight",
+    "model.language_model.layers.0.self_attn.k_norm.weight",
+    "model.language_model.layers.0.self_attn.q_norm.weight",
+    "model.language_model.layers.1.input_layernorm.weight",
+    "model.language_model.layers.1.post_attention_layernorm.weight",
+    "model.language_model.layers.1.self_attn.k_norm.weight",
+    "model.language_model.layers.1.self_attn.q_norm.weight",
+    "model.language_model.layers.10.input_layernorm.weight",
+    "model.language_model.layers.10.post_attention_layernorm.weight",
+    "model.language_model.layers.10.self_attn.k_norm.weight",
+    "model.language_model.layers.10.self_attn.q_norm.weight",
+    "model.language_model.layers.11.input_layernorm.weight",
+    "model.language_model.layers.11.post_attention_layernorm.weight",
+    "model.language_model.layers.11.self_attn.k_norm.weight",
+    "model.language_model.layers.11.self_attn.q_norm.weight",
+    "model.language_model.layers.12.input_layernorm.weight",
+    "model.language_model.layers.12.post_attention_layernorm.weight",
+    "model.language_model.layers.12.self_attn.k_norm.weight",
+    "model.language_model.layers.12.self_attn.q_norm.weight",
+    "model.language_model.layers.13.input_layernorm.weight",
+    "model.language_model.layers.13.post_attention_layernorm.weight",
+    "model.language_model.layers.13.self_attn.k_norm.weight",
+    "model.language_model.layers.13.self_attn.q_norm.weight",
+    "model.language_model.layers.14.input_layernorm.weight",
+    "model.language_model.layers.14.post_attention_layernorm.weight",
+    "model.language_model.layers.14.self_attn.k_norm.weight",
+    "model.language_model.layers.14.self_attn.q_norm.weight",
+    "model.language_model.layers.15.input_layernorm.weight",
+    "model.language_model.layers.15.post_attention_layernorm.weight",
+    "model.language_model.layers.15.self_attn.k_norm.weight",
+    "model.language_model.layers.15.self_attn.q_norm.weight",
+    "model.language_model.layers.16.input_layernorm.weight",
+    "model.language_model.layers.16.post_attention_layernorm.weight",
+    "model.language_model.layers.16.self_attn.k_norm.weight",
+    "model.language_model.layers.16.self_attn.q_norm.weight",
+    "model.language_model.layers.17.input_layernorm.weight",
+    "model.language_model.layers.17.post_attention_layernorm.weight",
+    "model.language_model.layers.17.self_attn.k_norm.weight",
+    "model.language_model.layers.17.self_attn.q_norm.weight",
+    "model.language_model.layers.18.input_layernorm.weight",
+    "model.language_model.layers.18.post_attention_layernorm.weight",
+    "model.language_model.layers.18.self_attn.k_norm.weight",
+    "model.language_model.layers.18.self_attn.q_norm.weight",
+    "model.language_model.layers.19.input_layernorm.weight",
+    "model.language_model.layers.19.post_attention_layernorm.weight",
+    "model.language_model.layers.19.self_attn.k_norm.weight",
+    "model.language_model.layers.19.self_attn.q_norm.weight",
+    "model.language_model.layers.2.input_layernorm.weight",
+    "model.language_model.layers.2.post_attention_layernorm.weight",
+    "model.language_model.layers.2.self_attn.k_norm.weight",
+    "model.language_model.layers.2.self_attn.q_norm.weight",
+    "model.language_model.layers.20.input_layernorm.weight",
+    "model.language_model.layers.20.post_attention_layernorm.weight",
+    "model.language_model.layers.20.self_attn.k_norm.weight",
+    "model.language_model.layers.20.self_attn.q_norm.weight",
+    "model.language_model.layers.21.input_layernorm.weight",
+    "model.language_model.layers.21.post_attention_layernorm.weight",
+    "model.language_model.layers.21.self_attn.k_norm.weight",
+    "model.language_model.layers.21.self_attn.q_norm.weight",
+    "model.language_model.layers.22.input_layernorm.weight",
+    "model.language_model.layers.22.post_attention_layernorm.weight",
+    "model.language_model.layers.22.self_attn.k_norm.weight",
+    "model.language_model.layers.22.self_attn.q_norm.weight",
+    "model.language_model.layers.23.input_layernorm.weight",
+    "model.language_model.layers.23.post_attention_layernorm.weight",
+    "model.language_model.layers.23.self_attn.k_norm.weight",
+    "model.language_model.layers.23.self_attn.q_norm.weight",
+    "model.language_model.layers.24.input_layernorm.weight",
+    "model.language_model.layers.24.post_attention_layernorm.weight",
+    "model.language_model.layers.24.self_attn.k_norm.weight",
+    "model.language_model.layers.24.self_attn.q_norm.weight",
+    "model.language_model.layers.25.input_layernorm.weight",
+    "model.language_model.layers.25.post_attention_layernorm.weight",
+    "model.language_model.layers.25.self_attn.k_norm.weight",
+    "model.language_model.layers.25.self_attn.q_norm.weight",
+    "model.language_model.layers.26.input_layernorm.weight",
+    "model.language_model.layers.26.post_attention_layernorm.weight",
+    "model.language_model.layers.26.self_attn.k_norm.weight",
+    "model.language_model.layers.26.self_attn.q_norm.weight",
+    "model.language_model.layers.27.input_layernorm.weight",
+    "model.language_model.layers.27.post_attention_layernorm.weight",
+    "model.language_model.layers.27.self_attn.k_norm.weight",
+    "model.language_model.layers.27.self_attn.q_norm.weight",
+    "model.language_model.layers.28.input_layernorm.weight",
+    "model.language_model.layers.28.post_attention_layernorm.weight",
+    "model.language_model.layers.28.self_attn.k_norm.weight",
+    "model.language_model.layers.28.self_attn.q_norm.weight",
+    "model.language_model.layers.29.input_layernorm.weight",
+    "model.language_model.layers.29.post_attention_layernorm.weight",
+    "model.language_model.layers.29.self_attn.k_norm.weight",
+    "model.language_model.layers.29.self_attn.q_norm.weight",
+    "model.language_model.layers.3.input_layernorm.weight",
+    "model.language_model.layers.3.post_attention_layernorm.weight",
+    "model.language_model.layers.3.self_attn.k_norm.weight",
+    "model.language_model.layers.3.self_attn.q_norm.weight",
+    "model.language_model.layers.30.input_layernorm.weight",
+    "model.language_model.layers.30.post_attention_layernorm.weight",
+    "model.language_model.layers.30.self_attn.k_norm.weight",
+    "model.language_model.layers.30.self_attn.q_norm.weight",
+    "model.language_model.layers.31.input_layernorm.weight",
+    "model.language_model.layers.31.post_attention_layernorm.weight",
+    "model.language_model.layers.31.self_attn.k_norm.weight",
+    "model.language_model.layers.31.self_attn.q_norm.weight",
+    "model.language_model.layers.32.input_layernorm.weight",
+    "model.language_model.layers.32.post_attention_layernorm.weight",
+    "model.language_model.layers.32.self_attn.k_norm.weight",
+    "model.language_model.layers.32.self_attn.q_norm.weight",
+    "model.language_model.layers.33.input_layernorm.weight",
+    "model.language_model.layers.33.post_attention_layernorm.weight",
+    "model.language_model.layers.33.self_attn.k_norm.weight",
+    "model.language_model.layers.33.self_attn.q_norm.weight",
+    "model.language_model.layers.34.input_layernorm.weight",
+    "model.language_model.layers.34.post_attention_layernorm.weight",
+    "model.language_model.layers.34.self_attn.k_norm.weight",
+    "model.language_model.layers.34.self_attn.q_norm.weight",
+    "model.language_model.layers.35.input_layernorm.weight",
+    "model.language_model.layers.35.post_attention_layernorm.weight",
+    "model.language_model.layers.35.self_attn.k_norm.weight",
+    "model.language_model.layers.35.self_attn.q_norm.weight",
+    "model.language_model.layers.36.input_layernorm.weight",
+    "model.language_model.layers.36.post_attention_layernorm.weight",
+    "model.language_model.layers.36.self_attn.k_norm.weight",
+    "model.language_model.layers.36.self_attn.q_norm.weight",
+    "model.language_model.layers.37.input_layernorm.weight",
+    "model.language_model.layers.37.post_attention_layernorm.weight",
+    "model.language_model.layers.37.self_attn.k_norm.weight",
+    "model.language_model.layers.37.self_attn.q_norm.weight",
+    "model.language_model.layers.38.input_layernorm.weight",
+    "model.language_model.layers.38.post_attention_layernorm.weight",
+    "model.language_model.layers.38.self_attn.k_norm.weight",
+    "model.language_model.layers.38.self_attn.q_norm.weight",
+    "model.language_model.layers.39.input_layernorm.weight",
+    "model.language_model.layers.39.post_attention_layernorm.weight",
+    "model.language_model.layers.39.self_attn.k_norm.weight",
+    "model.language_model.layers.39.self_attn.q_norm.weight",
+    "model.language_model.layers.4.input_layernorm.weight",
+    "model.language_model.layers.4.post_attention_layernorm.weight",
+    "model.language_model.layers.4.self_attn.k_norm.weight",
+    "model.language_model.layers.4.self_attn.q_norm.weight",
+    "model.language_model.layers.40.input_layernorm.weight",
+    "model.language_model.layers.40.post_attention_layernorm.weight",
+    "model.language_model.layers.40.self_attn.k_norm.weight",
+    "model.language_model.layers.40.self_attn.q_norm.weight",
+    "model.language_model.layers.41.input_layernorm.weight",
+    "model.language_model.layers.41.post_attention_layernorm.weight",
+    "model.language_model.layers.41.self_attn.k_norm.weight",
+    "model.language_model.layers.41.self_attn.q_norm.weight",
+    "model.language_model.layers.42.input_layernorm.weight",
+    "model.language_model.layers.42.post_attention_layernorm.weight",
+    "model.language_model.layers.42.self_attn.k_norm.weight",
+    "model.language_model.layers.42.self_attn.q_norm.weight",
+    "model.language_model.layers.43.input_layernorm.weight",
+    "model.language_model.layers.43.post_attention_layernorm.weight",
+    "model.language_model.layers.43.self_attn.k_norm.weight",
+    "model.language_model.layers.43.self_attn.q_norm.weight",
+    "model.language_model.layers.44.input_layernorm.weight",
+    "model.language_model.layers.44.post_attention_layernorm.weight",
+    "model.language_model.layers.44.self_attn.k_norm.weight",
+    "model.language_model.layers.44.self_attn.q_norm.weight",
+    "model.language_model.layers.45.input_layernorm.weight",
+    "model.language_model.layers.45.post_attention_layernorm.weight",
+    "model.language_model.layers.45.self_attn.k_norm.weight",
+    "model.language_model.layers.45.self_attn.q_norm.weight",
+    "model.language_model.layers.46.input_layernorm.weight",
+    "model.language_model.layers.46.post_attention_layernorm.weight",
+    "model.language_model.layers.46.self_attn.k_norm.weight",
+    "model.language_model.layers.46.self_attn.q_norm.weight",
+    "model.language_model.layers.47.input_layernorm.weight",
+    "model.language_model.layers.47.post_attention_layernorm.weight",
+    "model.language_model.layers.47.self_attn.k_norm.weight",
+    "model.language_model.layers.47.self_attn.q_norm.weight",
+    "model.language_model.layers.48.input_layernorm.weight",
+    "model.language_model.layers.48.post_attention_layernorm.weight",
+    "model.language_model.layers.48.self_attn.k_norm.weight",
+    "model.language_model.layers.48.self_attn.q_norm.weight",
+    "model.language_model.layers.49.input_layernorm.weight",
+    "model.language_model.layers.49.post_attention_layernorm.weight",
+    "model.language_model.layers.49.self_attn.k_norm.weight",
+    "model.language_model.layers.49.self_attn.q_norm.weight",
+    "model.language_model.layers.5.input_layernorm.weight",
+    "model.language_model.layers.5.post_attention_layernorm.weight",
+    "model.language_model.layers.5.self_attn.k_norm.weight",
+    "model.language_model.layers.5.self_attn.q_norm.weight",
+    "model.language_model.layers.6.input_layernorm.weight",
+    "model.language_model.layers.6.post_attention_layernorm.weight",
+    "model.language_model.layers.6.self_attn.k_norm.weight",
+    "model.language_model.layers.6.self_attn.q_norm.weight",
+    "model.language_model.layers.7.input_layernorm.weight",
+    "model.language_model.layers.7.post_attention_layernorm.weight",
+    "model.language_model.layers.7.self_attn.k_norm.weight",
+    "model.language_model.layers.7.self_attn.q_norm.weight",
+    "model.language_model.layers.8.input_layernorm.weight",
+    "model.language_model.layers.8.post_attention_layernorm.weight",
+    "model.language_model.layers.8.self_attn.k_norm.weight",
+    "model.language_model.layers.8.self_attn.q_norm.weight",
+    "model.language_model.layers.9.input_layernorm.weight",
+    "model.language_model.layers.9.post_attention_layernorm.weight",
+    "model.language_model.layers.9.self_attn.k_norm.weight",
+    "model.language_model.layers.9.self_attn.q_norm.weight"
+  ],
+  "source_dir": "/Users/aryank/os/FastH3-NVFP4-encoder"
+}

text_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7d75934b4f730012a26ce14dfe228b645187c403a35215c16b6493a9cdc1490
+size 15271450842

tokenizer/tokenizer_config.json

@@ -0,0 +1,246 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>",
+    "<d>",
+    "</d>",
+    "<|cutoff|>",
+    "<|lyrics_start|>",
+    "<|lyrics_end|>",
+    "<|caption_start|>",
+    "<|caption_end|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {%- if messages[0].content is string %}\n            {{- messages[0].content }}\n        {%- else %}\n            {%- for content in messages[0].content %}\n                {%- if 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' }}\n        {%- if messages[0].content is string %}\n            {{- messages[0].content }}\n        {%- else %}\n            {%- for content in messages[0].content %}\n                {%- if 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set image_count = namespace(value=0) %}\n{%- set video_count = namespace(value=0) %}\n{%- for message in messages %}\n    {%- if message.role == \"user\" %}\n        {{- '<|im_start|>' + message.role + '\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content in message.content %}\n                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}\n                    {%- set image_count.value = image_count.value + 1 %}\n                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}\n                    <|vision_start|><|image_pad|><|vision_end|>\n                {%- elif content.type == 'video' or 'video' in content %}\n                    {%- set video_count.value = video_count.value + 1 %}\n                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}\n                    <|vision_start|><|video_pad|><|vision_end|>\n                {%- elif 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role + '\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content_item in message.content %}\n                {%- if 'text' in content_item %}\n                    {{- content_item.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and message.content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {%- if message.content is string %}\n            {{- message.content }}\n        {%- else %}\n            {%- for content in message.content %}\n                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}\n                    {%- set image_count.value = image_count.value + 1 %}\n                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}\n                    <|vision_start|><|image_pad|><|vision_end|>\n                {%- elif content.type == 'video' or 'video' in content %}\n                    {%- set video_count.value = video_count.value + 1 %}\n                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}\n                    <|vision_start|><|video_pad|><|vision_end|>\n                {%- elif 'text' in content %}\n                    {{- content.text }}\n                {%- endif %}\n            {%- endfor %}\n        {%- endif %}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

transformer/mlx_h3_dit.json

@@ -0,0 +1,1338 @@
+{
+  "format_version": 1,
+  "config": {
+    "_class_name": "MiniMaxH3Transformer3DModel",
+    "_diffusers_version": "0.36.0.dev0",
+    "num_attention_heads": 56,
+    "attention_head_dim": 128,
+    "hidden_size": 5376,
+    "num_layers": 42,
+    "num_refiner_layers": 2,
+    "ffn_dim": 14336,
+    "in_channels": 24,
+    "audio_in_channels": 32,
+    "patch_size": [
+      1,
+      2,
+      2
+    ],
+    "text_dim": 5120,
+    "freq_dim": 256,
+    "time_embed_hidden_dim": 5376,
+    "time_embed_dim": 2688,
+    "rope_freq_dim": 16,
+    "rope_theta": 10000.0,
+    "norm_eps": 1e-05,
+    "qk_norm_eps": 1e-05,
+    "final_norm_eps": 1e-05,
+    "adaln_rank": 16,
+    "source_num_layers": 50,
+    "kept_block_indices": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      5,
+      8,
+      10,
+      11,
+      12,
+      14,
+      17,
+      18,
+      19,
+      20,
+      21,
+      24,
+      25,
+      26,
+      27,
+      28,
+      29,
+      30,
+      31,
+      32,
+      33,
+      34,
+      35,
+      36,
+      37,
+      38,
+      39,
+      40,
+      41,
+      42,
+      43,
+      44,
+      45,
+      46,
+      47,
+      48,
+      49
+    ]
+  },
+  "num_blocks": 42,
+  "num_refiner_blocks": 2,
+  "quantization": {
+    "mode": "affine",
+    "bits": 6,
+    "group_size": 64
+  },
+  "quantized_keys": {
+    "blocks.0.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.0.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.0.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.0.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.0.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.0.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.0.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.1.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.2.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.3.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.4.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.5.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.6.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.7.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.8.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.9.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.10.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.11.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.12.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.13.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.14.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.15.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.16.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.17.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.18.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.19.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.20.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.21.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.22.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.23.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.24.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.25.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.26.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.27.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.28.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.29.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.30.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.31.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.32.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.33.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.34.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.35.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.36.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.37.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.38.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.39.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.40.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.attn.to_gate_compress.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "blocks.41.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.0.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.0.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.0.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.0.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.0.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.0.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.1.attn.to_q.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.1.ff.net.2.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.1.attn.to_v.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.1.attn.to_out.0.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.1.attn.to_k.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    },
+    "refiner.1.ff.net.0.proj.weight": {
+      "dequantized_dtype": "bf16",
+      "has_biases": true
+    }
+  },
+  "adaln_cache": {
+    "timesteps": [
+      0.00010009008110500872,
+      0.00033355571213178337,
+      0.014211595058441162,
+      0.03242475166916847,
+      0.04585152864456177,
+      0.05683192238211632,
+      0.09090909361839294,
+      0.10048038512468338,
+      0.1428571492433548,
+      0.16725978255271912,
+      0.23076923191547394,
+      0.25,
+      0.3571428656578064,
+      0.4117647111415863,
+      0.5,
+      0.699999988079071,
+      1.0
+    ],
+    "num_blocks": 42,
+    "tables_per_block": 6
+  },
+  "vsa": {
+    "capable": true,
+    "num_gate_matrices": 42,
+    "gate_key_suffix": "attn.to_gate_compress.weight",
+    "attention_activations": "bf16",
+    "note": "Gate projections are quantized with the checkpoint's affine INT8/INT6/INT4 weight-only grid. VSA attention Q/K/V remain unquantized activations."
+  }
+}

transformer/mlx_h3_dit.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0783bb0ca0b9801201b9aa5ded895670ec4ee79abd285fe536625e2d29059aea
+size 15355424607

vae/config.json

@@ -0,0 +1,97 @@
+{
+  "_class_name": "AutoencoderKLMiniMaxH3",
+  "_diffusers_version": "0.36.0.dev0",
+  "in_channels": 3,
+  "out_channels": 3,
+  "latent_channels": 24,
+  "block_out_channels": [
+    128,
+    256,
+    256,
+    512,
+    512,
+    1024
+  ],
+  "layers_per_block": 2,
+  "spatial_downsample_factors": [
+    2,
+    2,
+    2,
+    2,
+    1,
+    1
+  ],
+  "temporal_downsample_factors": [
+    1,
+    2,
+    2,
+    1,
+    1,
+    1
+  ],
+  "norm_num_groups": 32,
+  "norm_eps": 1e-06,
+  "spatial_padding_mode": "reflect",
+  "decoder_num_layers": 26,
+  "decoder_num_attention_heads": 32,
+  "decoder_attention_head_dim": 64,
+  "decoder_num_register_tokens": 4,
+  "decoder_ffn_mult": 4,
+  "decoder_rope_theta": 100.0,
+  "decoder_rope_dim_ratio": 0.75,
+  "decoder_norm_eps": 1e-05,
+  "clip_length": 17,
+  "token_drop": 3,
+  "latents_mean": [
+    0.858090341091156,
+    -0.9606591463088989,
+    1.0661640167236328,
+    -0.5090325474739075,
+    -0.2727581858634949,
+    -1.3675414323806763,
+    -0.2553254961967468,
+    -0.26907554268836975,
+    -0.5376840829849243,
+    -0.0464097298681736,
+    0.6657370328903198,
+    0.19690127670764923,
+    -0.5460608005523682,
+    -0.4035342037677765,
+    -0.23683024942874908,
+    0.25928452610969543,
+    -0.30133944749832153,
+    0.211341992020607,
+    -1.1206848621368408,
+    0.3581933379173279,
+    -0.04225143790245056,
+    0.2604829967021942,
+    0.22864092886447906,
+    0.7056031823158264
+  ],
+  "latents_std": [
+    1.2223774194717407,
+    1.2767263650894165,
+    1.6831774711608887,
+    1.7549455165863037,
+    1.5636216402053833,
+    2.194143533706665,
+    0.9653137922286987,
+    1.0569885969161987,
+    0.841948926448822,
+    0.7729952931404114,
+    1.8955937623977661,
+    0.946841835975647,
+    0.7996809482574463,
+    0.44988900423049927,
+    0.7197399735450745,
+    0.6936293244361877,
+    2.961095094680786,
+    2.7694199085235596,
+    3.0496184825897217,
+    2.1088054180145264,
+    3.276226282119751,
+    3.1627357006073,
+    2.2816812992095947,
+    2.6127843856811523
+  ]
+}

vae/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73fc5a61caeba438633eb8f8efd3a57f4915adfa46731a1c991312c724a3b5cc
+size 4225603880