Image-Text-to-Text
Transformers
Safetensors
English
nemotron_h_omni
conversational
custom_code
modelopt
Instructions to use MultiverseComputingCAI/Comet_60B_A12B_FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use MultiverseComputingCAI/Comet_60B_A12B_FP8 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="MultiverseComputingCAI/Comet_60B_A12B_FP8", trust_remote_code=True) messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForMultimodalLM model = AutoModelForMultimodalLM.from_pretrained("MultiverseComputingCAI/Comet_60B_A12B_FP8", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use MultiverseComputingCAI/Comet_60B_A12B_FP8 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "MultiverseComputingCAI/Comet_60B_A12B_FP8" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "MultiverseComputingCAI/Comet_60B_A12B_FP8", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/MultiverseComputingCAI/Comet_60B_A12B_FP8
- SGLang
How to use MultiverseComputingCAI/Comet_60B_A12B_FP8 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "MultiverseComputingCAI/Comet_60B_A12B_FP8" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "MultiverseComputingCAI/Comet_60B_A12B_FP8", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "MultiverseComputingCAI/Comet_60B_A12B_FP8" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "MultiverseComputingCAI/Comet_60B_A12B_FP8", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }' - Docker Model Runner
How to use MultiverseComputingCAI/Comet_60B_A12B_FP8 with Docker Model Runner:
docker model run hf.co/MultiverseComputingCAI/Comet_60B_A12B_FP8
Download config.json from MultiverseComputingCAI/Comet_60B_A12B_FP8: direct link, hf CLI and curl.
- Browser
- Download file 51 kB
-
https://huggingface.co/MultiverseComputingCAI/Comet_60B_A12B_FP8/resolve/main/config.json
- Command line
-
hf download hf://MultiverseComputingCAI/Comet_60B_A12B_FP8/config.json
-
curl -L -o config.json https://huggingface.co/MultiverseComputingCAI/Comet_60B_A12B_FP8/resolve/main/config.json
51 kB
| { | |
| "architectures": [ | |
| "NemotronH_Omni_Reasoning_V3" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_nemotron_h_omni.NemotronH_Omni_Reasoning_V3_Config", | |
| "AutoModel": "modeling_nemotron_h_omni.NemotronH_Omni_Reasoning_V3", | |
| "AutoModelForCausalLM": "modeling_nemotron_h_omni.NemotronH_Omni_Reasoning_V3", | |
| "AutoModelForImageTextToText": "modeling_nemotron_h_omni.NemotronH_Omni_Reasoning_V3" | |
| }, | |
| "downsample_ratio": 0.5, | |
| "dtype": "bfloat16", | |
| "eos_token_id": 11, | |
| "force_image_size": 512, | |
| "image_tag_type": "internvl", | |
| "img_context_token": "<image>", | |
| "img_context_token_id": 18, | |
| "llm_config": { | |
| "architectures": [ | |
| "NemotronHForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "chunk_size": 128, | |
| "conv_kernel": 4, | |
| "dtype": "bfloat16", | |
| "eos_token_id": 11, | |
| "expand": 2, | |
| "head_dim": 128, | |
| "hidden_dropout": 0.0, | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 2688, | |
| "layer_norm_epsilon": 1e-05, | |
| "layers_block_type": [ | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "attention", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe", | |
| "mamba", | |
| "moe" | |
| ], | |
| "mamba_head_dim": 64, | |
| "mamba_hidden_act": "silu", | |
| "mamba_num_heads": 128, | |
| "mamba_proj_bias": false, | |
| "mamba_ssm_cache_dtype": "float32", | |
| "max_position_embeddings": 1048576, | |
| "mlp_bias": false, | |
| "mlp_hidden_act": "relu2", | |
| "model_type": "nemotron_h", | |
| "moe_intermediate_size": 2432, | |
| "moe_latent_size": 1024, | |
| "moe_shared_expert_intermediate_size": 5376, | |
| "moe_shared_expert_overlap": false, | |
| "mtp_layers_block_type": [ | |
| "attention", | |
| "moe" | |
| ], | |
| "n_group": 1, | |
| "n_groups": 8, | |
| "n_routed_experts": 256, | |
| "n_shared_experts": 1, | |
| "norm_topk_prob": true, | |
| "num_attention_heads": 32, | |
| "num_experts_per_tok": 22, | |
| "num_key_value_heads": 2, | |
| "num_logits_to_keep": 1, | |
| "num_nextn_predict_layers": 1, | |
| "pad_token_id": 0, | |
| "rescale_prenorm_residual": true, | |
| "residual_in_fp32": false, | |
| "routed_scaling_factor": 5.0, | |
| "sliding_window": null, | |
| "ssm_state_size": 128, | |
| "tie_word_embeddings": false, | |
| "time_step_floor": 0.0001, | |
| "time_step_limit": [ | |
| 0.0, | |
| { | |
| "__float__": "Infinity" | |
| } | |
| ], | |
| "time_step_max": 0.1, | |
| "time_step_min": 0.001, | |
| "topk_group": 1, | |
| "use_bias": false, | |
| "use_cache": true, | |
| "use_conv_bias": true, | |
| "use_mamba_kernels": true, | |
| "vocab_size": 131072 | |
| }, | |
| "model_type": "nemotron_h_omni", | |
| "norm_mean": [ | |
| 0.48145466, | |
| 0.4578275, | |
| 0.40821073 | |
| ], | |
| "norm_std": [ | |
| 0.26862954, | |
| 0.26130258, | |
| 0.27577711 | |
| ], | |
| "patch_size": 16, | |
| "processor_class": "NemotronH_Omni_Reasoning_V3Processor", | |
| "projector_hidden_size": 20480, | |
| "ps_version": "v2", | |
| "sound_config": null, | |
| "sound_context_token": "<audio>", | |
| "sound_context_token_id": null, | |
| "template": "n5h_5p5_nanov2", | |
| "tie_word_embeddings": false, | |
| "transformers_version": "5.12.1", | |
| "use_thumbnail": false, | |
| "video_context_token": "<video>", | |
| "video_context_token_id": null, | |
| "video_pruning_rate": 0.7, | |
| "video_temporal_patch_size": 2, | |
| "vision_config": { | |
| "args": { | |
| "cls_token_per_teacher": true, | |
| "cpe_max_size": 2048, | |
| "max_num_patches": 13312, | |
| "min_num_patches": 1024, | |
| "model": "vit_huge_patch16_224", | |
| "register_multiple": 10, | |
| "teachers": [ | |
| { | |
| "name": 0, | |
| "use_summary": true | |
| }, | |
| { | |
| "name": 1, | |
| "use_summary": true | |
| }, | |
| { | |
| "name": 2, | |
| "use_summary": false | |
| } | |
| ] | |
| }, | |
| "attention_probs_dropout_prob": 0.0, | |
| "drop_path_rate": 0.0, | |
| "dtype": "bfloat16", | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.0, | |
| "hidden_size": 1280, | |
| "image_size": 224, | |
| "initializer_range": 0.02, | |
| "layer_norm_eps": 1e-06, | |
| "layerscale_value": 1.0, | |
| "max_img_size": 2048, | |
| "max_num_patches": 13312, | |
| "min_num_patches": 1024, | |
| "mlp_ratio": 4.0, | |
| "model_type": "radio", | |
| "norm_mean": [ | |
| 0.48145466, | |
| 0.4578275, | |
| 0.40821073 | |
| ], | |
| "norm_std": [ | |
| 0.26862954, | |
| 0.26130258, | |
| 0.27577711 | |
| ], | |
| "num_attention_heads": 16, | |
| "num_channels": 3, | |
| "num_cls_tokens": 3, | |
| "num_hidden_layers": 32, | |
| "num_registers": 7, | |
| "patch_size": 16, | |
| "qkv_bias": true, | |
| "summary_idxs": [ | |
| 0, | |
| 1 | |
| ], | |
| "use_swiglu_ffn": false, | |
| "video_temporal_patch_size": 2 | |
| }, | |
| "vit_hidden_size": 1280, | |
| "quantization_config": { | |
| "config_groups": { | |
| "group_0": { | |
| "input_activations": { | |
| "dynamic": false, | |
| "num_bits": 8, | |
| "type": "float" | |
| }, | |
| "weights": { | |
| "dynamic": false, | |
| "num_bits": 8, | |
| "type": "float" | |
| }, | |
| "targets": [ | |
| "Linear" | |
| ] | |
| } | |
| }, | |
| "ignore": [ | |
| "language_model.backbone.embeddings", | |
| "language_model.backbone.layers.0.mixer.conv1d", | |
| "language_model.backbone.layers.1.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.1.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.1.mixer.gate", | |
| "language_model.backbone.layers.10.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.10.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.10.mixer.gate", | |
| "language_model.backbone.layers.11.mixer.conv1d", | |
| "language_model.backbone.layers.12.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.12.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.12.mixer.gate", | |
| "language_model.backbone.layers.13.mixer.conv1d", | |
| "language_model.backbone.layers.14.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.14.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.14.mixer.gate", | |
| "language_model.backbone.layers.15.mixer.conv1d", | |
| "language_model.backbone.layers.16.mixer.k_proj", | |
| "language_model.backbone.layers.16.mixer.o_proj", | |
| "language_model.backbone.layers.16.mixer.q_proj", | |
| "language_model.backbone.layers.16.mixer.v_proj", | |
| "language_model.backbone.layers.17.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.17.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.17.mixer.gate", | |
| "language_model.backbone.layers.18.mixer.conv1d", | |
| "language_model.backbone.layers.19.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.19.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.19.mixer.gate", | |
| "language_model.backbone.layers.2.mixer.conv1d", | |
| "language_model.backbone.layers.20.mixer.conv1d", | |
| "language_model.backbone.layers.21.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.21.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.21.mixer.gate", | |
| "language_model.backbone.layers.22.mixer.conv1d", | |
| "language_model.backbone.layers.23.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.23.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.23.mixer.gate", | |
| "language_model.backbone.layers.24.mixer.conv1d", | |
| "language_model.backbone.layers.25.mixer.k_proj", | |
| "language_model.backbone.layers.25.mixer.o_proj", | |
| "language_model.backbone.layers.25.mixer.q_proj", | |
| "language_model.backbone.layers.25.mixer.v_proj", | |
| "language_model.backbone.layers.26.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.26.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.26.mixer.gate", | |
| "language_model.backbone.layers.27.mixer.conv1d", | |
| "language_model.backbone.layers.28.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.28.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.28.mixer.gate", | |
| "language_model.backbone.layers.29.mixer.conv1d", | |
| "language_model.backbone.layers.3.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.3.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.3.mixer.gate", | |
| "language_model.backbone.layers.30.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.30.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.30.mixer.gate", | |
| "language_model.backbone.layers.31.mixer.conv1d", | |
| "language_model.backbone.layers.32.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.32.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.32.mixer.gate", | |
| "language_model.backbone.layers.33.mixer.conv1d", | |
| "language_model.backbone.layers.34.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.34.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.34.mixer.gate", | |
| "language_model.backbone.layers.35.mixer.conv1d", | |
| "language_model.backbone.layers.36.mixer.k_proj", | |
| "language_model.backbone.layers.36.mixer.o_proj", | |
| "language_model.backbone.layers.36.mixer.q_proj", | |
| "language_model.backbone.layers.36.mixer.v_proj", | |
| "language_model.backbone.layers.37.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.37.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.37.mixer.gate", | |
| "language_model.backbone.layers.38.mixer.conv1d", | |
| "language_model.backbone.layers.39.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.39.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.39.mixer.gate", | |
| "language_model.backbone.layers.4.mixer.conv1d", | |
| "language_model.backbone.layers.40.mixer.conv1d", | |
| "language_model.backbone.layers.41.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.41.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.41.mixer.gate", | |
| "language_model.backbone.layers.42.mixer.conv1d", | |
| "language_model.backbone.layers.43.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.43.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.43.mixer.gate", | |
| "language_model.backbone.layers.44.mixer.conv1d", | |
| "language_model.backbone.layers.45.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.45.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.45.mixer.gate", | |
| "language_model.backbone.layers.46.mixer.conv1d", | |
| "language_model.backbone.layers.47.mixer.k_proj", | |
| "language_model.backbone.layers.47.mixer.o_proj", | |
| "language_model.backbone.layers.47.mixer.q_proj", | |
| "language_model.backbone.layers.47.mixer.v_proj", | |
| "language_model.backbone.layers.48.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.48.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.48.mixer.gate", | |
| "language_model.backbone.layers.49.mixer.conv1d", | |
| "language_model.backbone.layers.5.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.5.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.5.mixer.gate", | |
| "language_model.backbone.layers.50.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.50.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.50.mixer.gate", | |
| "language_model.backbone.layers.51.mixer.conv1d", | |
| "language_model.backbone.layers.52.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.52.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.52.mixer.gate", | |
| "language_model.backbone.layers.53.mixer.conv1d", | |
| "language_model.backbone.layers.54.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.54.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.54.mixer.gate", | |
| "language_model.backbone.layers.55.mixer.conv1d", | |
| "language_model.backbone.layers.56.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.56.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.56.mixer.gate", | |
| "language_model.backbone.layers.57.mixer.conv1d", | |
| "language_model.backbone.layers.58.mixer.k_proj", | |
| "language_model.backbone.layers.58.mixer.o_proj", | |
| "language_model.backbone.layers.58.mixer.q_proj", | |
| "language_model.backbone.layers.58.mixer.v_proj", | |
| "language_model.backbone.layers.59.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.59.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.59.mixer.gate", | |
| "language_model.backbone.layers.6.mixer.conv1d", | |
| "language_model.backbone.layers.60.mixer.conv1d", | |
| "language_model.backbone.layers.61.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.61.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.61.mixer.gate", | |
| "language_model.backbone.layers.62.mixer.conv1d", | |
| "language_model.backbone.layers.63.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.63.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.63.mixer.gate", | |
| "language_model.backbone.layers.64.mixer.conv1d", | |
| "language_model.backbone.layers.65.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.65.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.65.mixer.gate", | |
| "language_model.backbone.layers.66.mixer.conv1d", | |
| "language_model.backbone.layers.67.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.67.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.67.mixer.gate", | |
| "language_model.backbone.layers.68.mixer.conv1d", | |
| "language_model.backbone.layers.69.mixer.k_proj", | |
| "language_model.backbone.layers.69.mixer.o_proj", | |
| "language_model.backbone.layers.69.mixer.q_proj", | |
| "language_model.backbone.layers.69.mixer.v_proj", | |
| "language_model.backbone.layers.7.mixer.k_proj", | |
| "language_model.backbone.layers.7.mixer.o_proj", | |
| "language_model.backbone.layers.7.mixer.q_proj", | |
| "language_model.backbone.layers.7.mixer.v_proj", | |
| "language_model.backbone.layers.70.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.70.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.70.mixer.gate", | |
| "language_model.backbone.layers.71.mixer.conv1d", | |
| "language_model.backbone.layers.72.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.72.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.72.mixer.gate", | |
| "language_model.backbone.layers.73.mixer.conv1d", | |
| "language_model.backbone.layers.74.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.74.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.74.mixer.gate", | |
| "language_model.backbone.layers.75.mixer.conv1d", | |
| "language_model.backbone.layers.76.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.76.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.76.mixer.gate", | |
| "language_model.backbone.layers.77.mixer.conv1d", | |
| "language_model.backbone.layers.78.mixer.k_proj", | |
| "language_model.backbone.layers.78.mixer.o_proj", | |
| "language_model.backbone.layers.78.mixer.q_proj", | |
| "language_model.backbone.layers.78.mixer.v_proj", | |
| "language_model.backbone.layers.79.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.79.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.79.mixer.gate", | |
| "language_model.backbone.layers.8.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.8.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.8.mixer.gate", | |
| "language_model.backbone.layers.80.mixer.conv1d", | |
| "language_model.backbone.layers.81.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.81.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.81.mixer.gate", | |
| "language_model.backbone.layers.82.mixer.conv1d", | |
| "language_model.backbone.layers.83.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.83.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.83.mixer.gate", | |
| "language_model.backbone.layers.84.mixer.conv1d", | |
| "language_model.backbone.layers.85.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.85.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.85.mixer.gate", | |
| "language_model.backbone.layers.86.mixer.conv1d", | |
| "language_model.backbone.layers.87.mixer.fc1_latent_proj", | |
| "language_model.backbone.layers.87.mixer.fc2_latent_proj", | |
| "language_model.backbone.layers.87.mixer.gate", | |
| "language_model.backbone.layers.9.mixer.conv1d", | |
| "language_model.lm_head", | |
| "language_model.mtp.layers.0.eh_proj", | |
| "language_model.mtp.layers.0.mixer.k_proj", | |
| "language_model.mtp.layers.0.mixer.o_proj", | |
| "language_model.mtp.layers.0.mixer.q_proj", | |
| "language_model.mtp.layers.0.mixer.v_proj", | |
| "language_model.mtp.layers.1.mixer.experts.0.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.0.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.1.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.1.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.10.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.10.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.100.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.100.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.101.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.101.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.102.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.102.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.103.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.103.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.104.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.104.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.105.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.105.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.106.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.106.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.107.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.107.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.108.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.108.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.109.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.109.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.11.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.11.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.110.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.110.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.111.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.111.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.112.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.112.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.113.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.113.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.114.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.114.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.115.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.115.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.116.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.116.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.117.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.117.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.118.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.118.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.119.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.119.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.12.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.12.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.120.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.120.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.121.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.121.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.122.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.122.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.123.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.123.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.124.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.124.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.125.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.125.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.126.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.126.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.127.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.127.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.128.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.128.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.129.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.129.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.13.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.13.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.130.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.130.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.131.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.131.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.132.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.132.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.133.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.133.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.134.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.134.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.135.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.135.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.136.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.136.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.137.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.137.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.138.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.138.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.139.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.139.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.14.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.14.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.140.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.140.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.141.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.141.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.142.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.142.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.143.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.143.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.144.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.144.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.145.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.145.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.146.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.146.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.147.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.147.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.148.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.148.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.149.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.149.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.15.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.15.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.150.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.150.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.151.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.151.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.152.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.152.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.153.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.153.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.154.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.154.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.155.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.155.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.156.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.156.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.157.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.157.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.158.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.158.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.159.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.159.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.16.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.16.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.160.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.160.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.161.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.161.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.162.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.162.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.163.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.163.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.164.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.164.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.165.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.165.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.166.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.166.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.167.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.167.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.168.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.168.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.169.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.169.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.17.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.17.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.170.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.170.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.171.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.171.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.172.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.172.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.173.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.173.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.174.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.174.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.175.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.175.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.176.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.176.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.177.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.177.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.178.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.178.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.179.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.179.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.18.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.18.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.180.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.180.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.181.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.181.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.182.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.182.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.183.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.183.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.184.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.184.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.185.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.185.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.186.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.186.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.187.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.187.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.188.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.188.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.189.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.189.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.19.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.19.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.190.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.190.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.191.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.191.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.192.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.192.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.193.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.193.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.194.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.194.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.195.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.195.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.196.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.196.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.197.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.197.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.198.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.198.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.199.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.199.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.2.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.2.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.20.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.20.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.200.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.200.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.201.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.201.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.202.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.202.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.203.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.203.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.204.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.204.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.205.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.205.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.206.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.206.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.207.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.207.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.208.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.208.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.209.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.209.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.21.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.21.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.210.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.210.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.211.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.211.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.212.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.212.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.213.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.213.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.214.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.214.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.215.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.215.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.216.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.216.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.217.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.217.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.218.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.218.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.219.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.219.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.22.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.22.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.220.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.220.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.221.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.221.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.222.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.222.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.223.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.223.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.224.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.224.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.225.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.225.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.226.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.226.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.227.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.227.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.228.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.228.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.229.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.229.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.23.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.23.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.230.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.230.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.231.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.231.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.232.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.232.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.233.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.233.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.234.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.234.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.235.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.235.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.236.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.236.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.237.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.237.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.238.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.238.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.239.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.239.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.24.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.24.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.240.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.240.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.241.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.241.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.242.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.242.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.243.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.243.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.244.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.244.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.245.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.245.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.246.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.246.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.247.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.247.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.248.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.248.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.249.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.249.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.25.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.25.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.250.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.250.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.251.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.251.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.252.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.252.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.253.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.253.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.254.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.254.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.255.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.255.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.26.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.26.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.27.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.27.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.28.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.28.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.29.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.29.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.3.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.3.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.30.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.30.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.31.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.31.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.32.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.32.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.33.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.33.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.34.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.34.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.35.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.35.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.36.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.36.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.37.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.37.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.38.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.38.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.39.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.39.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.4.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.4.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.40.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.40.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.41.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.41.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.42.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.42.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.43.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.43.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.44.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.44.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.45.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.45.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.46.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.46.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.47.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.47.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.48.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.48.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.49.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.49.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.5.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.5.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.50.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.50.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.51.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.51.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.52.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.52.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.53.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.53.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.54.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.54.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.55.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.55.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.56.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.56.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.57.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.57.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.58.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.58.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.59.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.59.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.6.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.6.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.60.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.60.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.61.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.61.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.62.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.62.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.63.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.63.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.64.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.64.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.65.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.65.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.66.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.66.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.67.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.67.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.68.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.68.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.69.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.69.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.7.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.7.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.70.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.70.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.71.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.71.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.72.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.72.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.73.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.73.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.74.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.74.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.75.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.75.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.76.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.76.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.77.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.77.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.78.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.78.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.79.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.79.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.8.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.8.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.80.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.80.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.81.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.81.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.82.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.82.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.83.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.83.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.84.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.84.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.85.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.85.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.86.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.86.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.87.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.87.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.88.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.88.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.89.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.89.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.9.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.9.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.90.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.90.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.91.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.91.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.92.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.92.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.93.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.93.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.94.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.94.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.95.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.95.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.96.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.96.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.97.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.97.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.98.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.98.up_proj", | |
| "language_model.mtp.layers.1.mixer.experts.99.down_proj", | |
| "language_model.mtp.layers.1.mixer.experts.99.up_proj", | |
| "language_model.mtp.layers.1.mixer.fc1_latent_proj", | |
| "language_model.mtp.layers.1.mixer.fc2_latent_proj", | |
| "language_model.mtp.layers.1.mixer.gate", | |
| "language_model.mtp.layers.1.mixer.shared_experts.down_proj", | |
| "language_model.mtp.layers.1.mixer.shared_experts.up_proj", | |
| "vision_model.*", | |
| "mlp1.*", | |
| "vision_projector.*" | |
| ], | |
| "quant_algo": "FP8", | |
| "kv_cache_scheme": { | |
| "dynamic": false, | |
| "num_bits": 8, | |
| "type": "float" | |
| }, | |
| "producer": { | |
| "name": "modelopt", | |
| "version": "0.47.0rc0" | |
| }, | |
| "quant_method": "modelopt" | |
| } | |
| } |