thisnick's picture
Add files using upload-large-folder tool
3c06106 verified
raw
history blame
16.1 kB
{
"_name_or_path": "/content/llama-nsfw-video-still-captioner",
"architectures": [
"MllamaForConditionalGeneration"
],
"image_token_index": 128256,
"model_type": "mllama",
"pad_token_id": 128004,
"quantization_config": {
"config_groups": {
"group_0": {
"input_activations": {
"actorder": null,
"block_structure": null,
"dynamic": true,
"group_size": null,
"num_bits": 8,
"observer": null,
"observer_kwargs": {},
"strategy": "token",
"symmetric": true,
"type": "float"
},
"output_activations": null,
"targets": [
"Linear"
],
"weights": {
"actorder": null,
"block_structure": null,
"dynamic": false,
"group_size": null,
"num_bits": 8,
"observer": "minmax",
"observer_kwargs": {},
"strategy": "channel",
"symmetric": true,
"type": "float"
}
}
},
"format": "float-quantized",
"global_compression_ratio": 1.2163620952885337,
"ignore": [
"vision_model.transformer.layers.0.self_attn.q_proj",
"vision_model.transformer.layers.0.self_attn.k_proj",
"vision_model.transformer.layers.0.self_attn.v_proj",
"vision_model.transformer.layers.0.self_attn.o_proj",
"vision_model.transformer.layers.0.mlp.fc1",
"vision_model.transformer.layers.0.mlp.fc2",
"vision_model.transformer.layers.1.self_attn.q_proj",
"vision_model.transformer.layers.1.self_attn.k_proj",
"vision_model.transformer.layers.1.self_attn.v_proj",
"vision_model.transformer.layers.1.self_attn.o_proj",
"vision_model.transformer.layers.1.mlp.fc1",
"vision_model.transformer.layers.1.mlp.fc2",
"vision_model.transformer.layers.2.self_attn.q_proj",
"vision_model.transformer.layers.2.self_attn.k_proj",
"vision_model.transformer.layers.2.self_attn.v_proj",
"vision_model.transformer.layers.2.self_attn.o_proj",
"vision_model.transformer.layers.2.mlp.fc1",
"vision_model.transformer.layers.2.mlp.fc2",
"vision_model.transformer.layers.3.self_attn.q_proj",
"vision_model.transformer.layers.3.self_attn.k_proj",
"vision_model.transformer.layers.3.self_attn.v_proj",
"vision_model.transformer.layers.3.self_attn.o_proj",
"vision_model.transformer.layers.3.mlp.fc1",
"vision_model.transformer.layers.3.mlp.fc2",
"vision_model.transformer.layers.4.self_attn.q_proj",
"vision_model.transformer.layers.4.self_attn.k_proj",
"vision_model.transformer.layers.4.self_attn.v_proj",
"vision_model.transformer.layers.4.self_attn.o_proj",
"vision_model.transformer.layers.4.mlp.fc1",
"vision_model.transformer.layers.4.mlp.fc2",
"vision_model.transformer.layers.5.self_attn.q_proj",
"vision_model.transformer.layers.5.self_attn.k_proj",
"vision_model.transformer.layers.5.self_attn.v_proj",
"vision_model.transformer.layers.5.self_attn.o_proj",
"vision_model.transformer.layers.5.mlp.fc1",
"vision_model.transformer.layers.5.mlp.fc2",
"vision_model.transformer.layers.6.self_attn.q_proj",
"vision_model.transformer.layers.6.self_attn.k_proj",
"vision_model.transformer.layers.6.self_attn.v_proj",
"vision_model.transformer.layers.6.self_attn.o_proj",
"vision_model.transformer.layers.6.mlp.fc1",
"vision_model.transformer.layers.6.mlp.fc2",
"vision_model.transformer.layers.7.self_attn.q_proj",
"vision_model.transformer.layers.7.self_attn.k_proj",
"vision_model.transformer.layers.7.self_attn.v_proj",
"vision_model.transformer.layers.7.self_attn.o_proj",
"vision_model.transformer.layers.7.mlp.fc1",
"vision_model.transformer.layers.7.mlp.fc2",
"vision_model.transformer.layers.8.self_attn.q_proj",
"vision_model.transformer.layers.8.self_attn.k_proj",
"vision_model.transformer.layers.8.self_attn.v_proj",
"vision_model.transformer.layers.8.self_attn.o_proj",
"vision_model.transformer.layers.8.mlp.fc1",
"vision_model.transformer.layers.8.mlp.fc2",
"vision_model.transformer.layers.9.self_attn.q_proj",
"vision_model.transformer.layers.9.self_attn.k_proj",
"vision_model.transformer.layers.9.self_attn.v_proj",
"vision_model.transformer.layers.9.self_attn.o_proj",
"vision_model.transformer.layers.9.mlp.fc1",
"vision_model.transformer.layers.9.mlp.fc2",
"vision_model.transformer.layers.10.self_attn.q_proj",
"vision_model.transformer.layers.10.self_attn.k_proj",
"vision_model.transformer.layers.10.self_attn.v_proj",
"vision_model.transformer.layers.10.self_attn.o_proj",
"vision_model.transformer.layers.10.mlp.fc1",
"vision_model.transformer.layers.10.mlp.fc2",
"vision_model.transformer.layers.11.self_attn.q_proj",
"vision_model.transformer.layers.11.self_attn.k_proj",
"vision_model.transformer.layers.11.self_attn.v_proj",
"vision_model.transformer.layers.11.self_attn.o_proj",
"vision_model.transformer.layers.11.mlp.fc1",
"vision_model.transformer.layers.11.mlp.fc2",
"vision_model.transformer.layers.12.self_attn.q_proj",
"vision_model.transformer.layers.12.self_attn.k_proj",
"vision_model.transformer.layers.12.self_attn.v_proj",
"vision_model.transformer.layers.12.self_attn.o_proj",
"vision_model.transformer.layers.12.mlp.fc1",
"vision_model.transformer.layers.12.mlp.fc2",
"vision_model.transformer.layers.13.self_attn.q_proj",
"vision_model.transformer.layers.13.self_attn.k_proj",
"vision_model.transformer.layers.13.self_attn.v_proj",
"vision_model.transformer.layers.13.self_attn.o_proj",
"vision_model.transformer.layers.13.mlp.fc1",
"vision_model.transformer.layers.13.mlp.fc2",
"vision_model.transformer.layers.14.self_attn.q_proj",
"vision_model.transformer.layers.14.self_attn.k_proj",
"vision_model.transformer.layers.14.self_attn.v_proj",
"vision_model.transformer.layers.14.self_attn.o_proj",
"vision_model.transformer.layers.14.mlp.fc1",
"vision_model.transformer.layers.14.mlp.fc2",
"vision_model.transformer.layers.15.self_attn.q_proj",
"vision_model.transformer.layers.15.self_attn.k_proj",
"vision_model.transformer.layers.15.self_attn.v_proj",
"vision_model.transformer.layers.15.self_attn.o_proj",
"vision_model.transformer.layers.15.mlp.fc1",
"vision_model.transformer.layers.15.mlp.fc2",
"vision_model.transformer.layers.16.self_attn.q_proj",
"vision_model.transformer.layers.16.self_attn.k_proj",
"vision_model.transformer.layers.16.self_attn.v_proj",
"vision_model.transformer.layers.16.self_attn.o_proj",
"vision_model.transformer.layers.16.mlp.fc1",
"vision_model.transformer.layers.16.mlp.fc2",
"vision_model.transformer.layers.17.self_attn.q_proj",
"vision_model.transformer.layers.17.self_attn.k_proj",
"vision_model.transformer.layers.17.self_attn.v_proj",
"vision_model.transformer.layers.17.self_attn.o_proj",
"vision_model.transformer.layers.17.mlp.fc1",
"vision_model.transformer.layers.17.mlp.fc2",
"vision_model.transformer.layers.18.self_attn.q_proj",
"vision_model.transformer.layers.18.self_attn.k_proj",
"vision_model.transformer.layers.18.self_attn.v_proj",
"vision_model.transformer.layers.18.self_attn.o_proj",
"vision_model.transformer.layers.18.mlp.fc1",
"vision_model.transformer.layers.18.mlp.fc2",
"vision_model.transformer.layers.19.self_attn.q_proj",
"vision_model.transformer.layers.19.self_attn.k_proj",
"vision_model.transformer.layers.19.self_attn.v_proj",
"vision_model.transformer.layers.19.self_attn.o_proj",
"vision_model.transformer.layers.19.mlp.fc1",
"vision_model.transformer.layers.19.mlp.fc2",
"vision_model.transformer.layers.20.self_attn.q_proj",
"vision_model.transformer.layers.20.self_attn.k_proj",
"vision_model.transformer.layers.20.self_attn.v_proj",
"vision_model.transformer.layers.20.self_attn.o_proj",
"vision_model.transformer.layers.20.mlp.fc1",
"vision_model.transformer.layers.20.mlp.fc2",
"vision_model.transformer.layers.21.self_attn.q_proj",
"vision_model.transformer.layers.21.self_attn.k_proj",
"vision_model.transformer.layers.21.self_attn.v_proj",
"vision_model.transformer.layers.21.self_attn.o_proj",
"vision_model.transformer.layers.21.mlp.fc1",
"vision_model.transformer.layers.21.mlp.fc2",
"vision_model.transformer.layers.22.self_attn.q_proj",
"vision_model.transformer.layers.22.self_attn.k_proj",
"vision_model.transformer.layers.22.self_attn.v_proj",
"vision_model.transformer.layers.22.self_attn.o_proj",
"vision_model.transformer.layers.22.mlp.fc1",
"vision_model.transformer.layers.22.mlp.fc2",
"vision_model.transformer.layers.23.self_attn.q_proj",
"vision_model.transformer.layers.23.self_attn.k_proj",
"vision_model.transformer.layers.23.self_attn.v_proj",
"vision_model.transformer.layers.23.self_attn.o_proj",
"vision_model.transformer.layers.23.mlp.fc1",
"vision_model.transformer.layers.23.mlp.fc2",
"vision_model.transformer.layers.24.self_attn.q_proj",
"vision_model.transformer.layers.24.self_attn.k_proj",
"vision_model.transformer.layers.24.self_attn.v_proj",
"vision_model.transformer.layers.24.self_attn.o_proj",
"vision_model.transformer.layers.24.mlp.fc1",
"vision_model.transformer.layers.24.mlp.fc2",
"vision_model.transformer.layers.25.self_attn.q_proj",
"vision_model.transformer.layers.25.self_attn.k_proj",
"vision_model.transformer.layers.25.self_attn.v_proj",
"vision_model.transformer.layers.25.self_attn.o_proj",
"vision_model.transformer.layers.25.mlp.fc1",
"vision_model.transformer.layers.25.mlp.fc2",
"vision_model.transformer.layers.26.self_attn.q_proj",
"vision_model.transformer.layers.26.self_attn.k_proj",
"vision_model.transformer.layers.26.self_attn.v_proj",
"vision_model.transformer.layers.26.self_attn.o_proj",
"vision_model.transformer.layers.26.mlp.fc1",
"vision_model.transformer.layers.26.mlp.fc2",
"vision_model.transformer.layers.27.self_attn.q_proj",
"vision_model.transformer.layers.27.self_attn.k_proj",
"vision_model.transformer.layers.27.self_attn.v_proj",
"vision_model.transformer.layers.27.self_attn.o_proj",
"vision_model.transformer.layers.27.mlp.fc1",
"vision_model.transformer.layers.27.mlp.fc2",
"vision_model.transformer.layers.28.self_attn.q_proj",
"vision_model.transformer.layers.28.self_attn.k_proj",
"vision_model.transformer.layers.28.self_attn.v_proj",
"vision_model.transformer.layers.28.self_attn.o_proj",
"vision_model.transformer.layers.28.mlp.fc1",
"vision_model.transformer.layers.28.mlp.fc2",
"vision_model.transformer.layers.29.self_attn.q_proj",
"vision_model.transformer.layers.29.self_attn.k_proj",
"vision_model.transformer.layers.29.self_attn.v_proj",
"vision_model.transformer.layers.29.self_attn.o_proj",
"vision_model.transformer.layers.29.mlp.fc1",
"vision_model.transformer.layers.29.mlp.fc2",
"vision_model.transformer.layers.30.self_attn.q_proj",
"vision_model.transformer.layers.30.self_attn.k_proj",
"vision_model.transformer.layers.30.self_attn.v_proj",
"vision_model.transformer.layers.30.self_attn.o_proj",
"vision_model.transformer.layers.30.mlp.fc1",
"vision_model.transformer.layers.30.mlp.fc2",
"vision_model.transformer.layers.31.self_attn.q_proj",
"vision_model.transformer.layers.31.self_attn.k_proj",
"vision_model.transformer.layers.31.self_attn.v_proj",
"vision_model.transformer.layers.31.self_attn.o_proj",
"vision_model.transformer.layers.31.mlp.fc1",
"vision_model.transformer.layers.31.mlp.fc2",
"vision_model.global_transformer.layers.0.self_attn.q_proj",
"vision_model.global_transformer.layers.0.self_attn.k_proj",
"vision_model.global_transformer.layers.0.self_attn.v_proj",
"vision_model.global_transformer.layers.0.self_attn.o_proj",
"vision_model.global_transformer.layers.0.mlp.fc1",
"vision_model.global_transformer.layers.0.mlp.fc2",
"vision_model.global_transformer.layers.1.self_attn.q_proj",
"vision_model.global_transformer.layers.1.self_attn.k_proj",
"vision_model.global_transformer.layers.1.self_attn.v_proj",
"vision_model.global_transformer.layers.1.self_attn.o_proj",
"vision_model.global_transformer.layers.1.mlp.fc1",
"vision_model.global_transformer.layers.1.mlp.fc2",
"vision_model.global_transformer.layers.2.self_attn.q_proj",
"vision_model.global_transformer.layers.2.self_attn.k_proj",
"vision_model.global_transformer.layers.2.self_attn.v_proj",
"vision_model.global_transformer.layers.2.self_attn.o_proj",
"vision_model.global_transformer.layers.2.mlp.fc1",
"vision_model.global_transformer.layers.2.mlp.fc2",
"vision_model.global_transformer.layers.3.self_attn.q_proj",
"vision_model.global_transformer.layers.3.self_attn.k_proj",
"vision_model.global_transformer.layers.3.self_attn.v_proj",
"vision_model.global_transformer.layers.3.self_attn.o_proj",
"vision_model.global_transformer.layers.3.mlp.fc1",
"vision_model.global_transformer.layers.3.mlp.fc2",
"vision_model.global_transformer.layers.4.self_attn.q_proj",
"vision_model.global_transformer.layers.4.self_attn.k_proj",
"vision_model.global_transformer.layers.4.self_attn.v_proj",
"vision_model.global_transformer.layers.4.self_attn.o_proj",
"vision_model.global_transformer.layers.4.mlp.fc1",
"vision_model.global_transformer.layers.4.mlp.fc2",
"vision_model.global_transformer.layers.5.self_attn.q_proj",
"vision_model.global_transformer.layers.5.self_attn.k_proj",
"vision_model.global_transformer.layers.5.self_attn.v_proj",
"vision_model.global_transformer.layers.5.self_attn.o_proj",
"vision_model.global_transformer.layers.5.mlp.fc1",
"vision_model.global_transformer.layers.5.mlp.fc2",
"vision_model.global_transformer.layers.6.self_attn.q_proj",
"vision_model.global_transformer.layers.6.self_attn.k_proj",
"vision_model.global_transformer.layers.6.self_attn.v_proj",
"vision_model.global_transformer.layers.6.self_attn.o_proj",
"vision_model.global_transformer.layers.6.mlp.fc1",
"vision_model.global_transformer.layers.6.mlp.fc2",
"vision_model.global_transformer.layers.7.self_attn.q_proj",
"vision_model.global_transformer.layers.7.self_attn.k_proj",
"vision_model.global_transformer.layers.7.self_attn.v_proj",
"vision_model.global_transformer.layers.7.self_attn.o_proj",
"vision_model.global_transformer.layers.7.mlp.fc1",
"vision_model.global_transformer.layers.7.mlp.fc2",
"language_model.lm_head",
"multi_modal_projector"
],
"kv_cache_scheme": null,
"quant_method": "compressed-tensors",
"quantization_status": "compressed"
},
"text_config": {
"eos_token_id": [
128001,
128008,
128009
],
"model_type": "mllama_text_model",
"rope_scaling": {
"factor": 8.0,
"high_freq_factor": 4.0,
"low_freq_factor": 1.0,
"original_max_position_embeddings": 8192,
"rope_type": "llama3"
},
"torch_dtype": "bfloat16"
},
"torch_dtype": "bfloat16",
"transformers_version": "4.49.0",
"unsloth_fixed": true,
"unsloth_version": "2025.2.12",
"vision_config": {
"image_size": 560,
"model_type": "mllama_vision_model",
"torch_dtype": "bfloat16"
}
}