{ "_name_or_path": "/home/mgoin/.cache/huggingface/hub/models--llava-hf--llava-onevision-qwen2-7b-ov-hf/snapshots/cc1d22d900237602e3c1093cee0c8b8ed689c42a", "architectures": [ "LlavaOnevisionForConditionalGeneration" ], "ignore_index": -100, "image_grid_pinpoints": [ [ 384, 384 ], [ 384, 768 ], [ 384, 1152 ], [ 384, 1536 ], [ 384, 1920 ], [ 384, 2304 ], [ 768, 384 ], [ 768, 768 ], [ 768, 1152 ], [ 768, 1536 ], [ 768, 1920 ], [ 768, 2304 ], [ 1152, 384 ], [ 1152, 768 ], [ 1152, 1152 ], [ 1152, 1536 ], [ 1152, 1920 ], [ 1152, 2304 ], [ 1536, 384 ], [ 1536, 768 ], [ 1536, 1152 ], [ 1536, 1536 ], [ 1536, 1920 ], [ 1536, 2304 ], [ 1920, 384 ], [ 1920, 768 ], [ 1920, 1152 ], [ 1920, 1536 ], [ 1920, 1920 ], [ 1920, 2304 ], [ 2304, 384 ], [ 2304, 768 ], [ 2304, 1152 ], [ 2304, 1536 ], [ 2304, 1920 ], [ 2304, 2304 ] ], "image_token_index": 151646, "model_type": "llava_onevision", "projector_hidden_act": "gelu", "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": 1.2158363142513449, "ignore": [ "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.0.mlp.fc1", "vision_tower.vision_model.encoder.layers.0.mlp.fc2", "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.1.mlp.fc1", "vision_tower.vision_model.encoder.layers.1.mlp.fc2", "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.2.mlp.fc1", "vision_tower.vision_model.encoder.layers.2.mlp.fc2", "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.3.mlp.fc1", "vision_tower.vision_model.encoder.layers.3.mlp.fc2", "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.4.mlp.fc1", "vision_tower.vision_model.encoder.layers.4.mlp.fc2", "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.5.mlp.fc1", "vision_tower.vision_model.encoder.layers.5.mlp.fc2", "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.6.mlp.fc1", "vision_tower.vision_model.encoder.layers.6.mlp.fc2", "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.7.mlp.fc1", "vision_tower.vision_model.encoder.layers.7.mlp.fc2", "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.8.mlp.fc1", "vision_tower.vision_model.encoder.layers.8.mlp.fc2", "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.9.mlp.fc1", "vision_tower.vision_model.encoder.layers.9.mlp.fc2", "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.10.mlp.fc1", "vision_tower.vision_model.encoder.layers.10.mlp.fc2", "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.11.mlp.fc1", "vision_tower.vision_model.encoder.layers.11.mlp.fc2", "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.12.mlp.fc1", "vision_tower.vision_model.encoder.layers.12.mlp.fc2", "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.13.mlp.fc1", "vision_tower.vision_model.encoder.layers.13.mlp.fc2", "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.14.mlp.fc1", "vision_tower.vision_model.encoder.layers.14.mlp.fc2", "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.15.mlp.fc1", "vision_tower.vision_model.encoder.layers.15.mlp.fc2", "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.16.mlp.fc1", "vision_tower.vision_model.encoder.layers.16.mlp.fc2", "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.17.mlp.fc1", "vision_tower.vision_model.encoder.layers.17.mlp.fc2", "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.18.mlp.fc1", "vision_tower.vision_model.encoder.layers.18.mlp.fc2", "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.19.mlp.fc1", "vision_tower.vision_model.encoder.layers.19.mlp.fc2", "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.20.mlp.fc1", "vision_tower.vision_model.encoder.layers.20.mlp.fc2", "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.21.mlp.fc1", "vision_tower.vision_model.encoder.layers.21.mlp.fc2", "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.22.mlp.fc1", "vision_tower.vision_model.encoder.layers.22.mlp.fc2", "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.23.mlp.fc1", "vision_tower.vision_model.encoder.layers.23.mlp.fc2", "vision_tower.vision_model.encoder.layers.24.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.24.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.24.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.24.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.24.mlp.fc1", "vision_tower.vision_model.encoder.layers.24.mlp.fc2", "vision_tower.vision_model.encoder.layers.25.self_attn.k_proj", "vision_tower.vision_model.encoder.layers.25.self_attn.v_proj", "vision_tower.vision_model.encoder.layers.25.self_attn.q_proj", "vision_tower.vision_model.encoder.layers.25.self_attn.out_proj", "vision_tower.vision_model.encoder.layers.25.mlp.fc1", "vision_tower.vision_model.encoder.layers.25.mlp.fc2", "multi_modal_projector.linear_1", "multi_modal_projector.linear_2", "language_model.lm_head" ], "kv_cache_scheme": null, "quant_method": "compressed-tensors", "quantization_status": "compressed" }, "text_config": { "_name_or_path": "Qwen/Qwen2-7B-Instruct", "architectures": [ "Qwen2ForCausalLM" ], "bos_token_id": 151643, "eos_token_id": 151645, "hidden_size": 3584, "intermediate_size": 18944, "model_type": "qwen2", "num_attention_heads": 28, "num_hidden_layers": 28, "num_key_value_heads": 4, "rope_theta": 1000000.0, "torch_dtype": "bfloat16", "vocab_size": 152128 }, "tie_word_embeddings": false, "torch_dtype": "float16", "transformers_version": "4.46.0.dev0", "use_image_newline_parameter": true, "video_token_index": 151647, "vision_aspect_ratio": "anyres_max_9", "vision_config": { "hidden_size": 1152, "image_size": 384, "intermediate_size": 4304, "model_type": "siglip_vision_model", "num_attention_heads": 16, "num_hidden_layers": 26, "patch_size": 14, "vision_use_head": false }, "vision_feature_layer": -1, "vision_feature_select_strategy": "full" }