{ "architectures": [ "StepVLForConditionalGeneration" ], "auto_map": { "AutoConfig": "configuration_step_vl.StepRoboticsConfig", "AutoModelForCausalLM": "modeling_step_vl.Step3VL10BForCausalLM" }, "model_type": "step_robotics", "im_end_token": "", "im_patch_token": "", "im_start_token": "", "image_token_len": 169, "patch_token_len": 81, "image_token_id": 151679, "understand_projector_stride": 2, "use_im_start_end": "true", "vision_select_layer": -1, "projector_bias": false, "vision_config": { "image_size": 728, "patch_size": 14, "width": 1536, "layers": 47, "heads": 16, "pool_type": "none", "output_dim": null, "use_cls_token": false, "ls_init_value": 0.1, "use_ln_post": false, "hidden_act": "quick_gelu" }, "text_config": { "architectures": [ "Qwen3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 151643, "eos_token_id": [ 151643, 151679 ], "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 12288, "max_position_embeddings": 65536, "max_window_layers": 36, "model_type": "qwen3", "num_attention_heads": 32, "num_hidden_layers": 36, "num_key_value_heads": 8, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.51.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151936 }, "quantization_config": { "quant_method": "fp8", "activation_scheme": "dynamic", "fmt": "e4m3", "weight_block_size": [ 128, 128 ], "modules_to_not_convert": [ "lm_head", "model.visual" ], "ignored_layers": [ "lm_head", "model.layers.0.mlp.down_proj", "model.layers.0.mlp.gate_proj", "model.layers.0.mlp.up_proj", "model.layers.0.self_attn.k_proj", "model.layers.0.self_attn.o_proj", "model.layers.0.self_attn.q_proj", "model.layers.0.self_attn.v_proj", "model.layers.1.self_attn.k_proj", "model.layers.1.self_attn.o_proj", "model.layers.1.self_attn.q_proj", "model.layers.1.self_attn.v_proj", "model.layers.10.self_attn.k_proj", "model.layers.10.self_attn.o_proj", "model.layers.10.self_attn.q_proj", "model.layers.10.self_attn.v_proj", "model.layers.11.self_attn.k_proj", "model.layers.11.self_attn.o_proj", "model.layers.11.self_attn.q_proj", "model.layers.11.self_attn.v_proj", "model.layers.12.self_attn.k_proj", "model.layers.12.self_attn.o_proj", "model.layers.12.self_attn.q_proj", "model.layers.12.self_attn.v_proj", "model.layers.13.self_attn.k_proj", "model.layers.13.self_attn.o_proj", "model.layers.13.self_attn.q_proj", "model.layers.13.self_attn.v_proj", "model.layers.14.self_attn.k_proj", "model.layers.14.self_attn.o_proj", "model.layers.14.self_attn.q_proj", "model.layers.14.self_attn.v_proj", "model.layers.15.self_attn.k_proj", "model.layers.15.self_attn.o_proj", "model.layers.15.self_attn.q_proj", "model.layers.15.self_attn.v_proj", "model.layers.16.self_attn.k_proj", "model.layers.16.self_attn.o_proj", "model.layers.16.self_attn.q_proj", "model.layers.16.self_attn.v_proj", "model.layers.17.self_attn.k_proj", "model.layers.17.self_attn.o_proj", "model.layers.17.self_attn.q_proj", "model.layers.17.self_attn.v_proj", "model.layers.18.self_attn.k_proj", "model.layers.18.self_attn.o_proj", "model.layers.18.self_attn.q_proj", "model.layers.18.self_attn.v_proj", "model.layers.19.self_attn.k_proj", "model.layers.19.self_attn.o_proj", "model.layers.19.self_attn.q_proj", "model.layers.19.self_attn.v_proj", "model.layers.2.self_attn.k_proj", "model.layers.2.self_attn.o_proj", "model.layers.2.self_attn.q_proj", "model.layers.2.self_attn.v_proj", "model.layers.20.self_attn.k_proj", "model.layers.20.self_attn.o_proj", "model.layers.20.self_attn.q_proj", "model.layers.20.self_attn.v_proj", "model.layers.21.self_attn.k_proj", "model.layers.21.self_attn.o_proj", "model.layers.21.self_attn.q_proj", "model.layers.21.self_attn.v_proj", "model.layers.22.self_attn.k_proj", "model.layers.22.self_attn.o_proj", "model.layers.22.self_attn.q_proj", "model.layers.22.self_attn.v_proj", "model.layers.23.self_attn.k_proj", "model.layers.23.self_attn.o_proj", "model.layers.23.self_attn.q_proj", "model.layers.23.self_attn.v_proj", "model.layers.24.self_attn.k_proj", "model.layers.24.self_attn.o_proj", "model.layers.24.self_attn.q_proj", "model.layers.24.self_attn.v_proj", "model.layers.25.self_attn.k_proj", "model.layers.25.self_attn.o_proj", "model.layers.25.self_attn.q_proj", "model.layers.25.self_attn.v_proj", "model.layers.26.self_attn.k_proj", "model.layers.26.self_attn.o_proj", "model.layers.26.self_attn.q_proj", "model.layers.26.self_attn.v_proj", "model.layers.27.self_attn.k_proj", "model.layers.27.self_attn.o_proj", "model.layers.27.self_attn.q_proj", "model.layers.27.self_attn.v_proj", "model.layers.28.self_attn.k_proj", "model.layers.28.self_attn.o_proj", "model.layers.28.self_attn.q_proj", "model.layers.28.self_attn.v_proj", "model.layers.29.self_attn.k_proj", "model.layers.29.self_attn.o_proj", "model.layers.29.self_attn.q_proj", "model.layers.29.self_attn.v_proj", "model.layers.3.self_attn.k_proj", "model.layers.3.self_attn.o_proj", "model.layers.3.self_attn.q_proj", "model.layers.3.self_attn.v_proj", "model.layers.30.self_attn.k_proj", "model.layers.30.self_attn.o_proj", "model.layers.30.self_attn.q_proj", "model.layers.30.self_attn.v_proj", "model.layers.31.self_attn.k_proj", "model.layers.31.self_attn.o_proj", "model.layers.31.self_attn.q_proj", "model.layers.31.self_attn.v_proj", "model.layers.32.self_attn.k_proj", "model.layers.32.self_attn.o_proj", "model.layers.32.self_attn.q_proj", "model.layers.32.self_attn.v_proj", "model.layers.33.self_attn.k_proj", "model.layers.33.self_attn.o_proj", "model.layers.33.self_attn.q_proj", "model.layers.33.self_attn.v_proj", "model.layers.34.self_attn.k_proj", "model.layers.34.self_attn.o_proj", "model.layers.34.self_attn.q_proj", "model.layers.34.self_attn.v_proj", "model.layers.35.self_attn.k_proj", "model.layers.35.self_attn.o_proj", "model.layers.35.self_attn.q_proj", "model.layers.35.self_attn.v_proj", "model.layers.4.self_attn.k_proj", "model.layers.4.self_attn.o_proj", "model.layers.4.self_attn.q_proj", "model.layers.4.self_attn.v_proj", "model.layers.5.self_attn.k_proj", "model.layers.5.self_attn.o_proj", "model.layers.5.self_attn.q_proj", "model.layers.5.self_attn.v_proj", "model.layers.6.self_attn.k_proj", "model.layers.6.self_attn.o_proj", "model.layers.6.self_attn.q_proj", "model.layers.6.self_attn.v_proj", "model.layers.7.self_attn.k_proj", "model.layers.7.self_attn.o_proj", "model.layers.7.self_attn.q_proj", "model.layers.7.self_attn.v_proj", "model.layers.8.self_attn.k_proj", "model.layers.8.self_attn.o_proj", "model.layers.8.self_attn.q_proj", "model.layers.8.self_attn.v_proj", "model.layers.9.self_attn.k_proj", "model.layers.9.self_attn.o_proj", "model.layers.9.self_attn.q_proj", "model.layers.9.self_attn.v_proj", "model.vision_model", "model.vision_model.conv1", "model.vision_model.ln_pre", "model.vision_model.transformer.resblocks.0.attn.out_proj", "model.vision_model.transformer.resblocks.0.attn.qkv_proj", "model.vision_model.transformer.resblocks.0.ln_1", "model.vision_model.transformer.resblocks.0.ln_2", "model.vision_model.transformer.resblocks.0.mlp.c_fc", "model.vision_model.transformer.resblocks.0.mlp.c_proj", "model.vision_model.transformer.resblocks.1.attn.out_proj", "model.vision_model.transformer.resblocks.1.attn.qkv_proj", "model.vision_model.transformer.resblocks.1.ln_1", "model.vision_model.transformer.resblocks.1.ln_2", "model.vision_model.transformer.resblocks.1.mlp.c_fc", "model.vision_model.transformer.resblocks.1.mlp.c_proj", "model.vision_model.transformer.resblocks.10.attn.out_proj", "model.vision_model.transformer.resblocks.10.attn.qkv_proj", "model.vision_model.transformer.resblocks.10.ln_1", "model.vision_model.transformer.resblocks.10.ln_2", "model.vision_model.transformer.resblocks.10.mlp.c_fc", "model.vision_model.transformer.resblocks.10.mlp.c_proj", "model.vision_model.transformer.resblocks.11.attn.out_proj", "model.vision_model.transformer.resblocks.11.attn.qkv_proj", "model.vision_model.transformer.resblocks.11.ln_1", "model.vision_model.transformer.resblocks.11.ln_2", "model.vision_model.transformer.resblocks.11.mlp.c_fc", "model.vision_model.transformer.resblocks.11.mlp.c_proj", "model.vision_model.transformer.resblocks.12.attn.out_proj", "model.vision_model.transformer.resblocks.12.attn.qkv_proj", "model.vision_model.transformer.resblocks.12.ln_1", "model.vision_model.transformer.resblocks.12.ln_2", "model.vision_model.transformer.resblocks.12.mlp.c_fc", "model.vision_model.transformer.resblocks.12.mlp.c_proj", "model.vision_model.transformer.resblocks.13.attn.out_proj", "model.vision_model.transformer.resblocks.13.attn.qkv_proj", "model.vision_model.transformer.resblocks.13.ln_1", "model.vision_model.transformer.resblocks.13.ln_2", "model.vision_model.transformer.resblocks.13.mlp.c_fc", "model.vision_model.transformer.resblocks.13.mlp.c_proj", "model.vision_model.transformer.resblocks.14.attn.out_proj", "model.vision_model.transformer.resblocks.14.attn.qkv_proj", "model.vision_model.transformer.resblocks.14.ln_1", "model.vision_model.transformer.resblocks.14.ln_2", "model.vision_model.transformer.resblocks.14.mlp.c_fc", "model.vision_model.transformer.resblocks.14.mlp.c_proj", "model.vision_model.transformer.resblocks.15.attn.out_proj", "model.vision_model.transformer.resblocks.15.attn.qkv_proj", "model.vision_model.transformer.resblocks.15.ln_1", "model.vision_model.transformer.resblocks.15.ln_2", "model.vision_model.transformer.resblocks.15.mlp.c_fc", "model.vision_model.transformer.resblocks.15.mlp.c_proj", "model.vision_model.transformer.resblocks.16.attn.out_proj", "model.vision_model.transformer.resblocks.16.attn.qkv_proj", "model.vision_model.transformer.resblocks.16.ln_1", "model.vision_model.transformer.resblocks.16.ln_2", "model.vision_model.transformer.resblocks.16.mlp.c_fc", "model.vision_model.transformer.resblocks.16.mlp.c_proj", "model.vision_model.transformer.resblocks.17.attn.out_proj", "model.vision_model.transformer.resblocks.17.attn.qkv_proj", "model.vision_model.transformer.resblocks.17.ln_1", "model.vision_model.transformer.resblocks.17.ln_2", "model.vision_model.transformer.resblocks.17.mlp.c_fc", "model.vision_model.transformer.resblocks.17.mlp.c_proj", "model.vision_model.transformer.resblocks.18.attn.out_proj", "model.vision_model.transformer.resblocks.18.attn.qkv_proj", "model.vision_model.transformer.resblocks.18.ln_1", "model.vision_model.transformer.resblocks.18.ln_2", "model.vision_model.transformer.resblocks.18.mlp.c_fc", "model.vision_model.transformer.resblocks.18.mlp.c_proj", "model.vision_model.transformer.resblocks.19.attn.out_proj", "model.vision_model.transformer.resblocks.19.attn.qkv_proj", "model.vision_model.transformer.resblocks.19.ln_1", "model.vision_model.transformer.resblocks.19.ln_2", "model.vision_model.transformer.resblocks.19.mlp.c_fc", "model.vision_model.transformer.resblocks.19.mlp.c_proj", "model.vision_model.transformer.resblocks.2.attn.out_proj", "model.vision_model.transformer.resblocks.2.attn.qkv_proj", "model.vision_model.transformer.resblocks.2.ln_1", "model.vision_model.transformer.resblocks.2.ln_2", "model.vision_model.transformer.resblocks.2.mlp.c_fc", "model.vision_model.transformer.resblocks.2.mlp.c_proj", "model.vision_model.transformer.resblocks.20.attn.out_proj", "model.vision_model.transformer.resblocks.20.attn.qkv_proj", "model.vision_model.transformer.resblocks.20.ln_1", "model.vision_model.transformer.resblocks.20.ln_2", "model.vision_model.transformer.resblocks.20.mlp.c_fc", "model.vision_model.transformer.resblocks.20.mlp.c_proj", "model.vision_model.transformer.resblocks.21.attn.out_proj", "model.vision_model.transformer.resblocks.21.attn.qkv_proj", "model.vision_model.transformer.resblocks.21.ln_1", "model.vision_model.transformer.resblocks.21.ln_2", "model.vision_model.transformer.resblocks.21.mlp.c_fc", "model.vision_model.transformer.resblocks.21.mlp.c_proj", "model.vision_model.transformer.resblocks.22.attn.out_proj", "model.vision_model.transformer.resblocks.22.attn.qkv_proj", "model.vision_model.transformer.resblocks.22.ln_1", "model.vision_model.transformer.resblocks.22.ln_2", "model.vision_model.transformer.resblocks.22.mlp.c_fc", "model.vision_model.transformer.resblocks.22.mlp.c_proj", "model.vision_model.transformer.resblocks.23.attn.out_proj", "model.vision_model.transformer.resblocks.23.attn.qkv_proj", "model.vision_model.transformer.resblocks.23.ln_1", "model.vision_model.transformer.resblocks.23.ln_2", "model.vision_model.transformer.resblocks.23.mlp.c_fc", "model.vision_model.transformer.resblocks.23.mlp.c_proj", "model.vision_model.transformer.resblocks.24.attn.out_proj", "model.vision_model.transformer.resblocks.24.attn.qkv_proj", "model.vision_model.transformer.resblocks.24.ln_1", "model.vision_model.transformer.resblocks.24.ln_2", "model.vision_model.transformer.resblocks.24.mlp.c_fc", "model.vision_model.transformer.resblocks.24.mlp.c_proj", "model.vision_model.transformer.resblocks.25.attn.out_proj", "model.vision_model.transformer.resblocks.25.attn.qkv_proj", "model.vision_model.transformer.resblocks.25.ln_1", "model.vision_model.transformer.resblocks.25.ln_2", "model.vision_model.transformer.resblocks.25.mlp.c_fc", "model.vision_model.transformer.resblocks.25.mlp.c_proj", "model.vision_model.transformer.resblocks.26.attn.out_proj", "model.vision_model.transformer.resblocks.26.attn.qkv_proj", "model.vision_model.transformer.resblocks.26.ln_1", "model.vision_model.transformer.resblocks.26.ln_2", "model.vision_model.transformer.resblocks.26.mlp.c_fc", "model.vision_model.transformer.resblocks.26.mlp.c_proj", "model.vision_model.transformer.resblocks.27.attn.out_proj", "model.vision_model.transformer.resblocks.27.attn.qkv_proj", "model.vision_model.transformer.resblocks.27.ln_1", "model.vision_model.transformer.resblocks.27.ln_2", "model.vision_model.transformer.resblocks.27.mlp.c_fc", "model.vision_model.transformer.resblocks.27.mlp.c_proj", "model.vision_model.transformer.resblocks.28.attn.out_proj", "model.vision_model.transformer.resblocks.28.attn.qkv_proj", "model.vision_model.transformer.resblocks.28.ln_1", "model.vision_model.transformer.resblocks.28.ln_2", "model.vision_model.transformer.resblocks.28.mlp.c_fc", "model.vision_model.transformer.resblocks.28.mlp.c_proj", "model.vision_model.transformer.resblocks.29.attn.out_proj", "model.vision_model.transformer.resblocks.29.attn.qkv_proj", "model.vision_model.transformer.resblocks.29.ln_1", "model.vision_model.transformer.resblocks.29.ln_2", "model.vision_model.transformer.resblocks.29.mlp.c_fc", "model.vision_model.transformer.resblocks.29.mlp.c_proj", "model.vision_model.transformer.resblocks.3.attn.out_proj", "model.vision_model.transformer.resblocks.3.attn.qkv_proj", "model.vision_model.transformer.resblocks.3.ln_1", "model.vision_model.transformer.resblocks.3.ln_2", "model.vision_model.transformer.resblocks.3.mlp.c_fc", "model.vision_model.transformer.resblocks.3.mlp.c_proj", "model.vision_model.transformer.resblocks.30.attn.out_proj", "model.vision_model.transformer.resblocks.30.attn.qkv_proj", "model.vision_model.transformer.resblocks.30.ln_1", "model.vision_model.transformer.resblocks.30.ln_2", "model.vision_model.transformer.resblocks.30.mlp.c_fc", "model.vision_model.transformer.resblocks.30.mlp.c_proj", "model.vision_model.transformer.resblocks.31.attn.out_proj", "model.vision_model.transformer.resblocks.31.attn.qkv_proj", "model.vision_model.transformer.resblocks.31.ln_1", "model.vision_model.transformer.resblocks.31.ln_2", "model.vision_model.transformer.resblocks.31.mlp.c_fc", "model.vision_model.transformer.resblocks.31.mlp.c_proj", "model.vision_model.transformer.resblocks.32.attn.out_proj", "model.vision_model.transformer.resblocks.32.attn.qkv_proj", "model.vision_model.transformer.resblocks.32.ln_1", "model.vision_model.transformer.resblocks.32.ln_2", "model.vision_model.transformer.resblocks.32.mlp.c_fc", "model.vision_model.transformer.resblocks.32.mlp.c_proj", "model.vision_model.transformer.resblocks.33.attn.out_proj", "model.vision_model.transformer.resblocks.33.attn.qkv_proj", "model.vision_model.transformer.resblocks.33.ln_1", "model.vision_model.transformer.resblocks.33.ln_2", "model.vision_model.transformer.resblocks.33.mlp.c_fc", "model.vision_model.transformer.resblocks.33.mlp.c_proj", "model.vision_model.transformer.resblocks.34.attn.out_proj", "model.vision_model.transformer.resblocks.34.attn.qkv_proj", "model.vision_model.transformer.resblocks.34.ln_1", "model.vision_model.transformer.resblocks.34.ln_2", "model.vision_model.transformer.resblocks.34.mlp.c_fc", "model.vision_model.transformer.resblocks.34.mlp.c_proj", "model.vision_model.transformer.resblocks.35.attn.out_proj", "model.vision_model.transformer.resblocks.35.attn.qkv_proj", "model.vision_model.transformer.resblocks.35.ln_1", "model.vision_model.transformer.resblocks.35.ln_2", "model.vision_model.transformer.resblocks.35.mlp.c_fc", "model.vision_model.transformer.resblocks.35.mlp.c_proj", "model.vision_model.transformer.resblocks.36.attn.out_proj", "model.vision_model.transformer.resblocks.36.attn.qkv_proj", "model.vision_model.transformer.resblocks.36.ln_1", "model.vision_model.transformer.resblocks.36.ln_2", "model.vision_model.transformer.resblocks.36.mlp.c_fc", "model.vision_model.transformer.resblocks.36.mlp.c_proj", "model.vision_model.transformer.resblocks.37.attn.out_proj", "model.vision_model.transformer.resblocks.37.attn.qkv_proj", "model.vision_model.transformer.resblocks.37.ln_1", "model.vision_model.transformer.resblocks.37.ln_2", "model.vision_model.transformer.resblocks.37.mlp.c_fc", "model.vision_model.transformer.resblocks.37.mlp.c_proj", "model.vision_model.transformer.resblocks.38.attn.out_proj", "model.vision_model.transformer.resblocks.38.attn.qkv_proj", "model.vision_model.transformer.resblocks.38.ln_1", "model.vision_model.transformer.resblocks.38.ln_2", "model.vision_model.transformer.resblocks.38.mlp.c_fc", "model.vision_model.transformer.resblocks.38.mlp.c_proj", "model.vision_model.transformer.resblocks.39.attn.out_proj", "model.vision_model.transformer.resblocks.39.attn.qkv_proj", "model.vision_model.transformer.resblocks.39.ln_1", "model.vision_model.transformer.resblocks.39.ln_2", "model.vision_model.transformer.resblocks.39.mlp.c_fc", "model.vision_model.transformer.resblocks.39.mlp.c_proj", "model.vision_model.transformer.resblocks.4.attn.out_proj", "model.vision_model.transformer.resblocks.4.attn.qkv_proj", "model.vision_model.transformer.resblocks.4.ln_1", "model.vision_model.transformer.resblocks.4.ln_2", "model.vision_model.transformer.resblocks.4.mlp.c_fc", "model.vision_model.transformer.resblocks.4.mlp.c_proj", "model.vision_model.transformer.resblocks.40.attn.out_proj", "model.vision_model.transformer.resblocks.40.attn.qkv_proj", "model.vision_model.transformer.resblocks.40.ln_1", "model.vision_model.transformer.resblocks.40.ln_2", "model.vision_model.transformer.resblocks.40.mlp.c_fc", "model.vision_model.transformer.resblocks.40.mlp.c_proj", "model.vision_model.transformer.resblocks.41.attn.out_proj", "model.vision_model.transformer.resblocks.41.attn.qkv_proj", "model.vision_model.transformer.resblocks.41.ln_1", "model.vision_model.transformer.resblocks.41.ln_2", "model.vision_model.transformer.resblocks.41.mlp.c_fc", "model.vision_model.transformer.resblocks.41.mlp.c_proj", "model.vision_model.transformer.resblocks.42.attn.out_proj", "model.vision_model.transformer.resblocks.42.attn.qkv_proj", "model.vision_model.transformer.resblocks.42.ln_1", "model.vision_model.transformer.resblocks.42.ln_2", "model.vision_model.transformer.resblocks.42.mlp.c_fc", "model.vision_model.transformer.resblocks.42.mlp.c_proj", "model.vision_model.transformer.resblocks.43.attn.out_proj", "model.vision_model.transformer.resblocks.43.attn.qkv_proj", "model.vision_model.transformer.resblocks.43.ln_1", "model.vision_model.transformer.resblocks.43.ln_2", "model.vision_model.transformer.resblocks.43.mlp.c_fc", "model.vision_model.transformer.resblocks.43.mlp.c_proj", "model.vision_model.transformer.resblocks.44.attn.out_proj", "model.vision_model.transformer.resblocks.44.attn.qkv_proj", "model.vision_model.transformer.resblocks.44.ln_1", "model.vision_model.transformer.resblocks.44.ln_2", "model.vision_model.transformer.resblocks.44.mlp.c_fc", "model.vision_model.transformer.resblocks.44.mlp.c_proj", "model.vision_model.transformer.resblocks.45.attn.out_proj", "model.vision_model.transformer.resblocks.45.attn.qkv_proj", "model.vision_model.transformer.resblocks.45.ln_1", "model.vision_model.transformer.resblocks.45.ln_2", "model.vision_model.transformer.resblocks.45.mlp.c_fc", "model.vision_model.transformer.resblocks.45.mlp.c_proj", "model.vision_model.transformer.resblocks.46.attn.out_proj", "model.vision_model.transformer.resblocks.46.attn.qkv_proj", "model.vision_model.transformer.resblocks.46.ln_1", "model.vision_model.transformer.resblocks.46.ln_2", "model.vision_model.transformer.resblocks.46.mlp.c_fc", "model.vision_model.transformer.resblocks.46.mlp.c_proj", "model.vision_model.transformer.resblocks.5.attn.out_proj", "model.vision_model.transformer.resblocks.5.attn.qkv_proj", "model.vision_model.transformer.resblocks.5.ln_1", "model.vision_model.transformer.resblocks.5.ln_2", "model.vision_model.transformer.resblocks.5.mlp.c_fc", "model.vision_model.transformer.resblocks.5.mlp.c_proj", "model.vision_model.transformer.resblocks.6.attn.out_proj", "model.vision_model.transformer.resblocks.6.attn.qkv_proj", "model.vision_model.transformer.resblocks.6.ln_1", "model.vision_model.transformer.resblocks.6.ln_2", "model.vision_model.transformer.resblocks.6.mlp.c_fc", "model.vision_model.transformer.resblocks.6.mlp.c_proj", "model.vision_model.transformer.resblocks.7.attn.out_proj", "model.vision_model.transformer.resblocks.7.attn.qkv_proj", "model.vision_model.transformer.resblocks.7.ln_1", "model.vision_model.transformer.resblocks.7.ln_2", "model.vision_model.transformer.resblocks.7.mlp.c_fc", "model.vision_model.transformer.resblocks.7.mlp.c_proj", "model.vision_model.transformer.resblocks.8.attn.out_proj", "model.vision_model.transformer.resblocks.8.attn.qkv_proj", "model.vision_model.transformer.resblocks.8.ln_1", "model.vision_model.transformer.resblocks.8.ln_2", "model.vision_model.transformer.resblocks.8.mlp.c_fc", "model.vision_model.transformer.resblocks.8.mlp.c_proj", "model.vision_model.transformer.resblocks.9.attn.out_proj", "model.vision_model.transformer.resblocks.9.attn.qkv_proj", "model.vision_model.transformer.resblocks.9.ln_1", "model.vision_model.transformer.resblocks.9.ln_2", "model.vision_model.transformer.resblocks.9.mlp.c_fc", "model.vision_model.transformer.resblocks.9.mlp.c_proj", "model.vision_model.vit_downsampler1", "model.vision_model.vit_downsampler2", "model.vit_large_projector", "vision_model", "vision_model.conv1", "vision_model.ln_pre", "vision_model.transformer.resblocks.0.attn.out_proj", "vision_model.transformer.resblocks.0.attn.qkv_proj", "vision_model.transformer.resblocks.0.ln_1", "vision_model.transformer.resblocks.0.ln_2", "vision_model.transformer.resblocks.0.mlp.c_fc", "vision_model.transformer.resblocks.0.mlp.c_proj", "vision_model.transformer.resblocks.1.attn.out_proj", "vision_model.transformer.resblocks.1.attn.qkv_proj", "vision_model.transformer.resblocks.1.ln_1", "vision_model.transformer.resblocks.1.ln_2", "vision_model.transformer.resblocks.1.mlp.c_fc", "vision_model.transformer.resblocks.1.mlp.c_proj", "vision_model.transformer.resblocks.10.attn.out_proj", "vision_model.transformer.resblocks.10.attn.qkv_proj", "vision_model.transformer.resblocks.10.ln_1", "vision_model.transformer.resblocks.10.ln_2", "vision_model.transformer.resblocks.10.mlp.c_fc", "vision_model.transformer.resblocks.10.mlp.c_proj", "vision_model.transformer.resblocks.11.attn.out_proj", "vision_model.transformer.resblocks.11.attn.qkv_proj", "vision_model.transformer.resblocks.11.ln_1", "vision_model.transformer.resblocks.11.ln_2", "vision_model.transformer.resblocks.11.mlp.c_fc", "vision_model.transformer.resblocks.11.mlp.c_proj", "vision_model.transformer.resblocks.12.attn.out_proj", "vision_model.transformer.resblocks.12.attn.qkv_proj", "vision_model.transformer.resblocks.12.ln_1", "vision_model.transformer.resblocks.12.ln_2", "vision_model.transformer.resblocks.12.mlp.c_fc", "vision_model.transformer.resblocks.12.mlp.c_proj", "vision_model.transformer.resblocks.13.attn.out_proj", "vision_model.transformer.resblocks.13.attn.qkv_proj", "vision_model.transformer.resblocks.13.ln_1", "vision_model.transformer.resblocks.13.ln_2", "vision_model.transformer.resblocks.13.mlp.c_fc", "vision_model.transformer.resblocks.13.mlp.c_proj", "vision_model.transformer.resblocks.14.attn.out_proj", "vision_model.transformer.resblocks.14.attn.qkv_proj", "vision_model.transformer.resblocks.14.ln_1", "vision_model.transformer.resblocks.14.ln_2", "vision_model.transformer.resblocks.14.mlp.c_fc", "vision_model.transformer.resblocks.14.mlp.c_proj", "vision_model.transformer.resblocks.15.attn.out_proj", "vision_model.transformer.resblocks.15.attn.qkv_proj", "vision_model.transformer.resblocks.15.ln_1", "vision_model.transformer.resblocks.15.ln_2", "vision_model.transformer.resblocks.15.mlp.c_fc", "vision_model.transformer.resblocks.15.mlp.c_proj", "vision_model.transformer.resblocks.16.attn.out_proj", "vision_model.transformer.resblocks.16.attn.qkv_proj", "vision_model.transformer.resblocks.16.ln_1", "vision_model.transformer.resblocks.16.ln_2", "vision_model.transformer.resblocks.16.mlp.c_fc", "vision_model.transformer.resblocks.16.mlp.c_proj", "vision_model.transformer.resblocks.17.attn.out_proj", "vision_model.transformer.resblocks.17.attn.qkv_proj", "vision_model.transformer.resblocks.17.ln_1", "vision_model.transformer.resblocks.17.ln_2", "vision_model.transformer.resblocks.17.mlp.c_fc", "vision_model.transformer.resblocks.17.mlp.c_proj", "vision_model.transformer.resblocks.18.attn.out_proj", "vision_model.transformer.resblocks.18.attn.qkv_proj", "vision_model.transformer.resblocks.18.ln_1", "vision_model.transformer.resblocks.18.ln_2", "vision_model.transformer.resblocks.18.mlp.c_fc", "vision_model.transformer.resblocks.18.mlp.c_proj", "vision_model.transformer.resblocks.19.attn.out_proj", "vision_model.transformer.resblocks.19.attn.qkv_proj", "vision_model.transformer.resblocks.19.ln_1", "vision_model.transformer.resblocks.19.ln_2", "vision_model.transformer.resblocks.19.mlp.c_fc", "vision_model.transformer.resblocks.19.mlp.c_proj", "vision_model.transformer.resblocks.2.attn.out_proj", "vision_model.transformer.resblocks.2.attn.qkv_proj", "vision_model.transformer.resblocks.2.ln_1", "vision_model.transformer.resblocks.2.ln_2", "vision_model.transformer.resblocks.2.mlp.c_fc", "vision_model.transformer.resblocks.2.mlp.c_proj", "vision_model.transformer.resblocks.20.attn.out_proj", "vision_model.transformer.resblocks.20.attn.qkv_proj", "vision_model.transformer.resblocks.20.ln_1", "vision_model.transformer.resblocks.20.ln_2", "vision_model.transformer.resblocks.20.mlp.c_fc", "vision_model.transformer.resblocks.20.mlp.c_proj", "vision_model.transformer.resblocks.21.attn.out_proj", "vision_model.transformer.resblocks.21.attn.qkv_proj", "vision_model.transformer.resblocks.21.ln_1", "vision_model.transformer.resblocks.21.ln_2", "vision_model.transformer.resblocks.21.mlp.c_fc", "vision_model.transformer.resblocks.21.mlp.c_proj", "vision_model.transformer.resblocks.22.attn.out_proj", "vision_model.transformer.resblocks.22.attn.qkv_proj", "vision_model.transformer.resblocks.22.ln_1", "vision_model.transformer.resblocks.22.ln_2", "vision_model.transformer.resblocks.22.mlp.c_fc", "vision_model.transformer.resblocks.22.mlp.c_proj", "vision_model.transformer.resblocks.23.attn.out_proj", "vision_model.transformer.resblocks.23.attn.qkv_proj", "vision_model.transformer.resblocks.23.ln_1", "vision_model.transformer.resblocks.23.ln_2", "vision_model.transformer.resblocks.23.mlp.c_fc", "vision_model.transformer.resblocks.23.mlp.c_proj", "vision_model.transformer.resblocks.24.attn.out_proj", "vision_model.transformer.resblocks.24.attn.qkv_proj", "vision_model.transformer.resblocks.24.ln_1", "vision_model.transformer.resblocks.24.ln_2", "vision_model.transformer.resblocks.24.mlp.c_fc", "vision_model.transformer.resblocks.24.mlp.c_proj", "vision_model.transformer.resblocks.25.attn.out_proj", "vision_model.transformer.resblocks.25.attn.qkv_proj", "vision_model.transformer.resblocks.25.ln_1", "vision_model.transformer.resblocks.25.ln_2", "vision_model.transformer.resblocks.25.mlp.c_fc", "vision_model.transformer.resblocks.25.mlp.c_proj", "vision_model.transformer.resblocks.26.attn.out_proj", "vision_model.transformer.resblocks.26.attn.qkv_proj", "vision_model.transformer.resblocks.26.ln_1", "vision_model.transformer.resblocks.26.ln_2", "vision_model.transformer.resblocks.26.mlp.c_fc", "vision_model.transformer.resblocks.26.mlp.c_proj", "vision_model.transformer.resblocks.27.attn.out_proj", "vision_model.transformer.resblocks.27.attn.qkv_proj", "vision_model.transformer.resblocks.27.ln_1", "vision_model.transformer.resblocks.27.ln_2", "vision_model.transformer.resblocks.27.mlp.c_fc", "vision_model.transformer.resblocks.27.mlp.c_proj", "vision_model.transformer.resblocks.28.attn.out_proj", "vision_model.transformer.resblocks.28.attn.qkv_proj", "vision_model.transformer.resblocks.28.ln_1", "vision_model.transformer.resblocks.28.ln_2", "vision_model.transformer.resblocks.28.mlp.c_fc", "vision_model.transformer.resblocks.28.mlp.c_proj", "vision_model.transformer.resblocks.29.attn.out_proj", "vision_model.transformer.resblocks.29.attn.qkv_proj", "vision_model.transformer.resblocks.29.ln_1", "vision_model.transformer.resblocks.29.ln_2", "vision_model.transformer.resblocks.29.mlp.c_fc", "vision_model.transformer.resblocks.29.mlp.c_proj", "vision_model.transformer.resblocks.3.attn.out_proj", "vision_model.transformer.resblocks.3.attn.qkv_proj", "vision_model.transformer.resblocks.3.ln_1", "vision_model.transformer.resblocks.3.ln_2", "vision_model.transformer.resblocks.3.mlp.c_fc", "vision_model.transformer.resblocks.3.mlp.c_proj", "vision_model.transformer.resblocks.30.attn.out_proj", "vision_model.transformer.resblocks.30.attn.qkv_proj", "vision_model.transformer.resblocks.30.ln_1", "vision_model.transformer.resblocks.30.ln_2", "vision_model.transformer.resblocks.30.mlp.c_fc", "vision_model.transformer.resblocks.30.mlp.c_proj", "vision_model.transformer.resblocks.31.attn.out_proj", "vision_model.transformer.resblocks.31.attn.qkv_proj", "vision_model.transformer.resblocks.31.ln_1", "vision_model.transformer.resblocks.31.ln_2", "vision_model.transformer.resblocks.31.mlp.c_fc", "vision_model.transformer.resblocks.31.mlp.c_proj", "vision_model.transformer.resblocks.32.attn.out_proj", "vision_model.transformer.resblocks.32.attn.qkv_proj", "vision_model.transformer.resblocks.32.ln_1", "vision_model.transformer.resblocks.32.ln_2", "vision_model.transformer.resblocks.32.mlp.c_fc", "vision_model.transformer.resblocks.32.mlp.c_proj", "vision_model.transformer.resblocks.33.attn.out_proj", "vision_model.transformer.resblocks.33.attn.qkv_proj", "vision_model.transformer.resblocks.33.ln_1", "vision_model.transformer.resblocks.33.ln_2", "vision_model.transformer.resblocks.33.mlp.c_fc", "vision_model.transformer.resblocks.33.mlp.c_proj", "vision_model.transformer.resblocks.34.attn.out_proj", "vision_model.transformer.resblocks.34.attn.qkv_proj", "vision_model.transformer.resblocks.34.ln_1", "vision_model.transformer.resblocks.34.ln_2", "vision_model.transformer.resblocks.34.mlp.c_fc", "vision_model.transformer.resblocks.34.mlp.c_proj", "vision_model.transformer.resblocks.35.attn.out_proj", "vision_model.transformer.resblocks.35.attn.qkv_proj", "vision_model.transformer.resblocks.35.ln_1", "vision_model.transformer.resblocks.35.ln_2", "vision_model.transformer.resblocks.35.mlp.c_fc", "vision_model.transformer.resblocks.35.mlp.c_proj", "vision_model.transformer.resblocks.36.attn.out_proj", "vision_model.transformer.resblocks.36.attn.qkv_proj", "vision_model.transformer.resblocks.36.ln_1", "vision_model.transformer.resblocks.36.ln_2", "vision_model.transformer.resblocks.36.mlp.c_fc", "vision_model.transformer.resblocks.36.mlp.c_proj", "vision_model.transformer.resblocks.37.attn.out_proj", "vision_model.transformer.resblocks.37.attn.qkv_proj", "vision_model.transformer.resblocks.37.ln_1", "vision_model.transformer.resblocks.37.ln_2", "vision_model.transformer.resblocks.37.mlp.c_fc", "vision_model.transformer.resblocks.37.mlp.c_proj", "vision_model.transformer.resblocks.38.attn.out_proj", "vision_model.transformer.resblocks.38.attn.qkv_proj", "vision_model.transformer.resblocks.38.ln_1", "vision_model.transformer.resblocks.38.ln_2", "vision_model.transformer.resblocks.38.mlp.c_fc", "vision_model.transformer.resblocks.38.mlp.c_proj", "vision_model.transformer.resblocks.39.attn.out_proj", "vision_model.transformer.resblocks.39.attn.qkv_proj", "vision_model.transformer.resblocks.39.ln_1", "vision_model.transformer.resblocks.39.ln_2", "vision_model.transformer.resblocks.39.mlp.c_fc", "vision_model.transformer.resblocks.39.mlp.c_proj", "vision_model.transformer.resblocks.4.attn.out_proj", "vision_model.transformer.resblocks.4.attn.qkv_proj", "vision_model.transformer.resblocks.4.ln_1", "vision_model.transformer.resblocks.4.ln_2", "vision_model.transformer.resblocks.4.mlp.c_fc", "vision_model.transformer.resblocks.4.mlp.c_proj", "vision_model.transformer.resblocks.40.attn.out_proj", "vision_model.transformer.resblocks.40.attn.qkv_proj", "vision_model.transformer.resblocks.40.ln_1", "vision_model.transformer.resblocks.40.ln_2", "vision_model.transformer.resblocks.40.mlp.c_fc", "vision_model.transformer.resblocks.40.mlp.c_proj", "vision_model.transformer.resblocks.41.attn.out_proj", "vision_model.transformer.resblocks.41.attn.qkv_proj", "vision_model.transformer.resblocks.41.ln_1", "vision_model.transformer.resblocks.41.ln_2", "vision_model.transformer.resblocks.41.mlp.c_fc", "vision_model.transformer.resblocks.41.mlp.c_proj", "vision_model.transformer.resblocks.42.attn.out_proj", "vision_model.transformer.resblocks.42.attn.qkv_proj", "vision_model.transformer.resblocks.42.ln_1", "vision_model.transformer.resblocks.42.ln_2", "vision_model.transformer.resblocks.42.mlp.c_fc", "vision_model.transformer.resblocks.42.mlp.c_proj", "vision_model.transformer.resblocks.43.attn.out_proj", "vision_model.transformer.resblocks.43.attn.qkv_proj", "vision_model.transformer.resblocks.43.ln_1", "vision_model.transformer.resblocks.43.ln_2", "vision_model.transformer.resblocks.43.mlp.c_fc", "vision_model.transformer.resblocks.43.mlp.c_proj", "vision_model.transformer.resblocks.44.attn.out_proj", "vision_model.transformer.resblocks.44.attn.qkv_proj", "vision_model.transformer.resblocks.44.ln_1", "vision_model.transformer.resblocks.44.ln_2", "vision_model.transformer.resblocks.44.mlp.c_fc", "vision_model.transformer.resblocks.44.mlp.c_proj", "vision_model.transformer.resblocks.45.attn.out_proj", "vision_model.transformer.resblocks.45.attn.qkv_proj", "vision_model.transformer.resblocks.45.ln_1", "vision_model.transformer.resblocks.45.ln_2", "vision_model.transformer.resblocks.45.mlp.c_fc", "vision_model.transformer.resblocks.45.mlp.c_proj", "vision_model.transformer.resblocks.46.attn.out_proj", "vision_model.transformer.resblocks.46.attn.qkv_proj", "vision_model.transformer.resblocks.46.ln_1", "vision_model.transformer.resblocks.46.ln_2", "vision_model.transformer.resblocks.46.mlp.c_fc", "vision_model.transformer.resblocks.46.mlp.c_proj", "vision_model.transformer.resblocks.5.attn.out_proj", "vision_model.transformer.resblocks.5.attn.qkv_proj", "vision_model.transformer.resblocks.5.ln_1", "vision_model.transformer.resblocks.5.ln_2", "vision_model.transformer.resblocks.5.mlp.c_fc", "vision_model.transformer.resblocks.5.mlp.c_proj", "vision_model.transformer.resblocks.6.attn.out_proj", "vision_model.transformer.resblocks.6.attn.qkv_proj", "vision_model.transformer.resblocks.6.ln_1", "vision_model.transformer.resblocks.6.ln_2", "vision_model.transformer.resblocks.6.mlp.c_fc", "vision_model.transformer.resblocks.6.mlp.c_proj", "vision_model.transformer.resblocks.7.attn.out_proj", "vision_model.transformer.resblocks.7.attn.qkv_proj", "vision_model.transformer.resblocks.7.ln_1", "vision_model.transformer.resblocks.7.ln_2", "vision_model.transformer.resblocks.7.mlp.c_fc", "vision_model.transformer.resblocks.7.mlp.c_proj", "vision_model.transformer.resblocks.8.attn.out_proj", "vision_model.transformer.resblocks.8.attn.qkv_proj", "vision_model.transformer.resblocks.8.ln_1", "vision_model.transformer.resblocks.8.ln_2", "vision_model.transformer.resblocks.8.mlp.c_fc", "vision_model.transformer.resblocks.8.mlp.c_proj", "vision_model.transformer.resblocks.9.attn.out_proj", "vision_model.transformer.resblocks.9.attn.qkv_proj", "vision_model.transformer.resblocks.9.ln_1", "vision_model.transformer.resblocks.9.ln_2", "vision_model.transformer.resblocks.9.mlp.c_fc", "vision_model.transformer.resblocks.9.mlp.c_proj", "vision_model.vit_downsampler1", "vision_model.vit_downsampler2", "vit_large_projector" ] } }