{ "metadata": { "total_size": 16923948362 }, "weight_map": { "lm_head.weight": "pytorch_model-00004-of-00004.bin", "model.embed_tokens.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.11.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.11.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.20.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.input_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.mlp.down_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.mlp.up_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.23.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.23.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.23.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00002-of-00004.bin", "model.layers.24.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.30.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.input_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.mlp.down_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.mlp.up_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00003-of-00004.bin", "model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.input_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.mlp.down_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.mlp.up_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00001-of-00004.bin", "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00001-of-00004.bin", "model.motion_tokenizer.decode_task_layer.0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decode_task_layer.0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decode_task_layer.2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decode_task_layer.2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.0.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.1.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.10.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.11.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.2.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.3.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.4.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.5.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.6.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.7.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.8.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.blocks.9.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.in_proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.in_proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.norm.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.norm.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.0.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.0.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.1.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.1.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.2.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.2.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.3.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.decoder.up_sample_layers.3.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encode_task_layer.0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encode_task_layer.0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encode_task_layer.2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encode_task_layer.2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.0.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.1.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.10.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.11.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.2.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.3.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.4.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.5.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.6.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.7.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.8.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.spatial_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.spatial_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.spatial_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.spatial_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.spatial_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.temporal_attention.proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.temporal_attention.proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.temporal_attention.q_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.temporal_attention.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.blocks.9.temporal_attention.v_bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.0.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.0.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.1.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.1.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.2.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.2.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.3.conv.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.down_sample_layers.3.conv.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.in_proj.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.in_proj.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.norm.bias": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.norm.weight": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.pos_embed": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.encoder.temp_pos": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.quantize.cluster_size": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.quantize.embedding.cluster_size": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.quantize.embedding.embed_avg": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.quantize.embedding.initted": "pytorch_model-00004-of-00004.bin", "model.motion_tokenizer.quantize.embedding.weight": "pytorch_model-00004-of-00004.bin", "model.norm.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.0.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.1.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.10.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.11.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.2.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.3.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.4.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.5.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.6.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.7.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.8.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.norm0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.norm0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_causal_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_causal_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_causal_attn.qkv.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_causal_attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.key.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.key.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.query.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.query.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.value.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.blocks.9.token_cross_attn.value.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.ln_vision.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.ln_vision.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.norm.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.causal_encoder.norm.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encode_task_layer.0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encode_task_layer.0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encode_task_layer.2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encode_task_layer.2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.0.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.1.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.10.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.11.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.12.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.13.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.14.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.15.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.16.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.17.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.18.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.19.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.2.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.20.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.21.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.22.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.23.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.24.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.25.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.26.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.27.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.28.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.29.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.3.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.30.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.31.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.32.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.33.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.34.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.35.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.36.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.37.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.38.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.attn.proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.attn.proj.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.attn.q_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.attn.qkv.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.attn.v_bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.mlp.fc1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.mlp.fc1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.mlp.fc2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.mlp.fc2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.norm1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.norm1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.norm2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.39.norm2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.4.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.5.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.6.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.7.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.8.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.attn.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.attn.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.attn.q_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.attn.qkv.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.attn.v_bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.mlp.fc1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.mlp.fc1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.mlp.fc2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.mlp.fc2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.norm1.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.norm1.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.norm2.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.blocks.9.norm2.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.cls_token": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.head.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.head.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.norm.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.norm.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.encoder.patch_embed.proj.bias": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.patch_embed.proj.weight": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.encoder.pos_embed": "pytorch_model-00003-of-00004.bin", "model.visual_tokenizer.quantize.embedding.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.in_conv.0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.in_conv.0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.in_conv.1.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.in_conv.1.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.out_conv.0.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.out_conv.0.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.out_conv.2.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.out_conv.2.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.out_conv.4.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.token_predictor.out_conv.4.weight": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.vit_proj.bias": "pytorch_model-00004-of-00004.bin", "model.visual_tokenizer.vit_proj.weight": "pytorch_model-00004-of-00004.bin" } }