{ "metadata": { "total_size": 5024453236 }, "weight_map": { "beit3.encoder.embed_positions.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.embed_positions.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layer_norm.A.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layer_norm.A.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layer_norm.B.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layer_norm.B.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layers.0.ffn.A.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.A.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.A.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.A.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.A.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.A.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.B.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.B.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.B.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.B.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.B.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.ffn.B.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.final_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.final_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.final_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.final_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.inner_attn_ln.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.inner_attn_ln.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.inner_attn_ln.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.inner_attn_ln.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.k_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.k_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.k_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.k_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.out_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.out_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.out_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.out_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.q_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.q_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.q_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.q_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.v_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.v_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.v_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn.v_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.0.self_attn_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.A.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.A.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.A.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.A.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.A.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.A.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.B.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.B.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.B.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.B.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.B.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.ffn.B.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.final_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.final_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.final_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.final_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.inner_attn_ln.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.inner_attn_ln.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.inner_attn_ln.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.inner_attn_ln.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.k_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.k_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.k_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.k_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.out_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.out_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.out_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.out_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.q_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.q_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.q_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.q_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.v_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.v_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.v_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn.v_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.1.self_attn_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.A.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.A.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.A.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.A.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.A.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.A.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.B.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.B.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.B.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.B.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.B.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.ffn.B.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.final_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.final_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.final_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.final_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.inner_attn_ln.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.inner_attn_ln.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.inner_attn_ln.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.inner_attn_ln.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.k_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.k_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.k_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.k_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.out_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.out_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.out_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.out_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.q_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.q_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.q_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.q_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.v_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.v_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.v_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn.v_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.2.self_attn_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.A.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.A.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.A.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.A.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.A.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.A.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.B.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.B.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.B.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.B.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.B.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.ffn.B.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.final_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.final_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.final_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.final_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.inner_attn_ln.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.inner_attn_ln.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.inner_attn_ln.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.inner_attn_ln.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.k_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.k_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.k_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.k_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.out_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.out_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.out_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.out_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.q_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.q_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.q_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.q_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.v_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.v_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.v_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn.v_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.3.self_attn_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.A.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.A.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.A.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.A.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.A.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.A.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.B.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.B.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.B.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.B.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.B.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.ffn.B.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.final_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.final_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.final_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.final_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.inner_attn_ln.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.inner_attn_ln.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.inner_attn_ln.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.inner_attn_ln.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.k_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.k_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.k_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.k_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.out_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.out_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.out_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.out_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.q_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.q_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.q_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.q_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.v_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.v_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.v_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn.v_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.4.self_attn_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.A.fc1.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.A.fc1.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.A.fc2.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.A.fc2.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.A.ffn_layernorm.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.A.ffn_layernorm.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.ffn.B.fc1.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.ffn.B.fc1.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.ffn.B.fc2.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.ffn.B.fc2.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.ffn.B.ffn_layernorm.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.ffn.B.ffn_layernorm.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.final_layer_norm.A.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.final_layer_norm.A.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.final_layer_norm.B.bias": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.final_layer_norm.B.weight": "model-00002-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.inner_attn_ln.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.inner_attn_ln.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.inner_attn_ln.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.inner_attn_ln.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.k_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.k_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.k_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.k_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.out_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.out_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.out_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.out_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.q_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.q_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.q_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.q_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.v_proj.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.v_proj.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.v_proj.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn.v_proj.B.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn_layer_norm.A.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn_layer_norm.A.weight": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn_layer_norm.B.bias": "model-00001-of-00002.safetensors", "beit3.encoder.layers.5.self_attn_layer_norm.B.weight": "model-00001-of-00002.safetensors", "beit3.linear.bias": "model-00001-of-00002.safetensors", "beit3.linear.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.embed_positions.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layernorm_embedding.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layernorm_embedding.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.0.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.1.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.10.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.11.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.2.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.3.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.4.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.5.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.6.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.7.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.8.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.encoder_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.decoder.layers.9.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.embed_positions.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layernorm_embedding.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layernorm_embedding.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.0.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.1.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.10.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.11.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.2.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.3.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.4.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.5.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.6.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.7.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.8.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.fc1.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.fc1.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.fc2.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.fc2.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.final_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.final_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.k_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn_layer_norm.bias": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.encoder.layers.9.self_attn_layer_norm.weight": "model-00001-of-00002.safetensors", "beit3.text_embed.bartpho_word.shared.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.embeddings.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.embeddings.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.embeddings.position_embeddings.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.embeddings.position_ids": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.embeddings.word_embeddings.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.crossattention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.0.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.1.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.crossattention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.10.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.11.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.crossattention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.2.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.3.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.crossattention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.4.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.5.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.crossattention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.6.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.7.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.crossattention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.8.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.self.key.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.self.key.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.self.query.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.self.query.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.self.value.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.attention.self.value.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.intermediate.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.intermediate.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.intermediate_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.intermediate_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output_query.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output_query.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output_query.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.bert.encoder.layer.9.output_query.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.cls.predictions.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.cls.predictions.transform.LayerNorm.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.cls.predictions.transform.LayerNorm.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.cls.predictions.transform.dense.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.Qformer.cls.predictions.transform.dense.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.itm_head.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.itm_head.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.ln_vision.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.ln_vision.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.query_tokens": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.temp": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.text_proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.text_proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.vision_proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.vision_proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.0.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.1.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.10.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.11.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.12.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.13.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.14.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.15.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.16.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.17.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.18.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.19.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.2.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.20.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.21.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.22.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.23.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.24.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.25.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.26.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.27.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.28.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.29.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.3.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.30.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.31.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.32.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.33.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.34.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.35.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.36.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.37.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.38.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.4.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.5.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.6.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.7.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.8.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.attn.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.attn.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.attn.q_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.attn.qkv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.attn.v_bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.mlp.fc1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.mlp.fc1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.mlp.fc2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.mlp.fc2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.norm1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.norm1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.norm2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.blocks.9.norm2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.cls_token": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.patch_embed.proj.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.patch_embed.proj.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_blip2.visual_encoder.pos_embed": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.0._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.1._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.10._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.11._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.12._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.13._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.14._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.15._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.16._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.17._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.18._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.19._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.2._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.20._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.21._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.22._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.23._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.24._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.25._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.26._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.27._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.28._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.29._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.3._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.30._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.31._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.32._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.33._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.34._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.35._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.36._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.37._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.38._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.39._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.4._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.40._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.41._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.42._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.43._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.44._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.45._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.46._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.47._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.48._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.49._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.5._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.50._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.51._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.52._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.53._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.54._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.6._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.7._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.8._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn2.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn2.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn2.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn2.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._bn2.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._depthwise_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._expand_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._project_conv.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._se_expand.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._se_expand.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._se_reduce.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._blocks.9._se_reduce.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn0.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn0.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn0.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn0.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn0.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn1.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn1.num_batches_tracked": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn1.running_mean": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn1.running_var": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._bn1.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._conv_head.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._conv_stem.weight": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._fc.bias": "model-00001-of-00002.safetensors", "beit3.vision_embed.model_efficientnet._fc.weight": "model-00001-of-00002.safetensors", "head.0.bias": "model-00002-of-00002.safetensors", "head.0.weight": "model-00002-of-00002.safetensors", "head.1.bias": "model-00002-of-00002.safetensors", "head.1.weight": "model-00002-of-00002.safetensors", "head.3.bias": "model-00002-of-00002.safetensors", "head.3.weight": "model-00002-of-00002.safetensors", "pooler.dense.bias": "model-00002-of-00002.safetensors", "pooler.dense.weight": "model-00002-of-00002.safetensors", "pooler.norm.bias": "model-00002-of-00002.safetensors", "pooler.norm.weight": "model-00002-of-00002.safetensors" } }