{ "metadata": { "total_size": 59886793728.0 }, "weight_map": { "/layers/embedding/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/mlp/down/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/mlp/gate/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/mlp/up/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/mlp_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/kv_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/kv_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/kv_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/output_proj/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/query_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/query_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer/attention/query_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/mlp/experts/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/mlp/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/mlp/shared_experts/down/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/mlp/shared_experts/gate/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/mlp/shared_experts/up/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/mlp_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/kv_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/kv_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/kv_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/output_proj/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/query_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/query_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_1/attention/query_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/mlp/experts/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/mlp/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/mlp/shared_experts/down/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/mlp/shared_experts/gate/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/mlp/shared_experts/up/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/mlp_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/kv_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/kv_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/kv_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/output_proj/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/query_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/query_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_2/attention/query_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/mlp/experts/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/mlp/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/mlp/shared_experts/down/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/mlp/shared_experts/gate/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/mlp/shared_experts/up/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/mlp_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/kv_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/kv_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/kv_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/output_proj/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/query_a_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/query_a/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_3/attention/query_b/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention_norm/vars": [ "model_00000.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/mlp/experts/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/mlp/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/mlp/shared_experts/down/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/mlp/shared_experts/gate/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/mlp/shared_experts/up/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/mlp_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/kv_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/kv_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/kv_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/output_proj/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/query_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/query_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_4/attention/query_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/mlp/experts/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/mlp/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/mlp/shared_experts/down/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/mlp/shared_experts/gate/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/mlp/shared_experts/up/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/mlp_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/kv_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/kv_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/kv_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/output_proj/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/query_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/query_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_5/attention/query_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/mlp/experts/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/mlp/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/mlp/shared_experts/down/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/mlp/shared_experts/gate/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/mlp/shared_experts/up/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/mlp_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/kv_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/kv_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/kv_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/output_proj/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/query_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/query_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_6/attention/query_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/mlp/experts/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/mlp/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/mlp/shared_experts/down/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/mlp/shared_experts/gate/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/mlp/shared_experts/up/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/mlp_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/kv_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/kv_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/kv_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/output_proj/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/query_a_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/query_a/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_7/attention/query_b/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention_norm/vars": [ "model_00001.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/mlp/experts/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/mlp/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/mlp/shared_experts/down/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/mlp/shared_experts/gate/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/mlp/shared_experts/up/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/mlp_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/kv_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/kv_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/kv_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/output_proj/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/query_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/query_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_8/attention/query_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/mlp/experts/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/mlp/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/mlp/shared_experts/down/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/mlp/shared_experts/gate/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/mlp/shared_experts/up/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/mlp_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/kv_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/kv_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/kv_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/output_proj/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/query_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/query_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_9/attention/query_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/mlp/experts/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/mlp/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/mlp/shared_experts/down/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/mlp/shared_experts/gate/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/mlp/shared_experts/up/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/mlp_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/kv_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/kv_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/kv_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/output_proj/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/query_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/query_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_10/attention/query_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/mlp/experts/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/mlp/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/mlp/shared_experts/down/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/mlp/shared_experts/gate/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/mlp/shared_experts/up/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/mlp_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/kv_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/kv_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/kv_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/output_proj/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/query_a_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/query_a/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_11/attention/query_b/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention_norm/vars": [ "model_00002.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/mlp/experts/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/mlp/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/mlp/shared_experts/down/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/mlp/shared_experts/gate/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/mlp/shared_experts/up/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/mlp_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/kv_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/kv_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/kv_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/output_proj/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/query_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/query_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_12/attention/query_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/mlp/experts/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/mlp/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/mlp/shared_experts/down/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/mlp/shared_experts/gate/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/mlp/shared_experts/up/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/mlp_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/kv_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/kv_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/kv_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/output_proj/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/query_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/query_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_13/attention/query_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/mlp/experts/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/mlp/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/mlp/shared_experts/down/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/mlp/shared_experts/gate/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/mlp/shared_experts/up/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/mlp_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/kv_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/kv_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/kv_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/output_proj/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/query_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/query_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_14/attention/query_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/mlp/experts/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/mlp/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/mlp/shared_experts/down/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/mlp/shared_experts/gate/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/mlp/shared_experts/up/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/mlp_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/kv_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/kv_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/kv_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/output_proj/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/query_a_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/query_a/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_15/attention/query_b/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention_norm/vars": [ "model_00003.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/mlp/experts/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/mlp/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/mlp/shared_experts/down/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/mlp/shared_experts/gate/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/mlp/shared_experts/up/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/mlp_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/kv_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/kv_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/kv_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/output_proj/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/query_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/query_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_16/attention/query_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/mlp/experts/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/mlp/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/mlp/shared_experts/down/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/mlp/shared_experts/gate/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/mlp/shared_experts/up/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/mlp_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/kv_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/kv_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/kv_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/output_proj/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/query_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/query_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_17/attention/query_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/mlp/experts/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/mlp/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/mlp/shared_experts/down/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/mlp/shared_experts/gate/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/mlp/shared_experts/up/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/mlp_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/kv_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/kv_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/kv_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/output_proj/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/query_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/query_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_18/attention/query_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/mlp/experts/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/mlp/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/mlp/shared_experts/down/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/mlp/shared_experts/gate/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/mlp/shared_experts/up/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/mlp_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/kv_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/kv_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/kv_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/output_proj/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/query_a_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/query_a/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_19/attention/query_b/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention_norm/vars": [ "model_00004.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/mlp/experts/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/mlp/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/mlp/shared_experts/down/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/mlp/shared_experts/gate/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/mlp/shared_experts/up/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/mlp_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/kv_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/kv_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/kv_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/output_proj/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/query_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/query_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_20/attention/query_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/mlp/experts/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/mlp/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/mlp/shared_experts/down/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/mlp/shared_experts/gate/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/mlp/shared_experts/up/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/mlp_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/kv_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/kv_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/kv_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/output_proj/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/query_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/query_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_21/attention/query_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/mlp/experts/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/mlp/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/mlp/shared_experts/down/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/mlp/shared_experts/gate/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/mlp/shared_experts/up/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/mlp_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/kv_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/kv_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/kv_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/output_proj/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/query_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/query_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_22/attention/query_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/mlp/experts/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/mlp/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/mlp/shared_experts/down/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/mlp/shared_experts/gate/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/mlp/shared_experts/up/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/mlp_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/kv_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/kv_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/kv_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/output_proj/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/query_a_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/query_a/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_23/attention/query_b/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention_norm/vars": [ "model_00005.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/mlp/experts/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/mlp/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/mlp/shared_experts/down/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/mlp/shared_experts/gate/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/mlp/shared_experts/up/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/mlp_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/kv_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/kv_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/kv_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/output_proj/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/query_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/query_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_24/attention/query_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/mlp/experts/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/mlp/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/mlp/shared_experts/down/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/mlp/shared_experts/gate/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/mlp/shared_experts/up/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/mlp_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/kv_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/kv_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/kv_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/output_proj/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/query_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/query_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_25/attention/query_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/mlp/experts/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/mlp/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/mlp/shared_experts/down/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/mlp/shared_experts/gate/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/mlp/shared_experts/up/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/mlp_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/kv_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/kv_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/kv_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/output_proj/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/query_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/query_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_26/attention/query_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/mlp/experts/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/mlp/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/mlp/shared_experts/down/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/mlp/shared_experts/gate/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/mlp/shared_experts/up/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/mlp_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/kv_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/kv_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/kv_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/output_proj/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/query_a_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/query_a/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_27/attention/query_b/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention_norm/vars": [ "model_00006.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/mlp/experts/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/mlp/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/mlp/shared_experts/down/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/mlp/shared_experts/gate/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/mlp/shared_experts/up/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/mlp_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/kv_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/kv_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/kv_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/output_proj/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/query_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/query_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_28/attention/query_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/mlp/experts/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/mlp/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/mlp/shared_experts/down/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/mlp/shared_experts/gate/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/mlp/shared_experts/up/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/mlp_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/kv_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/kv_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/kv_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/output_proj/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/query_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/query_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_29/attention/query_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/mlp/experts/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/mlp/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/mlp/shared_experts/down/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/mlp/shared_experts/gate/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/mlp/shared_experts/up/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/mlp_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/kv_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/kv_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/kv_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/output_proj/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/query_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/query_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_30/attention/query_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/mlp/experts/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/mlp/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/mlp/shared_experts/down/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/mlp/shared_experts/gate/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/mlp/shared_experts/up/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/mlp_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/kv_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/kv_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/kv_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/output_proj/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/query_a_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/query_a/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_31/attention/query_b/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention_norm/vars": [ "model_00007.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/mlp/experts/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/mlp/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/mlp/shared_experts/down/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/mlp/shared_experts/gate/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/mlp/shared_experts/up/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/mlp_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/kv_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/kv_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/kv_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/output_proj/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/query_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/query_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_32/attention/query_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/mlp/experts/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/mlp/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/mlp/shared_experts/down/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/mlp/shared_experts/gate/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/mlp/shared_experts/up/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/mlp_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/kv_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/kv_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/kv_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/output_proj/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/query_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/query_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_33/attention/query_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/mlp/experts/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/mlp/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/mlp/shared_experts/down/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/mlp/shared_experts/gate/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/mlp/shared_experts/up/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/mlp_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/kv_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/kv_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/kv_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/output_proj/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/query_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/query_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_34/attention/query_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/mlp/experts/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/mlp/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/mlp/shared_experts/down/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/mlp/shared_experts/gate/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/mlp/shared_experts/up/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/mlp_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/kv_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/kv_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/kv_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/output_proj/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/query_a_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/query_a/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_35/attention/query_b/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention_norm/vars": [ "model_00008.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/mlp/experts/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/mlp/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/mlp/shared_experts/down/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/mlp/shared_experts/gate/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/mlp/shared_experts/up/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/mlp_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/kv_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/kv_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/kv_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/output_proj/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/query_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/query_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_36/attention/query_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/mlp/experts/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/mlp/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/mlp/shared_experts/down/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/mlp/shared_experts/gate/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/mlp/shared_experts/up/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/mlp_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/kv_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/kv_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/kv_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/output_proj/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/query_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/query_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_37/attention/query_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/mlp/experts/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/mlp/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/mlp/shared_experts/down/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/mlp/shared_experts/gate/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/mlp/shared_experts/up/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/mlp_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/kv_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/kv_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/kv_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/output_proj/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/query_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/query_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_38/attention/query_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/mlp/experts/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/mlp/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/mlp/shared_experts/down/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/mlp/shared_experts/gate/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/mlp/shared_experts/up/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/mlp_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/kv_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/kv_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/kv_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/output_proj/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/query_a_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/query_a/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_39/attention/query_b/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention_norm/vars": [ "model_00009.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/mlp/experts/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/mlp/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/mlp/shared_experts/down/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/mlp/shared_experts/gate/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/mlp/shared_experts/up/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/mlp_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/kv_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/kv_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/kv_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/output_proj/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/query_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/query_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_40/attention/query_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/mlp/experts/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/mlp/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/mlp/shared_experts/down/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/mlp/shared_experts/gate/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/mlp/shared_experts/up/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/mlp_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/kv_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/kv_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/kv_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/output_proj/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/query_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/query_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_41/attention/query_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/mlp/experts/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/mlp/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/mlp/shared_experts/down/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/mlp/shared_experts/gate/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/mlp/shared_experts/up/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/mlp_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/kv_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/kv_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/kv_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/output_proj/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/query_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/query_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_42/attention/query_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/mlp/experts/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/mlp/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/mlp/shared_experts/down/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/mlp/shared_experts/gate/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/mlp/shared_experts/up/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/mlp_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/kv_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/kv_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/kv_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/output_proj/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/query_a_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/query_a/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_43/attention/query_b/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention_norm/vars": [ "model_00010.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/mlp/experts/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/mlp/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/mlp/shared_experts/down/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/mlp/shared_experts/gate/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/mlp/shared_experts/up/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/mlp_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/kv_a_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/kv_a/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/kv_b/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/output_proj/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/query_a_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/query_a/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_44/attention/query_b/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/mlp/experts/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/mlp/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/mlp/shared_experts/down/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/mlp/shared_experts/gate/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/mlp/shared_experts/up/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/mlp_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/kv_a_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/kv_a/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/kv_b/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/output_proj/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/query_a_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/query_a/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_45/attention/query_b/vars": [ "model_00011.weights.h5" ], "/layers/dense/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/mlp/experts/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/mlp/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/mlp/shared_experts/down/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/mlp/shared_experts/gate/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/mlp/shared_experts/up/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/mlp_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/kv_a_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/kv_a/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/kv_b/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/output_proj/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/query_a_norm/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/query_a/vars": [ "model_00011.weights.h5" ], "/decoder_layers/glm4_moe_lite_decoder_layer_46/attention/query_b/vars": [ "model_00011.weights.h5" ], "/final_norm/vars": [ "model_00011.weights.h5" ] } }