diff --git a/cosmos3_nano_stage_one_lora_r256_replicate_only_500k_20261103/attempt01/step_004000/checkpoint_validation.json b/cosmos3_nano_stage_one_lora_r256_replicate_only_500k_20261103/attempt01/step_004000/checkpoint_validation.json new file mode 100644 index 0000000000000000000000000000000000000000..bd7a041b4d0e49a0ae4d0f139dd1a8bcb0e63322 --- /dev/null +++ b/cosmos3_nano_stage_one_lora_r256_replicate_only_500k_20261103/attempt01/step_004000/checkpoint_validation.json @@ -0,0 +1,3723 @@ +{ + "all_finite": true, + "bytes": 2990807184, + "dtype": "float32", + "metadata": { + "base_identity": "nvidia/Cosmos3-Nano@e59a53c25979a090fa8706c9acc0c254a6e89b92", + "config_identity": "cosmos3_nano_omni4d_stage_one_lora_r256_native_layer_masks_fp32_rope_full_checkpoint_retained_forward_parameters_global32", + "exact_training_resume": "false", + "format": "omni4d_trainable_weights_v1", + "parameter_names": "[\"language_model.model.layers.0.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.0.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.0.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.0.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.0.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.0.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.0.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.0.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.0.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.0.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.0.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.0.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.0.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.0.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.1.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.1.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.1.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.1.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.1.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.1.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.1.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.1.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.1.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.1.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.1.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.1.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.1.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.1.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.2.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.2.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.2.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.2.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.2.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.2.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.2.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.2.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.2.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.2.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.2.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.2.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.2.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.2.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.3.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.3.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.3.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.3.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.3.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.3.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.3.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.3.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.3.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.3.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.3.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.3.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.3.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.3.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.4.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.4.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.4.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.4.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.4.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.4.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.4.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.4.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.4.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.4.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.4.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.4.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.4.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.4.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.5.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.5.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.5.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.5.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.5.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.5.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.5.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.5.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.5.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.5.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.5.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.5.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.5.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.5.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.6.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.6.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.6.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.6.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.6.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.6.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.6.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.6.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.6.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.6.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.6.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.6.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.6.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.6.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.7.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.7.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.7.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.7.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.7.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.7.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.7.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.7.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.7.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.7.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.7.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.7.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.7.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.7.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.8.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.8.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.8.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.8.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.8.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.8.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.8.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.8.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.8.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.8.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.8.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.8.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.8.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.8.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.9.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.9.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.9.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.9.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.9.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.9.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.9.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.9.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.9.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.9.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.9.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.9.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.9.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.9.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.10.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.10.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.10.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.10.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.10.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.10.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.10.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.10.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.10.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.10.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.10.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.10.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.10.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.10.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.11.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.11.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.11.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.11.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.11.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.11.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.11.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.11.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.11.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.11.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.11.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.11.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.11.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.11.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.12.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.12.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.12.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.12.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.12.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.12.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.12.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.12.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.12.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.12.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.12.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.12.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.12.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.12.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.13.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.13.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.13.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.13.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.13.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.13.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.13.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.13.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.13.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.13.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.13.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.13.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.13.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.13.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.14.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.14.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.14.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.14.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.14.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.14.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.14.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.14.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.14.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.14.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.14.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.14.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.14.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.14.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.15.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.15.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.15.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.15.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.15.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.15.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.15.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.15.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.15.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.15.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.15.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.15.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.15.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.15.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.16.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.16.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.16.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.16.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.16.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.16.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.16.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.16.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.16.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.16.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.16.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.16.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.16.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.16.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.17.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.17.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.17.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.17.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.17.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.17.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.17.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.17.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.17.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.17.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.17.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.17.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.17.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.17.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.18.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.18.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.18.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.18.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.18.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.18.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.18.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.18.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.18.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.18.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.18.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.18.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.18.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.18.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.19.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.19.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.19.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.19.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.19.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.19.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.19.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.19.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.19.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.19.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.19.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.19.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.19.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.19.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.20.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.20.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.20.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.20.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.20.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.20.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.20.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.20.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.20.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.20.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.20.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.20.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.20.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.20.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.21.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.21.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.21.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.21.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.21.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.21.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.21.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.21.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.21.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.21.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.21.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.21.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.21.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.21.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.22.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.22.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.22.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.22.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.22.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.22.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.22.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.22.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.22.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.22.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.22.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.22.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.22.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.22.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.23.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.23.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.23.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.23.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.23.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.23.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.23.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.23.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.23.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.23.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.23.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.23.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.23.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.23.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.24.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.24.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.24.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.24.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.24.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.24.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.24.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.24.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.24.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.24.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.24.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.24.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.24.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.24.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.25.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.25.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.25.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.25.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.25.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.25.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.25.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.25.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.25.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.25.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.25.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.25.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.25.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.25.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.26.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.26.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.26.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.26.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.26.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.26.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.26.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.26.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.26.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.26.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.26.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.26.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.26.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.26.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.27.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.27.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.27.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.27.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.27.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.27.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.27.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.27.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.27.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.27.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.27.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.27.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.27.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.27.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.28.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.28.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.28.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.28.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.28.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.28.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.28.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.28.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.28.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.28.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.28.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.28.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.28.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.28.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.29.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.29.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.29.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.29.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.29.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.29.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.29.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.29.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.29.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.29.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.29.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.29.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.29.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.29.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.30.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.30.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.30.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.30.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.30.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.30.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.30.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.30.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.30.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.30.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.30.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.30.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.30.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.30.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.31.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.31.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.31.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.31.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.31.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.31.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.31.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.31.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.31.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.31.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.31.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.31.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.31.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.31.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.32.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.32.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.32.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.32.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.32.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.32.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.32.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.32.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.32.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.32.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.32.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.32.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.32.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.32.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.33.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.33.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.33.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.33.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.33.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.33.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.33.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.33.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.33.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.33.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.33.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.33.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.33.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.33.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.34.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.34.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.34.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.34.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.34.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.34.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.34.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.34.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.34.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.34.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.34.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.34.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.34.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.34.mlp_moe_gen.down_proj.lora_B.weight\", \"language_model.model.layers.35.self_attn.q_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.35.self_attn.q_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.35.self_attn.k_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.35.self_attn.k_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.35.self_attn.v_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.35.self_attn.v_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.35.self_attn.o_proj_moe_gen.lora_A.weight\", \"language_model.model.layers.35.self_attn.o_proj_moe_gen.lora_B.weight\", \"language_model.model.layers.35.mlp_moe_gen.gate_proj.lora_A.weight\", \"language_model.model.layers.35.mlp_moe_gen.gate_proj.lora_B.weight\", \"language_model.model.layers.35.mlp_moe_gen.up_proj.lora_A.weight\", \"language_model.model.layers.35.mlp_moe_gen.up_proj.lora_B.weight\", \"language_model.model.layers.35.mlp_moe_gen.down_proj.lora_A.weight\", \"language_model.model.layers.35.mlp_moe_gen.down_proj.lora_B.weight\", \"time_embedder.mlp.0.weight\", \"time_embedder.mlp.0.bias\", \"time_embedder.mlp.2.weight\", \"time_embedder.mlp.2.bias\", \"omni4d_stream_projector.input.video.weight\", \"omni4d_stream_projector.input.video.bias\", \"omni4d_stream_projector.input.pointmap.weight\", \"omni4d_stream_projector.input.pointmap.bias\", \"omni4d_stream_projector.input.camera.weight\", \"omni4d_stream_projector.input.camera.bias\", \"omni4d_stream_projector.input.dino.weight\", \"omni4d_stream_projector.input.dino.bias\", \"omni4d_stream_projector.input.track_video.weight\", \"omni4d_stream_projector.input.track_video.bias\", \"omni4d_stream_projector.dino_feature_norm.weight\", \"omni4d_stream_projector.dino_feature_norm.bias\", \"omni4d_stream_projector.output.video.weight\", \"omni4d_stream_projector.output.video.bias\", \"omni4d_stream_projector.output.pointmap.weight\", \"omni4d_stream_projector.output.pointmap.bias\", \"omni4d_stream_projector.output.camera.weight\", \"omni4d_stream_projector.output.camera.bias\", \"omni4d_stream_projector.output.dino.weight\", \"omni4d_stream_projector.output.dino.bias\", \"omni4d_stream_projector.output.track_video.weight\", \"omni4d_stream_projector.output.track_video.bias\"]", + "update": "4000" + }, + "parameter_count": 747673344, + "sha256": "ca1b11df3fd31480aee3564c3a57782ea29e7e083f6e21922d1a27de532e1497", + "source": "/scratch/shared/engs2907/logs/vgg_forecast_4d/train/cosmos3_nano_five_stream_stage_one_lora_r256_replicate_only_500k_20261103/attempt01/trainable_overlays/step_004000.safetensors", + "source_stat": { + "device": 57, + "inode": 18133911060991891096, + "mtime_ns": 1793764496514355758, + "size": 2990807184 + }, + "status": "passed", + "tensor_count": 530, + "tensor_layout": { + "language_model.model.layers.0.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.0.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.0.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.0.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.0.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.0.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.0.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.0.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.0.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.0.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.0.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.0.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.0.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.0.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.1.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.1.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.1.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.1.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.1.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.1.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.1.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.1.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.1.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.1.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.1.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.1.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.1.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.1.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.10.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.10.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.10.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.10.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.10.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.10.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.10.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.10.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.10.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.10.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.10.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.10.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.10.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.10.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.11.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.11.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.11.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.11.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.11.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.11.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.11.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.11.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.11.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.11.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.11.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.11.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.11.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.11.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.12.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.12.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.12.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.12.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.12.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.12.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.12.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.12.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.12.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.12.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.12.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.12.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.12.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.12.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.13.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.13.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.13.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.13.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.13.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.13.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.13.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.13.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.13.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.13.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.13.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.13.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.13.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.13.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.14.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.14.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.14.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.14.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.14.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.14.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.14.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.14.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.14.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.14.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.14.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.14.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.14.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.14.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.15.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.15.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.15.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.15.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.15.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.15.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.15.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.15.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.15.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.15.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.15.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.15.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.15.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.15.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.16.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.16.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.16.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.16.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.16.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.16.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.16.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.16.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.16.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.16.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.16.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.16.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.16.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.16.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.17.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.17.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.17.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.17.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.17.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.17.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.17.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.17.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.17.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.17.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.17.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.17.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.17.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.17.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.18.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.18.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.18.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.18.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.18.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.18.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.18.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.18.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.18.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.18.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.18.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.18.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.18.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.18.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.19.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.19.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.19.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.19.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.19.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.19.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.19.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.19.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.19.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.19.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.19.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.19.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.19.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.19.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.2.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.2.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.2.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.2.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.2.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.2.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.2.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.2.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.2.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.2.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.2.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.2.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.2.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.2.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.20.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.20.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.20.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.20.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.20.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.20.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.20.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.20.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.20.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.20.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.20.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.20.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.20.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.20.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.21.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.21.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.21.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.21.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.21.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.21.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.21.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.21.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.21.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.21.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.21.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.21.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.21.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.21.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.22.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.22.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.22.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.22.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.22.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.22.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.22.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.22.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.22.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.22.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.22.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.22.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.22.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.22.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.23.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.23.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.23.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.23.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.23.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.23.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.23.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.23.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.23.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.23.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.23.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.23.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.23.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.23.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.24.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.24.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.24.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.24.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.24.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.24.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.24.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.24.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.24.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.24.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.24.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.24.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.24.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.24.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.25.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.25.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.25.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.25.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.25.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.25.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.25.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.25.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.25.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.25.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.25.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.25.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.25.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.25.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.26.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.26.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.26.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.26.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.26.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.26.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.26.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.26.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.26.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.26.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.26.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.26.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.26.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.26.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.27.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.27.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.27.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.27.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.27.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.27.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.27.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.27.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.27.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.27.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.27.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.27.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.27.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.27.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.28.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.28.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.28.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.28.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.28.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.28.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.28.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.28.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.28.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.28.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.28.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.28.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.28.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.28.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.29.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.29.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.29.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.29.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.29.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.29.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.29.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.29.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.29.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.29.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.29.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.29.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.29.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.29.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.3.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.3.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.3.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.3.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.3.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.3.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.3.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.3.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.3.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.3.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.3.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.3.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.3.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.3.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.30.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.30.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.30.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.30.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.30.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.30.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.30.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.30.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.30.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.30.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.30.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.30.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.30.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.30.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.31.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.31.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.31.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.31.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.31.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.31.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.31.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.31.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.31.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.31.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.31.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.31.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.31.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.31.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.32.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.32.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.32.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.32.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.32.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.32.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.32.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.32.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.32.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.32.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.32.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.32.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.32.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.32.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.33.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.33.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.33.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.33.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.33.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.33.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.33.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.33.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.33.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.33.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.33.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.33.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.33.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.33.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.34.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.34.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.34.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.34.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.34.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.34.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.34.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.34.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.34.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.34.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.34.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.34.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.34.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.34.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.35.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.35.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.35.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.35.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.35.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.35.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.35.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.35.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.35.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.35.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.35.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.35.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.35.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.35.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.4.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.4.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.4.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.4.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.4.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.4.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.4.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.4.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.4.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.4.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.4.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.4.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.4.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.4.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.5.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.5.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.5.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.5.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.5.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.5.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.5.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.5.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.5.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.5.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.5.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.5.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.5.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.5.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.6.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.6.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.6.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.6.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.6.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.6.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.6.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.6.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.6.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.6.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.6.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.6.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.6.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.6.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.7.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.7.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.7.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.7.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.7.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.7.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.7.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.7.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.7.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.7.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.7.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.7.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.7.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.7.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.8.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.8.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.8.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.8.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.8.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.8.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.8.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.8.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.8.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.8.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.8.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.8.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.8.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.8.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.9.mlp_moe_gen.down_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 12288 + ] + }, + "language_model.model.layers.9.mlp_moe_gen.down_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.9.mlp_moe_gen.gate_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.9.mlp_moe_gen.gate_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.9.mlp_moe_gen.up_proj.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.9.mlp_moe_gen.up_proj.lora_B.weight": { + "dtype": "F32", + "shape": [ + 12288, + 256 + ] + }, + "language_model.model.layers.9.self_attn.k_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.9.self_attn.k_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "language_model.model.layers.9.self_attn.o_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.9.self_attn.o_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.9.self_attn.q_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.9.self_attn.q_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "language_model.model.layers.9.self_attn.v_proj_moe_gen.lora_A.weight": { + "dtype": "F32", + "shape": [ + 256, + 4096 + ] + }, + "language_model.model.layers.9.self_attn.v_proj_moe_gen.lora_B.weight": { + "dtype": "F32", + "shape": [ + 1024, + 256 + ] + }, + "omni4d_stream_projector.dino_feature_norm.bias": { + "dtype": "F32", + "shape": [ + 3072 + ] + }, + "omni4d_stream_projector.dino_feature_norm.weight": { + "dtype": "F32", + "shape": [ + 3072 + ] + }, + "omni4d_stream_projector.input.camera.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "omni4d_stream_projector.input.camera.weight": { + "dtype": "F32", + "shape": [ + 4096, + 192 + ] + }, + "omni4d_stream_projector.input.dino.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "omni4d_stream_projector.input.dino.weight": { + "dtype": "F32", + "shape": [ + 4096, + 3072 + ] + }, + "omni4d_stream_projector.input.pointmap.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "omni4d_stream_projector.input.pointmap.weight": { + "dtype": "F32", + "shape": [ + 4096, + 192 + ] + }, + "omni4d_stream_projector.input.track_video.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "omni4d_stream_projector.input.track_video.weight": { + "dtype": "F32", + "shape": [ + 4096, + 192 + ] + }, + "omni4d_stream_projector.input.video.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "omni4d_stream_projector.input.video.weight": { + "dtype": "F32", + "shape": [ + 4096, + 192 + ] + }, + "omni4d_stream_projector.output.camera.bias": { + "dtype": "F32", + "shape": [ + 192 + ] + }, + "omni4d_stream_projector.output.camera.weight": { + "dtype": "F32", + "shape": [ + 192, + 4096 + ] + }, + "omni4d_stream_projector.output.dino.bias": { + "dtype": "F32", + "shape": [ + 3072 + ] + }, + "omni4d_stream_projector.output.dino.weight": { + "dtype": "F32", + "shape": [ + 3072, + 4096 + ] + }, + "omni4d_stream_projector.output.pointmap.bias": { + "dtype": "F32", + "shape": [ + 192 + ] + }, + "omni4d_stream_projector.output.pointmap.weight": { + "dtype": "F32", + "shape": [ + 192, + 4096 + ] + }, + "omni4d_stream_projector.output.track_video.bias": { + "dtype": "F32", + "shape": [ + 192 + ] + }, + "omni4d_stream_projector.output.track_video.weight": { + "dtype": "F32", + "shape": [ + 192, + 4096 + ] + }, + "omni4d_stream_projector.output.video.bias": { + "dtype": "F32", + "shape": [ + 192 + ] + }, + "omni4d_stream_projector.output.video.weight": { + "dtype": "F32", + "shape": [ + 192, + 4096 + ] + }, + "time_embedder.mlp.0.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "time_embedder.mlp.0.weight": { + "dtype": "F32", + "shape": [ + 4096, + 256 + ] + }, + "time_embedder.mlp.2.bias": { + "dtype": "F32", + "shape": [ + 4096 + ] + }, + "time_embedder.mlp.2.weight": { + "dtype": "F32", + "shape": [ + 4096, + 4096 + ] + } + }, + "validated_at_utc": "2026-11-04T04:10:29.216377+00:00" +}