{ "model_type": "glm4_moe_lite", "src": { "model_type": "glm4_moe_lite", "vocab_size": 154880, "hidden_size": 2048, "intermediate_size": 10240, "moe_intermediate_size": 1536, "num_hidden_layers": 47, "num_nextn_predict_layers": 1, "num_attention_heads": 20, "num_key_value_heads": 20, "q_lora_rank": 768, "kv_lora_rank": 512, "qk_nope_head_dim": 192, "qk_rope_head_dim": 64, "v_head_dim": 256, "n_routed_experts": 64, "n_shared_experts": 1, "num_experts_per_tok": 4, "first_k_dense_replace": 1, "moe_layer_freq": 1, "rms_norm_eps": 1e-05, "tie_word_embeddings": false }, "tgt": { "hidden_size": 2048, "intermediate_size": 10240, "moe_intermediate_size": 1536, "num_hidden_layers": 75, "rms_norm_eps": 1e-05 }, "operators": { "hidden_operator": "vzero", "mlp_operator": "dormant_prime", "new_layer_init": "copy_base", "depth_identity_operator": "inproj_zero_prime", "seed": 42 } }