{ "architectures": [ "DeepForCausalLM" ], "attention_dropout": 0.0, "attention_type": "gqa", "bos_token_id": 2, "clamp_mu_contextual": false, "disable_mu_guidance": false, "eos_token_id": 0, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 4096, "max_position_embeddings": 2048, "mlp_type": "swiglu", "model_type": "deep", "mu_alpha_init": 1.0, "mu_context_max": 2.0, "mu_context_min": -2.0, "mu_init_value": 0.0, "mu_max": 2.0, "mu_min": 0.0, "mup_base_width": 256, "norm_eps": 1e-06, "norm_type": "rmsnorm", "num_attention_heads": 16, "num_experts": 1, "num_hidden_layers": 18, "num_key_value_heads": 4, "pad_token_id": 1, "rms_norm_eps": 1e-06, "rope_fraction": 1.0, "rope_theta": 10000.0, "rope_type": "standard", "routed_gate_init": 1.0, "shared_expert": false, "shared_gate_init": 1.0, "shared_intermediate_size": 0, "sliding_window": null, "tie_word_embeddings": true, "top_k": 1, "top_k_primary_weight": null, "torch_dtype": "float32", "transformers_version": "checkpoint-export", "use_cache": true, "use_mu_guidance": false, "use_mu_norm": false, "use_mup_attn_scale": false, "use_mup_init": false, "use_mup_output_mult": false, "use_qk_norm": true, "use_sdpa": true, "use_shared_routed_gates": false, "use_token_routed_mlp": false, "vocab_size": 32000 }