| { | |
| "architectures": [ | |
| "DogeForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "configuration_doge.DogeConfig", | |
| "AutoModelForCausalLM": "SmallDoge/Doge-320M-Instruct--modeling_doge.DogeForCausalLM" | |
| }, | |
| "bos_token_id": 0, | |
| "dynamic_mask_ratio": 0.0, | |
| "eos_token_id": 1, | |
| "hidden_act": "silu", | |
| "hidden_bias": false, | |
| "hidden_dropout": 0.0, | |
| "hidden_size": 1024, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 2048, | |
| "is_moe": false, | |
| "keep_window_size": 2048, | |
| "max_position_embeddings": 8192, | |
| "model_type": "doge", | |
| "num_attention_heads": 8, | |
| "num_experts": 2048, | |
| "num_experts_per_tok": 8, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 4, | |
| "pad_token_id": 2, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 100000.0, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.51.3", | |
| "use_cache": false, | |
| "vocab_size": 32768 | |
| } | |