| { |
| "A_init_range": [ |
| 1, |
| 16 |
| ], |
| "architectures": [ |
| "DragonForCausalLM" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "auto_map": { |
| "AutoConfig": "configuration_dragon.DragonConfig", |
| "AutoModel": "modeling_dragon.DragonModel", |
| "AutoModelForCausalLM": "modeling_dragon.DragonForCausalLM" |
| }, |
| "bos_token_id": 0, |
| "conv_kernel": 4, |
| "dtype": "bfloat16", |
| "eos_token_id": 0, |
| "eot_token_id": 0, |
| "expand_factor": 2, |
| "hidden_dropout": 0.0, |
| "hidden_size": 2048, |
| "initializer_range": 0.006, |
| "intermediate_size": 8192, |
| "layers_config": "lrlrdlrlrlrlrdlrlrlrlrdlrlrlrlrdlrlr", |
| "max_position_embeddings": 32768, |
| "mlp_bias": false, |
| "mlp_hidden_act": "relu2", |
| "model_type": "dragon", |
| "norm_epsilon": 1e-06, |
| "num_attention_heads": 32, |
| "num_hidden_layers": 36, |
| "num_key_value_heads": 16, |
| "num_logits_to_keep": 1, |
| "old_lns": true, |
| "pad_token_id": 0, |
| "qk_norm": true, |
| "residual_in_fp32": false, |
| "rope_theta": 163, |
| "scalable_softmax": true, |
| "sliding_window_size": 1024, |
| "slw_wsize": 32768, |
| "softcap_global_attn": 150.0, |
| "softcap_local_attn": 0.0, |
| "tie_word_embeddings": false, |
| "time_step_floor": 0.0001, |
| "time_step_max": 0.1, |
| "time_step_min": 0.001, |
| "transformers_version": "4.56.1", |
| "uscaling_tau": 0.2, |
| "use_bias": false, |
| "use_cache": true, |
| "use_uscaling": false, |
| "vocab_size": 196736 |
| } |
|
|
