| { |
| "architectures": [ |
| "RoseX1ForCausalLM" |
| ], |
| "model_type": "rose_x1", |
| "auto_map": { |
| "AutoConfig": "configuration_rose_x1.RoseX1Config", |
| "AutoModelForCausalLM": "modeling_rose_x1.RoseX1ForCausalLM" |
| }, |
| "vocab_size": 32768, |
| "hidden_size": 640, |
| "intermediate_size": 1920, |
| "num_hidden_layers": 24, |
| "num_attention_heads": 10, |
| "num_key_value_heads": 10, |
| "head_dim": 64, |
| "max_position_embeddings": 2048, |
| "hidden_act": "silu", |
| "rms_norm_eps": 1e-05, |
| "attention_bias": false, |
| "mlp_bias": false, |
| "attention_dropout": 0.0, |
| "tie_word_embeddings": true, |
| "rope_theta": 100000.0, |
| "rope_scaling": null, |
| "use_qk_norm": true, |
| "refresh_gate_enabled": true, |
| "refresh_gate_inject_layers": [ |
| 9, |
| 18 |
| ], |
| "refresh_gate_kernel_size": 9, |
| "bos_token_id": 2, |
| "eos_token_id": 3, |
| "pad_token_id": 0, |
| "torch_dtype": "float32", |
| "transformers_version": "4.40.0", |
| "use_cache": true, |
| "pretraining_tp": 1, |
| "initializer_range": 0.02 |
| } |
