| { |
| "architectures": [ |
| "LingxiForCausalLM" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_lingxi.LingxiConfig", |
| "AutoModel": "modeling_lingxi.LingxiModel", |
| "AutoModelForCausalLM": "modeling_lingxi.LingxiForCausalLM" |
| }, |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 151643, |
| "eos_token_id": 151645, |
| "hidden_act": "silu", |
| "hidden_size": 896, |
| "initializer_range": 0.02, |
| "intermediate_size": 3840, |
| "max_position_embeddings": 4096, |
| "model_type": "lingxi", |
| "num_attention_heads": 14, |
| "num_hidden_layers": 24, |
| "num_key_value_heads": 14, |
| "pretraining_tp": 1, |
| "rms_norm_eps": 1e-06, |
| "rope_scaling": null, |
| "rope_theta": 500000.0, |
| "tie_word_embeddings": true, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.40.2", |
| "use_cache": true, |
| "vocab_size": 151936 |
| } |
|
|
