| { |
| "_name_or_path": "/home/pasha/train/rugpt3xl-ctx4k/ruGPT3XL-4k-v2", |
| "architectures": [ |
| "RuGPT3XLForCausalLM" |
| ], |
| "attention_dropout": 0.1, |
| "attn_implementation": null, |
| "auto_map": { |
| "AutoConfig": "configuration_rugpt3xl.RuGPT3XLConfig", |
| "AutoModel": "modeling_rugpt3xl.RuGPT3XLModel", |
| "AutoModelForCausalLM": "modeling_rugpt3xl.RuGPT3XLForCausalLM" |
| }, |
| "bos_token_id": 2, |
| "embedding_dropout": 0.1, |
| "eos_token_id": 1, |
| "hidden_act": "gelu_new", |
| "hidden_size": 2048, |
| "initializer_range": 0.02, |
| "intermediate_size": 8192, |
| "layer_norm_eps": 1e-05, |
| "max_position_embeddings": 8192, |
| "model_type": "rugpt3xl", |
| "num_attention_heads": 16, |
| "num_hidden_layers": 24, |
| "output_dropout": 0.1, |
| "pad_token_id": 0, |
| "sparse_block_size": 16, |
| "sparse_mode": "alternating", |
| "sparse_num_different_global_patterns": 8, |
| "sparse_num_global_blocks": 1, |
| "sparse_num_local_blocks": 8, |
| "tie_word_embeddings": false, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.47.1", |
| "use_cache": true, |
| "vocab_size": 50264 |
| } |
