config.json
| 1 | { |
| 2 | "architectures": [ |
| 3 | "GraniteForCausalLM" |
| 4 | ], |
| 5 | "attention_bias": false, |
| 6 | "attention_dropout": 0.0, |
| 7 | "attention_multiplier": 0.0078125, |
| 8 | "bos_token_id": 100283, |
| 9 | "dtype": "bfloat16", |
| 10 | "embedding_multiplier": 1.0, |
| 11 | "eos_token_id": 100257, |
| 12 | "hidden_act": "silu", |
| 13 | "hidden_size": 4096, |
| 14 | "initializer_range": 0.1, |
| 15 | "intermediate_size": 12800, |
| 16 | "logits_scaling": 1.0, |
| 17 | "max_position_embeddings": 131072, |
| 18 | "mlp_bias": false, |
| 19 | "model_type": "granite", |
| 20 | "num_attention_heads": 32, |
| 21 | "num_hidden_layers": 40, |
| 22 | "num_key_value_heads": 8, |
| 23 | "pad_token_id": 100257, |
| 24 | "residual_multiplier": 1.0, |
| 25 | "rms_norm_eps": 1e-05, |
| 26 | "rope_parameters": { |
| 27 | "rope_theta": 10000000, |
| 28 | "rope_type": "default" |
| 29 | }, |
| 30 | "tie_word_embeddings": false, |
| 31 | "transformers_version": "5.5.3", |
| 32 | "use_cache": true, |
| 33 | "vocab_size": 100352 |
| 34 | } |
| 35 | |