| { |
| "architectures": [ |
| "BETForCausalLM" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_bet.BETConfig", |
| "AutoModelForCausalLM": "modeling_bet.BETForCausalLM", |
| "AutoTokenizer": "tokenization_bet.BETByteTokenizer" |
| }, |
| "bos_token_id": 257, |
| "dtype": "float32", |
| "eos_token_id": 258, |
| "head_dim": 64, |
| "hidden_size": 384, |
| "intermediate_size": 960, |
| "is_encoder_decoder": false, |
| "max_position_embeddings": 512, |
| "mhc_sinkhorn_iters": 8, |
| "model_type": "bet", |
| "num_attention_heads": 12, |
| "num_hidden_layers": 6, |
| "num_key_value_heads": 1, |
| "pad_token_id": 256, |
| "refinement_cycles": 6, |
| "rms_norm_eps": 1e-06, |
| "rope_theta": 1000000.0, |
| "tie_word_embeddings": true, |
| "tokenizer_class": "BETByteTokenizer", |
| "use_cache": false, |
| "vocab_size": 259 |
| } |
|
|
