| { | |
| "architectures": [ | |
| "PlamoBiModel" | |
| ], | |
| "attention_dropout": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "modeling_plamo.PlamoConfig", | |
| "AutoModel": "modeling_plamo.PlamoBiModel" | |
| }, | |
| "bos_token_id": 1, | |
| "capacity_factor": 1.0, | |
| "eos_token_id": 1, | |
| "eval_attention_n_bit": null, | |
| "eval_mlp_n_bit": null, | |
| "eval_offload_moe": false, | |
| "expert_dropout": 0.0, | |
| "fp8_accum_dtype": "bfloat16", | |
| "group_size": 1024, | |
| "hidden_size": 2048, | |
| "hidden_size_per_head": 128, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 8192, | |
| "k_expert": null, | |
| "linear_type": "fp8", | |
| "max_length": 4096, | |
| "max_position_embeddings": 4096, | |
| "model_type": "plamo", | |
| "n_expert": null, | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 16, | |
| "num_key_value_heads": 1, | |
| "pad_token_id": 3, | |
| "rms_norm_eps": 1e-06, | |
| "shared_intermediate_size": null, | |
| "sparse_intermediate_size": null, | |
| "sparse_step": null, | |
| "tie_word_embeddings": false, | |
| "tokenizer_class": "PlamoTokenizer", | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.47.0", | |
| "use_cache": false, | |
| "vocab_size": 50112 | |
| } | |