| { | |
| "architectures": [ | |
| "LLM" | |
| ], | |
| "attention_bias": false, | |
| "aux_loss_coef": 0.01, | |
| "dropout": 0.0, | |
| "expert_num": 4, | |
| "flash_attn": true, | |
| "hidden_size": 512, | |
| "intermediate_size": 2048, | |
| "max_seq_len": 512, | |
| "mlp_bias": false, | |
| "model_type": "moe_model", | |
| "n_layers": 8, | |
| "num_attention_heads": 16, | |
| "num_key_value_heads": 8, | |
| "output_router_logits": true, | |
| "topk": 2, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.54.1", | |
| "vocab_size": 6400 | |
| } | |