| { | |
| "_name_or_path": "time_moe_200m", | |
| "apply_aux_loss": true, | |
| "architectures": [ | |
| "TimeMoeForPrediction" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_time_moe.TimeMoeConfig", | |
| "AutoModelForCausalLM": "modeling_time_moe.TimeMoeForPrediction" | |
| }, | |
| "attention_dropout": 0.0, | |
| "hidden_act": "silu", | |
| "hidden_size": 768, | |
| "horizon_lengths": [ | |
| 1, | |
| 8, | |
| 32, | |
| 64 | |
| ], | |
| "initializer_range": 0.02, | |
| "input_size": 1, | |
| "intermediate_size": 3072, | |
| "max_position_embeddings": 4096, | |
| "model_type": "time_moe", | |
| "num_attention_heads": 12, | |
| "num_experts": 8, | |
| "num_experts_per_tok": 2, | |
| "num_hidden_layers": 12, | |
| "num_key_value_heads": 12, | |
| "rms_norm_eps": 1e-06, | |
| "rope_theta": 10000, | |
| "router_aux_loss_factor": 0.02, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.40.1", | |
| "use_cache": true, | |
| "use_dense": false | |
| } | |