| { | |
| "module": "keras_hub.src.models.qwen3_moe.qwen3_moe_backbone", | |
| "class_name": "Qwen3MoeBackbone", | |
| "config": { | |
| "name": "qwen3_moe_backbone", | |
| "trainable": true, | |
| "dtype": { | |
| "module": "keras", | |
| "class_name": "DTypePolicy", | |
| "config": { | |
| "name": "float32" | |
| }, | |
| "registered_name": null | |
| }, | |
| "vocabulary_size": 151936, | |
| "num_layers": 48, | |
| "num_query_heads": 32, | |
| "head_dim": 128, | |
| "hidden_dim": 2048, | |
| "intermediate_dim": 6144, | |
| "moe_intermediate_dim": 768, | |
| "rope_max_wavelength": 10000000, | |
| "num_key_value_heads": 4, | |
| "rope_scaling_factor": 1.0, | |
| "layer_norm_epsilon": 1e-06, | |
| "dropout": 0, | |
| "tie_word_embeddings": false, | |
| "sliding_window_size": null, | |
| "num_experts": 128, | |
| "top_k": 8, | |
| "norm_top_k_prob": true, | |
| "decoder_sparse_step": 1, | |
| "mlp_only_layers": [], | |
| "router_aux_loss_coefficient": 0.001 | |
| }, | |
| "registered_name": "keras_hub>Qwen3MoeBackbone" | |
| } |