| { |
| "architectures": [ |
| "Cohere2MoeModel" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 2, |
| "dtype": "float32", |
| "eos_token_id": 255001, |
| "expert_selection_fn": "softmax", |
| "head_dim": 16, |
| "hidden_act": "gelu", |
| "hidden_size": 32, |
| "initializer_range": 0.02, |
| "intermediate_size": 32, |
| "layer_norm_eps": 1e-05, |
| "layer_types": [ |
| "full_attention", |
| "sliding_attention" |
| ], |
| "logit_scale": 1.0, |
| "max_position_embeddings": 512, |
| "mlp_layer_types": [ |
| "dense", |
| "sparse" |
| ], |
| "model_type": "cohere2_moe", |
| "norm_topk_prob": true, |
| "num_attention_heads": 2, |
| "num_experts": 4, |
| "num_experts_per_tok": 2, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 2, |
| "num_shared_experts": 0, |
| "pad_token_id": 0, |
| "prefix_dense_intermediate_size": null, |
| "prefix_dense_sliding_window_pattern": 1, |
| "rms_norm_eps": null, |
| "rope_parameters": { |
| "rope_theta": 10000.0, |
| "rope_type": "default" |
| }, |
| "rope_scaling": null, |
| "rope_theta": 10000.0, |
| "shared_expert_combination_strategy": "average", |
| "sliding_window": 4096, |
| "sliding_window_pattern": 4, |
| "tie_word_embeddings": false, |
| "transformers_version": "5.16.0.dev0", |
| "use_cache": true, |
| "vocab_size": 255032 |
| } |
|
|