hf-tiny-v2/tiny-random-Cohere2Model
08
1{2 "architectures": [3 "Cohere2Model"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "attention_probs_dropout_prob": 0.1,8 "bos_token_id": 5,9 "dtype": "float32",10 "eos_token_id": 255001,11 "head_dim": 8,12 "hidden_act": "gelu",13 "hidden_dropout_prob": 0.1,14 "hidden_size": 32,15 "initializer_range": 0.02,16 "intermediate_size": 37,17 "is_decoder": false,18 "layer_norm_eps": 1e-05,19 "layer_types": [20 "sliding_attention",21 "sliding_attention"22 ],23 "logit_scale": 0.0625,24 "max_position_embeddings": 512,25 "model_type": "cohere2",26 "num_attention_heads": 4,27 "num_hidden_layers": 2,28 "num_key_value_heads": 4,29 "pad_token_id": 0,30 "rope_parameters": {31 "rope_theta": 10000.0,32 "rope_type": "default"33 },34 "sliding_window": 4096,35 "tie_word_embeddings": true,36 "transformers_version": "5.16.0.dev0",37 "type_vocab_size": 16,38 "use_cache": true,39 "vocab_size": 25502940}41 