hf-tiny-v2/tiny-random-MixtralForCausalLM
096
1{2 "architectures": [3 "MixtralForCausalLM"4 ],5 "attention_dropout": 0.0,6 "bos_token_id": 1,7 "dtype": "float32",8 "eos_token_id": 2,9 "head_dim": 16,10 "hidden_act": "gelu",11 "hidden_size": 32,12 "initializer_range": 0.02,13 "intermediate_size": 32,14 "is_decoder": true,15 "max_position_embeddings": 512,16 "model_type": "mixtral",17 "num_attention_heads": 2,18 "num_experts_per_tok": 2,19 "num_hidden_layers": 2,20 "num_key_value_heads": 2,21 "num_local_experts": 8,22 "output_router_logits": false,23 "pad_token_id": 0,24 "rms_norm_eps": 1e-05,25 "rope_parameters": {26 "rope_theta": 1000000.0,27 "rope_type": "default"28 },29 "router_aux_loss_coef": 0.001,30 "router_jitter_noise": 0.0,31 "sliding_window": null,32 "tie_word_embeddings": false,33 "transformers_version": "5.16.0.dev0",34 "use_cache": true,35 "vocab_size": 9936}37 