tiny-random/phi-moe

1{2  "architectures": [3    "PhiMoEForCausalLM"4  ],5  "attention_bias": true,6  "attention_dropout": 0.0,7  "auto_map": {8    "AutoConfig": "microsoft/Phi-tiny-MoE-instruct--configuration_slimmoe.PhiMoEConfig",9    "AutoModelForCausalLM": "microsoft/Phi-tiny-MoE-instruct--modeling_slimmoe.PhiMoEForCausalLM"10  },11  "bos_token_id": 1,12  "eos_token_id": 32000,13  "expert_dropout": 0.0,14  "head_dim": 32,15  "hidden_act": "silu",16  "hidden_dropout": 0.0,17  "hidden_size": 64,18  "initializer_range": 0.02,19  "input_jitter_noise": 0.01,20  "intermediate_size": 128,21  "lm_head_bias": true,22  "max_position_embeddings": 4096,23  "model_type": "phimoe",24  "num_attention_heads": 2,25  "num_experts_per_tok": 2,26  "num_hidden_layers": 2,27  "num_key_value_heads": 1,28  "num_local_experts": 8,29  "output_router_logits": false,30  "rms_norm_eps": 1e-05,31  "rope_scaling": null,32  "rope_theta": 10000.0,33  "router_aux_loss_coef": 0.0,34  "router_jitter_noise": 0.01,35  "sliding_window": 2047,36  "tie_word_embeddings": true,37  "torch_dtype": "bfloat16",38  "transformers_version": "4.51.3",39  "use_cache": true,40  "vocab_size": 3206441}