phi_moe_embed_16x16 / config.json
yifeihe3's picture
Upload MixtralForCausalLM
8e9573a verified
{
"architectures": [
"MixtralForCausalLM"
],
"attention_bias": true,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"input_jitter_noise": 0.01,
"intermediate_size": 6400,
"lm_head_bias": true,
"lora_rank": 32,
"max_position_embeddings": 4096,
"model_type": "mixtral",
"num_attention_heads": 32,
"num_experts_per_tok": 2,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"num_local_experts": 16,
"output_router_logits": false,
"rms_norm_eps": 1e-05,
"rope_theta": 10000.0,
"router_aux_loss_coef": 0.0,
"router_jitter_noise": 0.01,
"separate_lora": false,
"sliding_window": 2047,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.39.2",
"use_cache": true,
"use_lora": true,
"vocab_size": 32064
}