BAAI
/

AquilaMoE-SFT / config.json
hyxmmm's picture
Upload config.json with huggingface_hub
49b8c02 verified
raw
history blame contribute delete
908 Bytes
{
"architectures": [
"AquilaMoeForCausalLM"
],
"attention_dropout": 0.0,
"auto_map": {
"AutoConfig": "configuration_aquilamoe.AquilaMoeConfig",
"AutoModelForCausalLM": "modeling_aquilamoe.AquilaMoeForCausalLM"
},
"bos_token_id": 151849,
"eos_token_id": 151850,
"hidden_act": "silu",
"hidden_size": 5120,
"initializer_range": 0.02,
"intermediate_size": 20480,
"max_position_embeddings": 4096,
"model_type": "aquilamoe",
"num_attention_heads": 40,
"num_experts_per_tok": 2,
"num_hidden_layers": 40,
"num_key_value_heads": 8,
"num_local_experts": 8,
"output_router_logits": false,
"pad_token_id": 151643,
"rms_norm_eps": 1e-05,
"rope_theta": 1000000.0,
"router_aux_loss_coef": 0.001,
"sliding_window": null,
"tie_word_embeddings": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.40.1",
"use_cache": true,
"vocab_size": 151851
}