tgpt-opt-nano / config.json
lixiangchun's picture
initial upload
56377c9 verified
raw
history blame contribute delete
738 Bytes
{
"_name_or_path": "facebook/opt-125m",
"_attn_implementation": "flash_attention_2",
"activation_dropout": 0.0,
"activation_function": "relu",
"architectures": [
"OPTForCausalLM"
],
"attention_dropout": 0.0,
"do_layer_norm_before": true,
"dropout": 0.1,
"pad_token_id": 0,
"unk_token_id": 1,
"bos_token_id": 2,
"eos_token_id": 3,
"mask_token_id": 4,
"ffn_dim": 1024,
"hidden_size": 256,
"init_std": 0.02,
"layerdrop": 0.0,
"max_position_embeddings": 512,
"model_type": "opt",
"num_attention_heads": 8,
"num_hidden_layers": 6,
"prefix": "</s>",
"torch_dtype": "float16",
"transformers_version": "4.21.0.dev0",
"use_cache": true,
"vocab_size": 21051,
"word_embed_proj_dim": 256
}