pretrain / config.json
Weijie Xu
Training in progress, step 500
09781a4
raw
history blame contribute delete
705 Bytes
{
"architectures": [
"Net"
],
"d_ff": 2048,
"d_kv": 64,
"d_model": 512,
"dense_act_fn": "relu",
"dropout": 0.2,
"dropout_rate": 0.1,
"eos_token_id": 1,
"feed_forward_proj": "relu",
"hidden_size1": 256,
"hidden_size2": 512,
"initializer_factor": 1.0,
"is_encoder_decoder": true,
"is_gated_act": false,
"layer_norm_epsilon": 1e-06,
"model_type": "t5",
"num_decoder_layers": 6,
"num_heads": 8,
"num_layer": 1,
"num_layers": 6,
"output_size": 1024,
"pad_token_id": 0,
"relative_attention_max_distance": 128,
"relative_attention_num_buckets": 32,
"torch_dtype": "float32",
"transformers_version": "4.26.0",
"use_cache": true,
"vocab_size": 32128
}