2019-05-09 03:05:02 +08:00
|
|
|
{
|
2020-04-25 00:00:23 +08:00
|
|
|
"activation_function": "gelu_new",
|
2020-02-01 07:00:25 +08:00
|
|
|
"architectures": [
|
|
|
|
|
"GPT2LMHeadModel"
|
|
|
|
|
],
|
2019-05-09 03:05:02 +08:00
|
|
|
"attn_pdrop": 0.1,
|
2020-04-25 00:00:23 +08:00
|
|
|
"bos_token_id": 50256,
|
2019-05-09 03:05:02 +08:00
|
|
|
"embd_pdrop": 0.1,
|
2020-04-25 00:00:23 +08:00
|
|
|
"eos_token_id": 50256,
|
2019-05-09 03:05:02 +08:00
|
|
|
"initializer_range": 0.02,
|
|
|
|
|
"layer_norm_epsilon": 1e-05,
|
2020-04-25 00:00:23 +08:00
|
|
|
"model_type": "gpt2",
|
2019-05-09 03:05:02 +08:00
|
|
|
"n_ctx": 1024,
|
|
|
|
|
"n_embd": 1024,
|
|
|
|
|
"n_head": 16,
|
|
|
|
|
"n_layer": 24,
|
|
|
|
|
"n_positions": 1024,
|
|
|
|
|
"n_special": 0,
|
|
|
|
|
"predict_special_tokens": true,
|
|
|
|
|
"resid_pdrop": 0.1,
|
2020-04-25 00:00:23 +08:00
|
|
|
"summary_activation": null,
|
|
|
|
|
"summary_first_dropout": 0.1,
|
|
|
|
|
"summary_proj_to_labels": true,
|
|
|
|
|
"summary_type": "cls_index",
|
|
|
|
|
"summary_use_proj": true,
|
2019-05-09 03:05:02 +08:00
|
|
|
"vocab_size": 50257
|
|
|
|
|
}
|
