VISOR-GPT / train /models /xlm-roberta /base_config.json
szukevin's picture
upload
7900c16
raw
history blame
319 Bytes
{
"emb_size": 768,
"feedforward_size": 3072,
"hidden_size": 768,
"hidden_act": "gelu",
"heads_num": 12,
"layers_num": 12,
"max_seq_length": 514,
"dropout": 0.1,
"data_processor": "mlm",
"embedding": ["word", "pos", "seg"],
"encoder": "transformer",
"mask": "fully_visible",
"target": ["mlm"]
}