dev_results_model8_new / config.json
Anish13's picture
Training in progress, step 5000
def169d verified
raw
history blame
343 Bytes
{
"architectures": [
"TransformerNet"
],
"d_model": 512,
"dim_feedforward": 1024,
"dropout": 0.1,
"input_dim": 30000,
"max_seq_len": 512,
"model_type": "transformer",
"nhead": 8,
"num_decoder_layers": 6,
"num_encoder_layers": 6,
"output_dim": 30000,
"torch_dtype": "float32",
"transformers_version": "4.40.2"
}