gpt2-51M-1.31B-USPTO / train_results.json
SudharsanSundar's picture
Uploaded 1.31B tokens trained checkpoint of USPTO model
c4bdf19
raw
history blame contribute delete
194 Bytes
{
"epoch": 1.0,
"train_loss": 1.823427587890625,
"train_runtime": 29054.993,
"train_samples": 20000,
"train_samples_per_second": 44.054,
"train_steps_per_second": 0.688
}