betonews-tweetcontext / train_results.json
finiteautomata's picture
New model
874e833
raw history blame
No virus
240 Bytes
{
"epoch": 2.24,
"total_flos": 3.370512610512077e+17,
"train_loss": 0.09681204452514648,
"train_runtime": 5680.8564,
"train_samples": 8369648,
"train_samples_per_second": 3605.09,
"train_steps_per_second": 1.76
}