pythia-70m-wikipedia-paragraphs / train_results.json
agentlans's picture
Upload 10 files
9fd0cb8 verified
raw
history blame contribute delete
316 Bytes
{
"epoch": 50.0,
"num_input_tokens_seen": 290048000,
"total_flos": 7.7740363481088e+16,
"train_loss": 3.9178111260633375,
"train_runtime": 4788.9263,
"train_samples": 5665,
"train_samples_per_second": 59.147,
"train_steps_per_second": 7.402,
"train_tokens_per_second": 60566.395
}