llama3.1-8b-gpt4o_100k_closedqa-k / train_results.json
chansung's picture
Model save
5d01e7a verified
raw
history blame
236 Bytes
{
"epoch": 1.0,
"total_flos": 7.558147382936863e+17,
"train_loss": 0.9328742581419647,
"train_runtime": 2769.5152,
"train_samples": 111440,
"train_samples_per_second": 5.905,
"train_steps_per_second": 0.092
}