d1_science_longest_10k / train_results.json
ryanmarten's picture
Upload model
f70edf8 verified
raw
history blame
209 Bytes
{
"epoch": 4.992,
"total_flos": 9.228474907497595e+17,
"train_loss": 0.413890665769577,
"train_runtime": 15456.3662,
"train_samples_per_second": 3.235,
"train_steps_per_second": 0.025
}