pythia-2_8b_sft-gpt-turbo / train_results.json
DatPySci's picture
Model save
881d74f verified
{
"epoch": 1.0,
"total_flos": 4.7710598211200614e+17,
"train_loss": 1.2026635146722562,
"train_runtime": 2875.3707,
"train_samples": 41954,
"train_samples_per_second": 14.591,
"train_steps_per_second": 0.456
}