chessgpt-medium-m / all_results.json
dakwi's picture
End of training
2d77a45 verified
raw
history blame contribute delete
444 Bytes
{
"epoch": 1.0,
"eval_loss": 0.9776549935340881,
"eval_runtime": 458.7571,
"eval_samples": 25000,
"eval_samples_per_second": 54.495,
"eval_steps_per_second": 6.812,
"perplexity": 2.658215395086949,
"total_flos": 1.2942285813252096e+17,
"train_loss": 1.2280796801757812,
"train_runtime": 22032.6461,
"train_samples": 200000,
"train_samples_per_second": 9.077,
"train_steps_per_second": 0.142
}