PELM-JointGPT / with_eval_results.json
GItaf's picture
End of training
1961feb
raw
history blame
447 Bytes
{
"epoch": 3.0,
"eval_CLS loss": 1.5777798891067505,
"eval_CLS_accuracy": 0.6201729106628242,
"eval_CLS_f1": 0.6125731879708048,
"eval_CLS_precision": 0.6216224034048498,
"eval_CLS_recall": 0.6201729106628242,
"eval_LM loss": 3.9608688354492188,
"eval_Perplexity": "52.50",
"eval_loss": 4.355612277984619,
"eval_runtime": 106.4477,
"eval_samples_per_second": 16.299,
"eval_steps_per_second": 8.154
}