llm-mdeberta-v3-swag / train_results.json
Paulo Vitor
fine tune swag
406e7e9
raw
history blame contribute delete
No virus
196 Bytes
{
"epoch": 3.0,
"train_loss": 0.6582008492790169,
"train_runtime": 15084.1881,
"train_samples": 73546,
"train_samples_per_second": 14.627,
"train_steps_per_second": 1.829
}