llm-deberta-v3-swag / train_results.json
Paulo Vitor
fine tune swag
9adc396
raw
history blame contribute delete
195 Bytes
{
"epoch": 3.0,
"train_loss": 0.4501488867961559,
"train_runtime": 3859.2073,
"train_samples": 73546,
"train_samples_per_second": 57.172,
"train_steps_per_second": 7.147
}