DeepScaleR-1.5B-L36-RollE3 / train_results.json
ZMC2019's picture
Model save
8c7403e verified
{
"total_flos": 1.2726481677031834e+17,
"train_loss": 0.3811905702956918,
"train_runtime": 3910.3734,
"train_samples": 93733,
"train_samples_per_second": 1.455,
"train_steps_per_second": 0.091
}