LiteLong_128K / all_results.json
jaber
upload initial model
baac3a9
{
"epoch": 0.0005,
"num_input_tokens_seen": 8396996608,
"train_loss": 0.0013505319258073469,
"train_runtime": 212.0248,
"train_samples_per_second": 301.852,
"train_steps_per_second": 4.716
}