Files
Terminal-data_processing/train_results.json

9 lines
243 B
JSON
Raw Normal View History

{
"epoch": 2.0,
"num_input_tokens_seen": 236749552,
"total_flos": 437831363461120.0,
"train_loss": 0.3453142216989031,
"train_runtime": 32003.0179,
"train_samples_per_second": 0.58,
"train_steps_per_second": 0.073
}