SLM_vs_LLM_experiments
/
max_seq_length_128_experiments
/google_bert
/bert_base_uncased_amazon
/all_results.json
{ | |
"epoch": 3.0, | |
"eval_accuracy": 0.7878787878787878, | |
"eval_f1_macro": 0.7307727509227439, | |
"eval_f1_micro": 0.7878787878787878, | |
"eval_loss": 0.7915631532669067, | |
"eval_runtime": 1.4904, | |
"eval_samples": 1518, | |
"eval_samples_per_second": 1018.532, | |
"eval_steps_per_second": 16.103, | |
"test_accuracy": 0.8201581027667985, | |
"test_f1_macro": 0.7636155764092812, | |
"test_f1_micro": 0.8201581027667985, | |
"test_loss": 0.7230327725410461, | |
"test_runtime": 1.4903, | |
"test_samples_per_second": 1018.573, | |
"test_steps_per_second": 16.104, | |
"train_loss": 1.1961698414986595, | |
"train_runtime": 135.7278, | |
"train_samples": 12144, | |
"train_samples_per_second": 268.42, | |
"train_steps_per_second": 4.2 | |
} |