SLM_vs_LLM_experiments
/
max_seq_length_128_experiments
/LoRA
/Qwen
/Qwen1.5_7B_LoRA_MAdAiLab
/twitter_disaster
/all_results.json
{ | |
"epoch": 3.0, | |
"eval_accuracy": 0.7766544117647058, | |
"eval_f1_macro": 0.7450627015924902, | |
"eval_f1_micro": 0.7766544117647058, | |
"eval_loss": 0.4901912808418274, | |
"eval_runtime": 13.0776, | |
"eval_samples": 1088, | |
"eval_samples_per_second": 83.196, | |
"eval_steps_per_second": 2.6, | |
"test_accuracy": 0.7766544117647058, | |
"test_f1_macro": 0.7415457166235069, | |
"test_f1_micro": 0.7766544117647058, | |
"test_loss": 0.4903779923915863, | |
"test_runtime": 13.1647, | |
"test_samples_per_second": 82.645, | |
"test_steps_per_second": 2.583, | |
"train_loss": 0.4493609409706265, | |
"train_runtime": 1238.3624, | |
"train_samples": 8700, | |
"train_samples_per_second": 21.076, | |
"train_steps_per_second": 0.659 | |
} |