max_seq_length_128_experiments/google_t5/t5_small_scotus/eval_results.json

{
    "epoch": 3.0,
    "eval_accuracy": 0.515,
    "eval_f1_macro": 0.2879233006589316,
    "eval_f1_micro": 0.515,
    "eval_loss": 1.6132301092147827,
    "eval_runtime": 1.1495,
    "eval_samples": 1400,
    "eval_samples_per_second": 1217.973,
    "eval_steps_per_second": 38.279
}