step_benchmark_bert / is_test_false_eval_results.txt
system's picture
system HF staff
Update is_test_false_eval_results.txt
2479ab2
raw
history blame contribute delete
178 Bytes
model =bert-base-uncased
total batch size=32
train num epochs=1
fp16 =False
max seq length =80
eval_acc = 0.8737777777777778
eval_loss = 0.3378378099742088