text2vec-base-chinese-law / training_progress_scores.csv
twang2218's picture
Retrain the model using positive cases and shuffled positive cases.
11f2b4a verified
raw
history blame contribute delete
305 Bytes
global_step,train_loss,eval_spearman,eval_pearson
2473,0.0003225283289793879,0.8228411643737183,0.8699263536976878
4946,0.001292585046030581,0.8332828581081788,0.8984922255029837
7419,0.002469111466780305,0.8340860653185219,0.903639799369768
9892,0.00259420252405107,0.8356478571164112,0.9086306182661994