xlmr-large-all-B / test_eval_ar.txt
HHansi's picture
Upload folder using huggingface_hub
9fae6d2 verified
Default classification report:
precision recall f1-score support
F 0.8753 0.6740 0.7616 500
T 0.7350 0.9040 0.8108 500
accuracy 0.7890 1000
macro avg 0.8051 0.7890 0.7862 1000
weighted avg 0.8051 0.7890 0.7862 1000
ADJ
Accuracy = 0.7142857142857143
Weighted Recall = 0.7142857142857143
Weighted Precision = 0.7466928286600417
Weighted F1 = 0.7104522312069482
Macro Recall = 0.7257861635220126
Macro Precision = 0.7385910500664599
Macro F1 = 0.7123689727463313
ADV
Accuracy = 0.8
Weighted Recall = 0.8
Weighted Precision = 0.8
Weighted F1 = 0.8
Macro Recall = 0.6875
Macro Precision = 0.6875
Macro F1 = 0.6875
NOUN
Accuracy = 0.7975708502024291
Weighted Recall = 0.7975708502024291
Weighted Precision = 0.8157128174844275
Weighted F1 = 0.7942263098436241
Macro Recall = 0.7960655737704918
Macro Precision = 0.8166199158485273
Macro F1 = 0.7938918558077436
VERB
Accuracy = 0.7964824120603015
Weighted Recall = 0.7964824120603015
Weighted Precision = 0.8086927102556094
Weighted F1 = 0.7947149907919534
Macro Recall = 0.7974467762709296
Macro Precision = 0.8080665411173886
Macro F1 = 0.7948962647681943