roberta-large-ner-ghtk-cs-6-label-new-data-3090-5Sep-1

This model is a fine-tuned version of FacebookAI/xlm-roberta-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1818
• Tk: {'precision': 0.8532110091743119, 'recall': 0.8017241379310345, 'f1': 0.8266666666666667, 'number': 116}
• Gày: {'precision': 0.725, 'recall': 0.8529411764705882, 'f1': 0.7837837837837837, 'number': 34}
• Gày trừu tượng: {'precision': 0.9247967479674797, 'recall': 0.9323770491803278, 'f1': 0.9285714285714285, 'number': 488}
• Ã đơn: {'precision': 0.8781725888324873, 'recall': 0.8522167487684729, 'f1': 0.8649999999999999, 'number': 203}
• Đt: {'precision': 0.9528508771929824, 'recall': 0.989749430523918, 'f1': 0.970949720670391, 'number': 878}
• Đt trừu tượng: {'precision': 0.7865612648221344, 'recall': 0.8540772532188842, 'f1': 0.8189300411522634, 'number': 233}
• Overall Precision: 0.9076
• Overall Recall: 0.9314
• Overall F1: 0.9193
• Overall Accuracy: 0.9689

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2.5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Tk	Gày	Gày trừu tượng	à đơn	Đt	Đt trừu tượng	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
No log	1.0	467	0.1589	{'precision': 0.43478260869565216, 'recall': 0.1724137931034483, 'f1': 0.24691358024691357, 'number': 116}	{'precision': 0.49122807017543857, 'recall': 0.8235294117647058, 'f1': 0.6153846153846154, 'number': 34}	{'precision': 0.7614213197969543, 'recall': 0.9221311475409836, 'f1': 0.8341056533827618, 'number': 488}	{'precision': 0.7719298245614035, 'recall': 0.8669950738916257, 'f1': 0.8167053364269142, 'number': 203}	{'precision': 0.8665997993981945, 'recall': 0.9840546697038725, 'f1': 0.9216, 'number': 878}	{'precision': 0.8, 'recall': 0.7725321888412017, 'f1': 0.7860262008733626, 'number': 233}	0.8013	0.8801	0.8389	0.9489
0.2556	2.0	934	0.1466	{'precision': 0.7111111111111111, 'recall': 0.5517241379310345, 'f1': 0.6213592233009709, 'number': 116}	{'precision': 0.7647058823529411, 'recall': 0.7647058823529411, 'f1': 0.7647058823529412, 'number': 34}	{'precision': 0.8391866913123844, 'recall': 0.930327868852459, 'f1': 0.8824101068999027, 'number': 488}	{'precision': 0.9256198347107438, 'recall': 0.5517241379310345, 'f1': 0.691358024691358, 'number': 203}	{'precision': 0.9391592920353983, 'recall': 0.9669703872437357, 'f1': 0.9528619528619529, 'number': 878}	{'precision': 0.603125, 'recall': 0.8283261802575107, 'f1': 0.6980108499095841, 'number': 233}	0.8448	0.8699	0.8571	0.9473
0.1013	3.0	1401	0.1158	{'precision': 0.8068181818181818, 'recall': 0.6120689655172413, 'f1': 0.6960784313725489, 'number': 116}	{'precision': 0.7894736842105263, 'recall': 0.8823529411764706, 'f1': 0.8333333333333333, 'number': 34}	{'precision': 0.893574297188755, 'recall': 0.9118852459016393, 'f1': 0.9026369168356998, 'number': 488}	{'precision': 0.8291457286432161, 'recall': 0.812807881773399, 'f1': 0.8208955223880597, 'number': 203}	{'precision': 0.917981072555205, 'recall': 0.9943052391799544, 'f1': 0.9546200109349371, 'number': 878}	{'precision': 0.84375, 'recall': 0.6952789699570815, 'f1': 0.7623529411764706, 'number': 233}	0.8881	0.8945	0.8913	0.9620
0.0708	4.0	1868	0.1508	{'precision': 0.7564102564102564, 'recall': 0.5086206896551724, 'f1': 0.6082474226804123, 'number': 116}	{'precision': 0.75, 'recall': 0.7941176470588235, 'f1': 0.7714285714285715, 'number': 34}	{'precision': 0.906832298136646, 'recall': 0.8975409836065574, 'f1': 0.90216271884655, 'number': 488}	{'precision': 0.8027522935779816, 'recall': 0.8620689655172413, 'f1': 0.831353919239905, 'number': 203}	{'precision': 0.9142259414225942, 'recall': 0.9954441913439636, 'f1': 0.9531079607415485, 'number': 878}	{'precision': 0.6872964169381107, 'recall': 0.9055793991416309, 'f1': 0.7814814814814814, 'number': 233}	0.8585	0.9139	0.8854	0.9455
0.0513	5.0	2335	0.1342	{'precision': 0.8518518518518519, 'recall': 0.7931034482758621, 'f1': 0.8214285714285715, 'number': 116}	{'precision': 0.7575757575757576, 'recall': 0.7352941176470589, 'f1': 0.746268656716418, 'number': 34}	{'precision': 0.9118852459016393, 'recall': 0.9118852459016393, 'f1': 0.9118852459016393, 'number': 488}	{'precision': 0.907103825136612, 'recall': 0.8177339901477833, 'f1': 0.8601036269430052, 'number': 203}	{'precision': 0.9448051948051948, 'recall': 0.9943052391799544, 'f1': 0.9689234184239734, 'number': 878}	{'precision': 0.7632508833922261, 'recall': 0.927038626609442, 'f1': 0.8372093023255813, 'number': 233}	0.9000	0.9308	0.9151	0.9680
0.0356	6.0	2802	0.1209	{'precision': 0.8416666666666667, 'recall': 0.8706896551724138, 'f1': 0.8559322033898306, 'number': 116}	{'precision': 0.7435897435897436, 'recall': 0.8529411764705882, 'f1': 0.7945205479452054, 'number': 34}	{'precision': 0.9319148936170213, 'recall': 0.8975409836065574, 'f1': 0.9144050104384135, 'number': 488}	{'precision': 0.8762886597938144, 'recall': 0.8374384236453202, 'f1': 0.8564231738035264, 'number': 203}	{'precision': 0.9623477297895903, 'recall': 0.989749430523918, 'f1': 0.9758562605277933, 'number': 878}	{'precision': 0.7848605577689243, 'recall': 0.8454935622317596, 'f1': 0.8140495867768595, 'number': 233}	0.9125	0.9242	0.9183	0.9679
0.0202	7.0	3269	0.1386	{'precision': 0.8518518518518519, 'recall': 0.7931034482758621, 'f1': 0.8214285714285715, 'number': 116}	{'precision': 0.7209302325581395, 'recall': 0.9117647058823529, 'f1': 0.8051948051948051, 'number': 34}	{'precision': 0.8912621359223301, 'recall': 0.9405737704918032, 'f1': 0.9152542372881355, 'number': 488}	{'precision': 0.8865979381443299, 'recall': 0.8472906403940886, 'f1': 0.8664987405541562, 'number': 203}	{'precision': 0.9538461538461539, 'recall': 0.9886104783599089, 'f1': 0.970917225950783, 'number': 878}	{'precision': 0.8518518518518519, 'recall': 0.7896995708154506, 'f1': 0.8195991091314031, 'number': 233}	0.9094	0.9252	0.9172	0.9698
0.0148	8.0	3736	0.1621	{'precision': 0.9021739130434783, 'recall': 0.7155172413793104, 'f1': 0.7980769230769231, 'number': 116}	{'precision': 0.7368421052631579, 'recall': 0.8235294117647058, 'f1': 0.7777777777777778, 'number': 34}	{'precision': 0.9249492900608519, 'recall': 0.9344262295081968, 'f1': 0.9296636085626911, 'number': 488}	{'precision': 0.8686868686868687, 'recall': 0.8472906403940886, 'f1': 0.85785536159601, 'number': 203}	{'precision': 0.9414316702819957, 'recall': 0.9886104783599089, 'f1': 0.9644444444444444, 'number': 878}	{'precision': 0.8130434782608695, 'recall': 0.8025751072961373, 'f1': 0.8077753779697624, 'number': 233}	0.9093	0.9191	0.9141	0.9684
0.0075	9.0	4203	0.1779	{'precision': 0.8598130841121495, 'recall': 0.7931034482758621, 'f1': 0.8251121076233183, 'number': 116}	{'precision': 0.725, 'recall': 0.8529411764705882, 'f1': 0.7837837837837837, 'number': 34}	{'precision': 0.926530612244898, 'recall': 0.930327868852459, 'f1': 0.9284253578732106, 'number': 488}	{'precision': 0.8826530612244898, 'recall': 0.8522167487684729, 'f1': 0.8671679197994987, 'number': 203}	{'precision': 0.9486899563318777, 'recall': 0.989749430523918, 'f1': 0.9687848383500557, 'number': 878}	{'precision': 0.7966804979253111, 'recall': 0.8240343347639485, 'f1': 0.8101265822784809, 'number': 233}	0.9090	0.9267	0.9178	0.9689
0.006	10.0	4670	0.1818	{'precision': 0.8532110091743119, 'recall': 0.8017241379310345, 'f1': 0.8266666666666667, 'number': 116}	{'precision': 0.725, 'recall': 0.8529411764705882, 'f1': 0.7837837837837837, 'number': 34}	{'precision': 0.9247967479674797, 'recall': 0.9323770491803278, 'f1': 0.9285714285714285, 'number': 488}	{'precision': 0.8781725888324873, 'recall': 0.8522167487684729, 'f1': 0.8649999999999999, 'number': 203}	{'precision': 0.9528508771929824, 'recall': 0.989749430523918, 'f1': 0.970949720670391, 'number': 878}	{'precision': 0.7865612648221344, 'recall': 0.8540772532188842, 'f1': 0.8189300411522634, 'number': 233}	0.9076	0.9314	0.9193	0.9689

Framework versions

• Transformers 4.44.0
• Pytorch 2.3.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1