xlm-roberta-base-language-detection-finetuned

This model is a fine-tuned version of papluca/xlm-roberta-base-language-detection on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1662
• Accuracy: 0.9619
• F1: 0.9619

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• 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	Accuracy	F1
No log	0.14	50	0.2578	0.9137	0.9135
No log	0.28	100	0.2252	0.9294	0.9294
No log	0.42	150	0.2141	0.9350	0.9351
No log	0.56	200	0.1996	0.9394	0.9395
No log	0.69	250	0.1767	0.9451	0.9451
No log	0.83	300	0.1669	0.9476	0.9477
No log	0.97	350	0.1935	0.9479	0.9479
0.2195	1.11	400	0.1823	0.9504	0.9505
0.2195	1.25	450	0.1641	0.9498	0.9499
0.2195	1.39	500	0.1906	0.9529	0.9530
0.2195	1.53	550	0.1868	0.9481	0.9483
0.2195	1.67	600	0.1581	0.9557	0.9557
0.2195	1.81	650	0.1539	0.9518	0.9518
0.2195	1.94	700	0.1476	0.9579	0.9580
0.1469	2.08	750	0.1500	0.9557	0.9558
0.1469	2.22	800	0.1645	0.9571	0.9571
0.1469	2.36	850	0.1470	0.9579	0.9580
0.1469	2.5	900	0.1506	0.9521	0.9522
0.1469	2.64	950	0.1511	0.9574	0.9574
0.1469	2.78	1000	0.1553	0.9596	0.9596
0.1469	2.92	1050	0.1467	0.9557	0.9558
0.1247	3.06	1100	0.1676	0.9579	0.9580
0.1247	3.19	1150	0.1508	0.9535	0.9536
0.1247	3.33	1200	0.1404	0.9563	0.9564
0.1247	3.47	1250	0.1394	0.9619	0.9619
0.1247	3.61	1300	0.1439	0.9644	0.9644
0.1247	3.75	1350	0.1444	0.9591	0.9591
0.1247	3.89	1400	0.1495	0.9577	0.9578
0.1082	4.03	1450	0.1361	0.9608	0.9608
0.1082	4.17	1500	0.1531	0.9588	0.9589
0.1082	4.31	1550	0.1711	0.9507	0.9508
0.1082	4.44	1600	0.1371	0.9585	0.9586
0.1082	4.58	1650	0.1408	0.9579	0.9580
0.1082	4.72	1700	0.1444	0.9636	0.9636
0.1082	4.86	1750	0.1504	0.9613	0.9614
0.0972	5.0	1800	0.1315	0.9599	0.9600
0.0972	5.14	1850	0.1521	0.9610	0.9611
0.0972	5.28	1900	0.1531	0.9577	0.9577
0.0972	5.42	1950	0.1534	0.9610	0.9611
0.0972	5.56	2000	0.1506	0.9622	0.9622
0.0972	5.69	2050	0.1487	0.9610	0.9611
0.0972	5.83	2100	0.1541	0.9610	0.9610
0.0972	5.97	2150	0.1376	0.9571	0.9572
0.0853	6.11	2200	0.1667	0.9588	0.9589
0.0853	6.25	2250	0.1548	0.9557	0.9558
0.0853	6.39	2300	0.1527	0.9622	0.9622
0.0853	6.53	2350	0.1469	0.9619	0.9619
0.0853	6.67	2400	0.1510	0.9596	0.9597
0.0853	6.81	2450	0.1531	0.9613	0.9613
0.0853	6.94	2500	0.1605	0.9619	0.9619
0.0784	7.08	2550	0.1740	0.9571	0.9572
0.0784	7.22	2600	0.1441	0.9633	0.9633
0.0784	7.36	2650	0.1596	0.9633	0.9633
0.0784	7.5	2700	0.1469	0.9613	0.9614
0.0784	7.64	2750	0.1643	0.9596	0.9597
0.0784	7.78	2800	0.1752	0.9619	0.9619
0.0784	7.92	2850	0.1591	0.9613	0.9614
0.0712	8.06	2900	0.1604	0.9608	0.9608
0.0712	8.19	2950	0.1565	0.9596	0.9597
0.0712	8.33	3000	0.1601	0.9605	0.9605
0.0712	8.47	3050	0.1668	0.9605	0.9605
0.0712	8.61	3100	0.1765	0.9624	0.9625
0.0712	8.75	3150	0.1616	0.9613	0.9614
0.0712	8.89	3200	0.1624	0.9616	0.9616
0.062	9.03	3250	0.1598	0.9613	0.9613
0.062	9.17	3300	0.1628	0.9624	0.9625
0.062	9.31	3350	0.1627	0.9624	0.9625
0.062	9.44	3400	0.1616	0.9613	0.9613
0.062	9.58	3450	0.1669	0.9610	0.9611
0.062	9.72	3500	0.1643	0.9608	0.9608
0.062	9.86	3550	0.1650	0.9610	0.9611
0.057	10.0	3600	0.1662	0.9619	0.9619

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2