MyDrive

This model was trained from scratch on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 1.1679
• Wer: 0.5567

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: 0.0001
• train_batch_size: 10
• eval_batch_size: 10
• seed: 42
• gradient_accumulation_steps: 3
• total_train_batch_size: 30
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1174	0.4228	200	1.2291	0.6153
0.098	0.8457	400	1.2325	0.6275
0.1301	1.2685	600	1.1969	0.6128
0.1514	1.6913	800	1.2293	0.6489
0.1494	2.1142	1000	1.3062	0.6701
0.1382	2.5370	1200	1.2223	0.6261
0.1382	2.9598	1400	1.3116	0.6506
0.1239	3.3827	1600	1.1977	0.6189
0.1228	3.8055	1800	1.1852	0.6281
0.1117	4.2283	2000	1.3370	0.6495
0.1118	4.6512	2200	1.3265	0.6432
0.1101	5.0740	2400	1.3458	0.6310
0.1328	5.4968	2600	1.2545	0.6342
0.1384	5.9197	2800	1.2806	0.6265
0.1334	6.3425	3000	1.2484	0.6369
0.1383	6.7653	3200	1.2701	0.6479
0.1281	7.1882	3400	1.1926	0.6314
0.1232	7.6110	3600	1.2255	0.6187
0.0727	8.0338	3800	1.2398	0.6014
0.0749	8.4567	4000	1.2319	0.5957
0.0734	8.8795	4200	1.2247	0.5879
0.0684	9.3023	4400	1.3474	0.6136
0.073	9.7252	4600	1.2837	0.5936
0.0728	10.1480	4800	1.2477	0.5910
0.0718	10.5708	5000	1.2472	0.5867
0.0685	10.9937	5200	1.2693	0.5789
0.0649	11.4165	5400	1.2165	0.5787
0.0632	11.8393	5600	1.2447	0.5842
0.0625	12.2622	5800	1.3088	0.5806
0.061	12.6850	6000	1.3399	0.5924
0.0595	13.1078	6200	1.3049	0.5769
0.0608	13.5307	6400	1.2737	0.5734
0.0596	13.9535	6600	1.2288	0.5747
0.0565	14.3763	6800	1.2599	0.5677
0.0568	14.7992	7000	1.2705	0.5622
0.0538	15.2220	7200	1.3540	0.5838
0.0585	15.6448	7400	1.3334	0.5798
0.0548	16.0677	7600	1.3313	0.5724
0.0526	16.4905	7800	1.3299	0.5720
0.0577	16.9133	8000	1.3206	0.5830
0.0513	17.3362	8200	1.3500	0.5787
0.0506	17.7590	8400	1.3185	0.5698
0.0498	18.1818	8600	1.3656	0.5800
0.0515	18.6047	8800	1.3253	0.5669
0.05	19.0275	9000	1.3411	0.5810
0.048	19.4503	9200	1.3628	0.5730
0.049	19.8732	9400	1.3700	0.5730
0.0469	20.2960	9600	1.3646	0.5718
0.0474	20.7188	9800	1.4191	0.5787
0.0488	21.1416	10000	1.3450	0.5753
0.0466	21.5645	10200	1.2961	0.5612
0.0462	21.9873	10400	1.3379	0.5732
0.0479	22.4101	10600	1.3641	0.5755
0.0475	22.8330	10800	1.3316	0.5751
0.0461	23.2558	11000	1.4021	0.5779
0.0443	23.6786	11200	1.3808	0.5767
0.0448	24.1015	11400	1.4157	0.5779
0.1948	16.3609	11600	0.8630	0.5620
0.1658	16.6429	11800	0.9330	0.5692
0.1632	16.9248	12000	0.8790	0.5518
0.1373	17.2068	12200	0.9279	0.5455
0.1233	17.4887	12400	1.0114	0.5634
0.1223	17.7707	12600	1.0203	0.5638
0.1207	18.0526	12800	1.0660	0.5724
0.1009	18.3346	13000	1.0873	0.5667
0.106	18.6165	13200	1.1188	0.5667
0.0989	18.8985	13400	1.0954	0.5689
0.0981	19.1805	13600	1.1168	0.5636
0.0858	19.4638	13800	1.1655	0.5669
0.0851	19.7458	14000	1.1516	0.5596
0.0929	20.0277	14200	1.1067	0.5545
0.0816	20.3097	14400	1.1479	0.5608
0.0853	20.5916	14600	1.1574	0.5626
0.0823	20.8736	14800	1.1786	0.5655
0.0837	21.1555	15000	1.1809	0.5618
0.0806	21.4375	15200	1.1776	0.5547
0.0819	21.7195	15400	1.1668	0.5581
0.079	22.0014	15600	1.2081	0.5573
0.0739	22.2834	15800	1.2005	0.5565
0.0751	22.5653	16000	1.1868	0.5539
0.0777	22.8473	16200	1.1831	0.5569
0.0705	23.1292	16400	1.2246	0.5579
0.0704	23.4112	16600	1.2922	0.5614
0.0684	23.6931	16800	1.2495	0.5555
0.0714	23.9751	17000	1.2268	0.5539
0.0669	24.2570	17200	1.3074	0.5647
0.067	24.5390	17400	1.2619	0.5555
0.0664	24.8210	17600	1.2757	0.5587
0.1389	21.5232	17800	1.1468	0.5704
0.1246	21.7648	18000	1.1285	0.5577
0.1292	22.0064	18200	1.1010	0.5524
0.1115	22.2481	18400	1.1428	0.5563
0.1129	22.4897	18600	1.1834	0.5647
0.1178	22.7314	18800	1.1346	0.5522
0.1119	22.9730	19000	1.1957	0.5587
0.1031	23.2147	19200	1.1525	0.5457
0.1066	23.4563	19400	1.1926	0.5583
0.103	23.6979	19600	1.2014	0.5563
0.1016	23.9396	19800	1.2301	0.5583
0.1009	24.1812	20000	1.2208	0.5530
0.0953	24.4229	20200	1.2250	0.5587
0.0981	24.6645	20400	1.2353	0.5543
0.0973	24.9062	20600	1.2359	0.5571
0.11	24.4417	20800	1.2127	0.5620
0.1119	24.6766	21000	1.1876	0.5571
0.1053	24.9115	21200	1.1679	0.5567

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1