xls-r-300m-model-F

dizzyme/xls-r-300m-model-F

xls-r-300m-model-F is a automatic speech recognition model from dizzyme. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

13% of all-time downloads

All-time downloads

Public

Repo size

8.8 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin1.3 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: apache-2.0
Model type: wav2vec2
Access: Public
Created: Jan 22, 2024
Updated: Jan 22, 2024
SHA: b35839d6

Model card

From the Hugging Face model README

xls-r-300m-model-F

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 2.4769
Wer: 1.0
Cer: 0.8567

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: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 60
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
32.1153	1.0	47	31.9339	1.0	1.9217
31.0227	2.0	94	30.0595	1.0	1.1023
26.5289	3.0	141	22.8137	1.0	1.0
16.2501	4.0	188	10.2118	1.0	1.0
10.5699	5.0	235	6.5570	1.0	1.0
7.3764	6.0	282	5.0657	1.0	1.0
5.3976	7.0	329	4.1925	1.0	1.0
4.3718	8.0	376	3.7204	1.0	1.0
3.8762	9.0	423	3.4513	1.0	1.0
3.5061	10.0	470	3.2995	1.0	1.0
3.3488	11.0	517	3.2259	1.0	1.0
3.2622	12.0	564	3.1920	1.0	1.0
3.2165	13.0	611	3.1745	1.0	1.0
3.18	14.0	658	3.1655	1.0	1.0
3.1688	15.0	705	3.1578	1.0	1.0
3.172	16.0	752	3.1574	1.0	1.0
3.1514	17.0	799	3.1510	1.0	1.0
3.1505	18.0	846	3.1493	1.0	1.0
3.148	19.0	893	3.1484	1.0	1.0
3.1424	20.0	940	3.1446	1.0	1.0
3.1448	21.0	987	3.1420	1.0	1.0
3.139	22.0	1034	3.1400	1.0	1.0
3.1408	23.0	1081	3.1358	1.0	1.0
3.1357	24.0	1128	3.1326	1.0	1.0
3.1322	25.0	1175	3.1284	1.0	1.0
3.1246	26.0	1222	3.1256	1.0	1.0
3.1294	27.0	1269	3.1162	1.0	1.0
3.1238	28.0	1316	3.1045	1.0	1.0
3.1183	29.0	1363	3.0778	1.0	1.0
3.0784	30.0	1410	3.1226	1.0	1.0
3.0611	31.0	1457	3.0231	1.0	1.0
3.0356	32.0	1504	3.0102	1.0	1.0
3.0014	33.0	1551	2.9689	1.0	1.0
2.9721	34.0	1598	2.9379	1.0	1.0
2.9472	35.0	1645	2.8771	1.0	1.0
2.9201	36.0	1692	2.8427	1.0	1.0
2.8779	37.0	1739	2.8147	1.0	1.0
2.8452	38.0	1786	2.7851	1.0	1.0
2.8276	39.0	1833	2.7496	1.0	0.9922
2.8059	40.0	1880	2.7214	1.0	0.9647
2.783	41.0	1927	2.6950	1.0	0.9273
2.7501	42.0	1974	2.6865	1.0	0.9097
2.7372	43.0	2021	2.6548	1.0	0.9125
2.7148	44.0	2068	2.6345	1.0	0.9083
2.687	45.0	2115	2.6206	1.0	0.8878
2.6763	46.0	2162	2.5913	1.0	0.8920
2.6653	47.0	2209	2.5812	1.0	0.8730
2.6526	48.0	2256	2.5677	1.0	0.8701
2.6442	49.0	2303	2.5460	1.0	0.8645
2.6226	50.0	2350	2.5414	1.0	0.8589
2.6065	51.0	2397	2.5251	1.0	0.8659
2.5939	52.0	2444	2.5120	1.0	0.8631
2.5973	53.0	2491	2.5066	1.0	0.8589
2.5888	54.0	2538	2.4987	1.0	0.8567
2.5878	55.0	2585	2.4962	1.0	0.8560
2.5608	56.0	2632	2.4851	1.0	0.8574
2.5687	57.0	2679	2.4839	1.0	0.8567
2.5594	58.0	2726	2.4864	1.0	0.8589
2.557	59.0	2773	2.4757	1.0	0.8567
2.5553	60.0	2820	2.4769	1.0	0.8567

Framework versions

Transformers 4.28.0
Pytorch 2.1.0+cu121
Datasets 2.16.1
Tokenizers 0.13.3

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: apache-2.0
Created: Jan 22, 2024
Updated: Jan 22, 2024