xls-r-300m-model-E

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

xls-r-300m-model-E 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

17% 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: eb1e49c1

Model card

From the Hugging Face model README

xls-r-300m-model-E

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: 2.3847
Wer: 1.0
Cer: 0.7819

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
32.0813	1.0	47	31.5461	1.0094	0.9548
29.4098	2.0	94	21.2317	1.0	1.0
21.4795	3.0	141	13.2108	1.0	1.0
14.1161	4.0	188	8.5594	1.0	1.0
10.2708	5.0	235	6.4335	1.0	1.0
7.5093	6.0	282	5.1505	1.0	1.0
5.5936	7.0	329	4.3389	1.0	1.0
4.5311	8.0	376	3.8477	1.0	1.0
4.0117	9.0	423	3.5432	1.0	1.0
3.5793	10.0	470	3.3558	1.0	1.0
3.4033	11.0	517	3.2559	1.0	1.0
3.3079	12.0	564	3.2058	1.0	1.0
3.2444	13.0	611	3.1807	1.0	1.0
3.1968	14.0	658	3.1692	1.0	1.0
3.1719	15.0	705	3.1606	1.0	1.0
3.186	16.0	752	3.1566	1.0	1.0
3.1522	17.0	799	3.1509	1.0	1.0
3.1505	18.0	846	3.1471	1.0	1.0
3.1512	19.0	893	3.1429	1.0	1.0
3.1391	20.0	940	3.1312	1.0	1.0
3.1303	21.0	987	3.0959	1.0	1.0
3.0991	22.0	1034	3.0370	1.0	1.0
3.0508	23.0	1081	3.0680	1.0	1.0
3.0225	24.0	1128	2.9567	1.0	1.0
2.9782	25.0	1175	2.8864	1.0	1.0
2.9184	26.0	1222	2.8570	1.0	1.0
2.8788	27.0	1269	2.8252	1.0	0.9760
2.8529	28.0	1316	2.7908	1.0	0.8525
2.8254	29.0	1363	2.8204	1.0	0.8758
2.7802	30.0	1410	2.7102	1.0	0.8426
2.7555	31.0	1457	2.6736	1.0	0.8405
2.7258	32.0	1504	2.6650	1.0	0.8433
2.7024	33.0	1551	2.6377	1.0	0.8447
2.6743	34.0	1598	2.6364	1.0	0.8391
2.6523	35.0	1645	2.6055	1.0	0.8299
2.6332	36.0	1692	2.5676	1.0	0.8342
2.6095	37.0	1739	2.5468	1.0	0.8215
2.5907	38.0	1786	2.5346	1.0	0.8243
2.5825	39.0	1833	2.5382	1.0	0.8158
2.5661	40.0	1880	2.5160	1.0	0.8250
2.5598	41.0	1927	2.5185	1.0	0.8285
2.5349	42.0	1974	2.5313	1.0	0.8285
2.5275	43.0	2021	2.5078	1.0	0.8109
2.5105	44.0	2068	2.4804	1.0	0.8102
2.4963	45.0	2115	2.4829	1.0	0.8080
2.4929	46.0	2162	2.4405	1.0	0.7946
2.4826	47.0	2209	2.4515	1.0	0.7848
2.4734	48.0	2256	2.4508	1.0	0.7876
2.4701	49.0	2303	2.3954	1.0	0.7791
2.4436	50.0	2350	2.4241	1.0	0.7812
2.4308	51.0	2397	2.4091	1.0	0.7798
2.4252	52.0	2444	2.4084	1.0	0.7833
2.4284	53.0	2491	2.3973	1.0	0.7826
2.4264	54.0	2538	2.3883	1.0	0.7833
2.4126	55.0	2585	2.3703	1.0	0.7805
2.3986	56.0	2632	2.3867	1.0	0.7819
2.3946	57.0	2679	2.3888	1.0	0.7819
2.3992	58.0	2726	2.4034	1.0	0.7819
2.3951	59.0	2773	2.3804	1.0	0.7805
2.3908	60.0	2820	2.3847	1.0	0.7819

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