em_ctc

sz-mp/em_ctc

em_ctc is a automatic speech recognition model from sz-mp. 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

29% of all-time downloads

All-time downloads

Public

Repo size

71.9 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: Sep 27, 2023
Updated: Oct 4, 2023
SHA: fdbbc884

Base models

facebook/wav2vec2-large-xlsr-53

Datasets

audiofolder

Model card

From the Hugging Face model README

em_ctc

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the WAV_SUB-P001 - TR dataset. It achieves the following results on the evaluation set:

Loss: 2.7286
Wer: 0.9843

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: 0.0003
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.16	100	3.8990	1.0
No log	0.32	200	2.8678	1.0
No log	0.48	300	2.7795	1.0
No log	0.65	400	2.7316	1.0
4.389	0.81	500	2.6889	0.9784
4.389	0.97	600	3.2727	0.9784
4.389	1.13	700	2.7057	0.9784
4.389	1.29	800	2.8525	0.9964
4.389	1.45	900	2.6685	0.9968
2.5649	1.61	1000	2.7403	1.0
2.5649	1.78	1100	2.7790	1.0
2.5649	1.94	1200	2.8130	0.9786
2.5649	2.1	1300	2.8031	1.0
2.5649	2.26	1400	2.9683	1.0
2.5421	2.42	1500	2.9459	0.9784
2.5421	2.58	1600	2.7052	0.9784
2.5421	2.74	1700	2.7879	0.9786
2.5421	2.91	1800	2.7956	1.0
2.5421	3.07	1900	2.7760	0.9784
2.5357	3.23	2000	2.8594	0.9995
2.5357	3.39	2100	2.9048	0.9796
2.5357	3.55	2200	3.0098	0.9784
2.5357	3.71	2300	2.7079	0.9784
2.5357	3.87	2400	3.2403	1.0
2.5203	4.04	2500	3.0476	0.9784
2.5203	4.2	2600	2.8510	1.0
2.5203	4.36	2700	2.7907	0.9784
2.5203	4.52	2800	2.7486	0.9784
2.5203	4.68	2900	3.1701	1.0
2.5191	4.84	3000	2.9529	0.9784
2.5191	5.0	3100	3.1192	0.9650
2.5191	5.17	3200	2.8596	1.0
2.5191	5.33	3300	2.9193	1.0
2.5191	5.49	3400	3.0367	0.9784
2.5422	5.65	3500	2.9162	0.9784
2.5422	5.81	3600	3.0334	1.0
2.5422	5.97	3700	2.8514	0.9784
2.5422	6.13	3800	2.9654	1.0
2.5422	6.3	3900	3.2616	0.9784
2.5062	6.46	4000	3.3320	0.9793
2.5062	6.62	4100	2.7141	0.9784
2.5062	6.78	4200	3.2108	0.9784
2.5062	6.94	4300	3.0015	0.9784
2.5062	7.1	4400	3.0244	1.0
2.5114	7.26	4500	2.8742	0.9784
2.5114	7.43	4600	3.1471	0.9784
2.5114	7.59	4700	2.7006	0.9773
2.5114	7.75	4800	3.1189	1.0
2.5114	7.91	4900	3.1604	0.9784
2.5065	8.07	5000	2.9297	0.9784
2.5065	8.23	5100	3.0998	0.9784
2.5065	8.39	5200	2.8184	0.9843
2.5065	8.56	5300	2.7133	0.9861
2.5065	8.72	5400	2.7399	0.9811
2.4956	8.88	5500	2.7186	0.9889
2.4956	9.04	5600	2.9872	0.9955
2.4956	9.2	5700	3.0825	0.9993
2.4956	9.36	5800	3.0589	0.9855
2.4956	9.52	5900	2.8177	0.9784
2.4774	9.69	6000	2.8104	0.9993
2.4774	9.85	6100	2.9498	0.9796
2.4774	10.01	6200	3.0006	0.9784
2.4774	10.17	6300	2.8100	0.9784
2.4774	10.33	6400	3.1577	0.9786
2.4689	10.49	6500	2.7814	0.9977
2.4689	10.65	6600	2.7271	0.9836
2.4689	10.82	6700	2.8403	0.9784
2.4689	10.98	6800	2.7257	0.9998
2.4689	11.14	6900	2.6728	0.9898
2.486	11.3	7000	2.7348	0.9809
2.486	11.46	7100	2.7054	0.9982
2.486	11.62	7200	2.7254	0.9948
2.486	11.78	7300	2.7498	0.9891
2.486	11.95	7400	2.7076	0.9898
2.4616	12.11	7500	2.6398	0.9995
2.4616	12.27	7600	2.7626	0.9846
2.4616	12.43	7700	2.6804	0.9814
2.4616	12.59	7800	2.8212	0.9834
2.4616	12.75	7900	2.6535	0.9959
2.4573	12.91	8000	2.7547	0.9993
2.4573	13.08	8100	2.7253	0.9798
2.4573	13.24	8200	2.6851	0.9936
2.4573	13.4	8300	2.7627	0.9907
2.4573	13.56	8400	2.6607	0.9857
2.4487	13.72	8500	2.6645	0.9800
2.4487	13.88	8600	2.7558	0.9973
2.4487	14.04	8700	2.7665	0.9961
2.4487	14.21	8800	2.7697	0.9827
2.4487	14.37	8900	2.8531	0.9918
2.4416	14.53	9000	2.8974	0.9920
2.4416	14.69	9100	2.7308	0.9975
2.4416	14.85	9200	2.7919	0.9816
2.4416	15.01	9300	2.6605	0.9893
2.4416	15.17	9400	2.6058	0.9816
2.4405	15.33	9500	2.6366	0.9911
2.4405	15.5	9600	2.5653	0.9818
2.4405	15.66	9700	2.7026	0.9807
2.4405	15.82	9800	2.7358	0.9796
2.4405	15.98	9900	2.6954	0.9848
2.4352	16.14	10000	2.6610	0.9857
2.4352	16.3	10100	2.7686	0.9811
2.4352	16.46	10200	2.7758	0.9798
2.4352	16.63	10300	2.7515	0.9848
2.4352	16.79	10400	2.7264	0.9911
2.4354	16.95	10500	2.7039	0.9791
2.4354	17.11	10600	2.7580	0.9843
2.4354	17.27	10700	2.7187	0.9855
2.4354	17.43	10800	2.7545	0.9798
2.4354	17.59	10900	2.7452	0.9809
2.4321	17.76	11000	2.6804	0.9836
2.4321	17.92	11100	2.6586	0.9891
2.4321	18.08	11200	2.6805	0.9830
2.4321	18.24	11300	2.6626	0.9871
2.4321	18.4	11400	2.7002	0.9809
2.4193	18.56	11500	2.7054	0.9839
2.4193	18.72	11600	2.7171	0.9900
2.4193	18.89	11700	2.7122	0.9852
2.4193	19.05	11800	2.7058	0.9871
2.4193	19.21	11900	2.7004	0.9839
2.4276	19.37	12000	2.7250	0.9852
2.4276	19.53	12100	2.7126	0.9861
2.4276	19.69	12200	2.7388	0.9834
2.4276	19.85	12300	2.7311	0.9850

Framework versions

Transformers 4.34.0.dev0
Pytorch 2.0.1+cu118
Datasets 2.14.5
Tokenizers 0.14.0

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: apache-2.0
Created: Sep 27, 2023
Updated: Oct 4, 2023