Shadowgraph
About
Sign in

Bajiyo/w2v-bert-2.0-nonstudio_and_studioRecords_final

Model

5

stars

53

commits

1

repos using this model

1

linked in READMEs

Jun 28, 2024

updated

automatic-speech-recognition
endpoints_compatible
generated_from_trainer
safetensors
tensorboard
transformers
wav2vec2-bert
Browse cluster: Multilingual Legal NLP Models →

README

w2v-bert-2.0-nonstudio_and_studioRecords_final

This model is a fine-tuned version of facebook/w2v-bert-2.0 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1772
  • Wer: 0.1266

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: 5e-05
  • 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: 10
  • mixed_precision_training: Native AMP

Training results

Training LossEpochStepValidation LossWer
1.0550.46016000.36830.4608
0.17340.920212000.26200.3546
0.12421.380418000.21150.3018
0.10751.840524000.20040.2889
0.08882.300630000.18700.2573
0.0782.760736000.17240.2267
0.06643.220942000.15720.2244
0.05763.681048000.17460.2217
0.05224.141154000.16430.1796
0.04154.601260000.17810.1851
0.03985.061366000.16700.1714
0.03015.521572000.15310.1617
0.02965.981678000.14630.1590
0.02116.441784000.15660.1473
0.02066.901890000.14230.1468
0.01477.362096000.14430.1413
0.01367.8221102000.15390.1418
0.01058.2822108000.16110.1383
0.00798.7423114000.17610.1351
0.00639.2025120000.18140.1304
0.00439.6626126000.17720.1266

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.3.1+cu121
  • Datasets 2.19.2
  • Tokenizers 0.19.1

Contributors

Bajiyo

53 commits

Similar repos

GoshKolotyan/w2v-bert-2.0-armenian-CV16.0-version-demo

Model

0

shared topics

speech31/wav2vec2-large-english-TIMIT-phoneme_v3

Model

3

shared topics

r-f/wav2vec-english-speech-emotion-recognition

Model

35

shared topics

ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition

Model

251

shared topics

sanchit-gandhi/whisper-small-hi

Model

10

shared topics

joelito/legal-finnish-roberta-base

Model

1

shared topics

joelito/legal-latvian-roberta-base

Model

0

shared topics

joelito/legal-swedish-roberta-base

Model

0

shared topics

Used in code

VishnuPJ/Malayalam_Speech2Speech_Vision_Text_Chatbot

This repository contains a fully functional speech-to-speech chatbot pipeline that supports…

18

1 file

Linked in READMEs

VishnuPJ/Malayalam_Speech2Speech_Vision_Text_Chatbot

This repository contains a fully functional speech-to-speech chatbot pipeline that supports…

18

README badge

Bajiyo/w2v-bert-2.0-nonstudio_and_studioRecords_final

Model

5

stars

53

commits

1

repos using this model

1

linked in READMEs

Jun 28, 2024

updated

automatic-speech-recognition
endpoints_compatible
generated_from_trainer
safetensors
tensorboard
transformers
wav2vec2-bert
Browse cluster: Multilingual Legal NLP Models →

README

w2v-bert-2.0-nonstudio_and_studioRecords_final

This model is a fine-tuned version of facebook/w2v-bert-2.0 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1772
  • Wer: 0.1266

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: 5e-05
  • 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: 10
  • mixed_precision_training: Native AMP

Training results

Training LossEpochStepValidation LossWer
1.0550.46016000.36830.4608
0.17340.920212000.26200.3546
0.12421.380418000.21150.3018
0.10751.840524000.20040.2889
0.08882.300630000.18700.2573
0.0782.760736000.17240.2267
0.06643.220942000.15720.2244
0.05763.681048000.17460.2217
0.05224.141154000.16430.1796
0.04154.601260000.17810.1851
0.03985.061366000.16700.1714
0.03015.521572000.15310.1617
0.02965.981678000.14630.1590
0.02116.441784000.15660.1473
0.02066.901890000.14230.1468
0.01477.362096000.14430.1413
0.01367.8221102000.15390.1418
0.01058.2822108000.16110.1383
0.00798.7423114000.17610.1351
0.00639.2025120000.18140.1304
0.00439.6626126000.17720.1266

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.3.1+cu121
  • Datasets 2.19.2
  • Tokenizers 0.19.1

Similar repos

GoshKolotyan/w2v-bert-2.0-armenian-CV16.0-version-demo

Model

0

shared topics

speech31/wav2vec2-large-english-TIMIT-phoneme_v3

Model

3

shared topics

r-f/wav2vec-english-speech-emotion-recognition

Model

35

shared topics

ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition

Model

251

shared topics

sanchit-gandhi/whisper-small-hi

Model

10

shared topics

joelito/legal-finnish-roberta-base

Model

1

shared topics

joelito/legal-latvian-roberta-base

Model

0

shared topics

joelito/legal-swedish-roberta-base

Model

0

shared topics

Used in code

VishnuPJ/Malayalam_Speech2Speech_Vision_Text_Chatbot

This repository contains a fully functional speech-to-speech chatbot pipeline that supports…

18

1 file

Linked in READMEs

VishnuPJ/Malayalam_Speech2Speech_Vision_Text_Chatbot

This repository contains a fully functional speech-to-speech chatbot pipeline that supports…

18

README badge

Contributors

Bajiyo

53 commits