Edit model card

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

One more step before getting this model.

This model is open access and available only for non-commercial use, with an SIL International AI & NLP RAIL-M license further specifying rights and usage.

The SIL RAIL-M License specifies:

  1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content. Particularly, you cannot use the model use with the intent or effect of harming or enabling discrimination against Indigenous People.
  2. SIL claims no rights on outputs you generate for non-commercial use, you are free to use them and are accountable for their use, which must not go against the provisions set in the license
  3. You may re-distribute the weights and use the model non-commercially including as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the SIL International AI & NLP RAIL-M to all your users (please read the license entirely and carefully). Please read the full license here: https://huggingface.co/spaces/sil-ai/model-license

By clicking on "Access repository" below, you accept that your contact information (email address and username) can be shared with the model authors as well.

If you would like to ask about commercial uses of this model, please email us.

Log in or Sign Up to review the conditions and access this model content.

wav2vec2-bloom-speech-fra

logo for Bloom Library sil-ai logo

Model description

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the SIL-AI/bloom-speech - FRA dataset. It achieves the following results on the evaluation set:

  • Loss: 0.9346
  • Wer: 0.3051
  • Cer: 0.0978

Users should refer to the original model for tutorials on using a trained model for inference.

Intended uses & limitations

Users of this model must abide by the SIL RAIL-M License.

This model is created as a proof of concept and no guarantees are made regarding the performance of the model is specific situations.

Training and evaluation data

Training, Validation, and Test datasets were generated from the same corpus, ensuring that no duplicate files were used.

Training procedure

Standard finetuning of XLS-R was used based on the examples in the Hugging Face Transformers Github

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: 250
  • num_epochs: 1000.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer Cer
No log 41.62 250 1.0210 0.7260 0.2412
2.4099 83.31 500 0.8338 0.4774 0.1514
2.4099 124.92 750 0.8023 0.4138 0.1205
0.0364 166.62 1000 0.7624 0.4068 0.1261
0.0364 208.31 1250 0.8166 0.3644 0.1149
0.0226 249.92 1500 0.8175 0.3686 0.1202
0.0226 291.62 1750 0.9210 0.3715 0.1170
0.0143 333.31 2000 0.8273 0.3729 0.1175
0.0143 374.92 2250 0.8743 0.3418 0.1109
0.0093 416.62 2500 0.8424 0.3432 0.1165
0.0093 458.31 2750 0.8365 0.3333 0.1053
0.0064 499.92 3000 0.9097 0.3432 0.1111
0.0064 541.62 3250 0.9744 0.3404 0.1098
0.0041 583.31 3500 0.9257 0.3489 0.1079
0.0041 624.92 3750 0.9047 0.3220 0.1031
0.0026 666.62 4000 0.8852 0.3263 0.1026
0.0026 708.31 4250 0.9000 0.3249 0.0986
0.0016 749.92 4500 0.9389 0.3376 0.1005
0.0016 791.62 4750 0.9329 0.3079 0.0970
0.0017 833.31 5000 0.9322 0.3164 0.0991
0.0017 874.92 5250 0.9351 0.3079 0.0991
0.0008 916.62 5500 0.9346 0.3051 0.0978
0.0008 958.31 5750 0.9391 0.3093 0.0991
0.0007 999.92 6000 0.9435 0.3065 0.0978

Framework versions

  • Transformers 4.21.0.dev0
  • Pytorch 1.9.0+cu111
  • Datasets 2.2.2
  • Tokenizers 0.12.1
Downloads last month
0
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Evaluation results