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.

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


license: mit base_model: facebook/w2v-bert-2.0 tags:

  • generated_from_trainer metrics:
  • wer model-index:
  • name: w2v2-bert-Wolof-5-hours-ALFFA-dataset results: []

w2v2-bert-Wolof-5-hours-ALFFA-dataset

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: 0.7038
  • Wer: 0.2426
  • Cer: 0.0697

Model description

  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Wer Cer
1.1501 3.11 400 0.4303 0.4107 0.1101
0.5594 6.23 800 0.5960 0.4697 0.1379
0.4583 9.34 1200 0.5268 0.3951 0.1164
0.352 12.45 1600 0.5083 0.3588 0.1102
0.2712 15.56 2000 0.4915 0.3799 0.1136
0.2112 18.68 2400 0.5377 0.3431 0.1029
0.1488 21.79 2800 0.4510 0.3121 0.0963
0.0956 24.9 3200 0.4963 0.2887 0.0848
0.0651 28.02 3600 0.6041 0.3017 0.0918
0.0399 31.13 4000 0.5339 0.2922 0.0871
0.0342 34.24 4400 0.5940 0.2837 0.0850
0.0145 37.35 4800 0.6339 0.2638 0.0782
0.0062 40.47 5200 0.6818 0.2661 0.0791
0.0037 43.58 5600 0.6538 0.2439 0.0699
0.002 46.69 6000 0.6837 0.2432 0.0702
0.0008 49.81 6400 0.7038 0.2426 0.0697

Framework versions

  • Transformers 4.38.1
Downloads last month
0
Safetensors
Model size
606M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.