Hugging Face's logo

KBLab
/
wav2vec2-large-voxrex-swedish

Automatic Speech Recognition
Transformers
PyTorch
Safetensors
Swedish
wav2vec2
audio
speech
hf-asr-leaderboard
Eval Results
Inference Endpoints
Model card Files Community
3
marma commited on
Commit
c9b8f60
•
1 Parent(s): 8da40f8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -0
README.md CHANGED
@@ -32,6 +32,7 @@ model-index:
32
  **Disclaimer:** This is a work in progress. See [VoxRex](https://huggingface.co/KBLab/wav2vec2-large-voxrex) for more details.
33
 
34
  **Update 2022-01-10:** Updated to VoxRex-C version.
 
35
 
36
  Finetuned version of KBs [VoxRex large](https://huggingface.co/KBLab/wav2vec2-large-voxrex) model using Swedish radio broadcasts, NST and Common Voice data. Evalutation without a language model gives the following: WER for NST + Common Voice test set (2% of total sentences) is **2.5%**. WER for Common Voice test set is **8.49%** directly and **7.37%** with a 4-gram language model.
37
 
 
32
  **Disclaimer:** This is a work in progress. See [VoxRex](https://huggingface.co/KBLab/wav2vec2-large-voxrex) for more details.
33
 
34
  **Update 2022-01-10:** Updated to VoxRex-C version.
35
+ **Update 2022-05-16:** Paper is is [here](https://arxiv.org/abs/2205.03026).
36
 
37
  Finetuned version of KBs [VoxRex large](https://huggingface.co/KBLab/wav2vec2-large-voxrex) model using Swedish radio broadcasts, NST and Common Voice data. Evalutation without a language model gives the following: WER for NST + Common Voice test set (2% of total sentences) is **2.5%**. WER for Common Voice test set is **8.49%** directly and **7.37%** with a 4-gram language model.
38