Update README.md

README.md

@@ -1,6 +1,10 @@
 ---
 language: en
 datasets:
+- libri_light
+- common_voice
+- switchboard
+- fisher
 - librispeech_asr
 tags:
 - speech
@@ -14,22 +18,32 @@ widget:
 license: apache-2.0
 ---
 
-# Wav2Vec2-Large-960h-Lv60 + Self-Training
+# Wav2Vec2-Large-Robust finetuned on Librispeech
 
-[Facebook's Wav2Vec2](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/)
+[Facebook's Wav2Vec2](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/).
 
-The large model pretrained and fine-tuned on 960 hours of Libri-Light and Librispeech on 16kHz sampled speech audio. Model was trained with [Self-Training objective](https://arxiv.org/abs/2010.11430). When using the model make sure that your speech input is also sampled at 16Khz.
+This model is a fine-tuned version of the (wav2vec2-large-robust)[https://huggingface.co/facebook/wav2vec2-large-robust] model.
+It has been pretrained on:
 
-[Paper](https://arxiv.org/abs/2006.11477)
+- [Libri-Light](https://github.com/facebookresearch/libri-light): open-source audio books from the LibriVox project; clean, read-out audio data
+- [CommonVoice](https://huggingface.co/datasets/common_voice): crowd-source collected audio data; read-out text snippets
+- [Switchboard](https://catalog.ldc.upenn.edu/LDC97S62): telephone speech corpus; noisy telephone data
+- [Fisher](https://catalog.ldc.upenn.edu/LDC2004T19): conversational telephone speech; noisy telephone data
 
-Authors: Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli
+and subsequently been finetuned on 960 hours of
 
-**Abstract**
+- [Librispeech](https://huggingface.co/datasets/librispeech_asr): open-source read-out audio data.
 
-We show for the first time that learning powerful representations from speech audio alone followed by fine-tuning on transcribed speech can outperform the best semi-supervised methods while being conceptually simpler. wav2vec 2.0 masks the speech input in the latent space and solves a contrastive task defined over a quantization of the latent representations which are jointly learned. Experiments using all labeled data of Librispeech achieve 1.8/3.3 WER on the clean/other test sets. When lowering the amount of labeled data to one hour, wav2vec 2.0 outperforms the previous state of the art on the 100 hour subset while using 100 times less labeled data. Using just ten minutes of labeled data and pre-training on 53k hours of unlabeled data still achieves 4.8/8.2 WER. This demonstrates the feasibility of speech recognition with limited amounts of labeled data.
+When using the model make sure that your speech input is also sampled at 16Khz. 
 
-The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20.
+[Paper Robust Wav2Vec2](https://arxiv.org/abs/2104.01027)
+
+Authors: Wei-Ning Hsu, Anuroop Sriram, Alexei Baevski, Tatiana Likhomanenko, Qiantong Xu, Vineel Pratap, Jacob Kahn, Ann Lee, Ronan Collobert, Gabriel Synnaeve, Michael Auli
 
+**Abstract**
+Self-supervised learning of speech representations has been a very active research area but most work is focused on a single domain such as read audio books for which there exist large quantities of labeled and unlabeled data. In this paper, we explore more general setups where the domain of the unlabeled data for pre-training data differs from the domain of the labeled data for fine-tuning, which in turn may differ from the test data domain. Our experiments show that using target domain data during pre-training leads to large performance improvements across a variety of setups. On a large-scale competitive setup, we show that pre-training on unlabeled in-domain data reduces the gap between models trained on in-domain and out-of-domain labeled data by 66%-73%. This has obvious practical implications since it is much easier to obtain unlabeled target domain data than labeled data. Moreover, we find that pre-training on multiple domains improves generalization performance on domains not seen during training. Code and models will be made available at this https URL.
+
+The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20.
 
 # Usage
 
@@ -42,8 +56,8 @@ To transcribe audio files the model can be used as a standalone acoustic model a
  import torch
  
  # load model and processor
- processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-large-960h-lv60-self")
- model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-large-960h-lv60-self")
+ processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-large-robust-ft-libri-960h")
+ model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-large-robust-ft-libri-960h")
  
  # define function to read in sound file
  def map_to_array(batch):