Update README.md

README.md

@@ -8,24 +8,25 @@ tags:
 license: apache-2.0
 ---
 
-# Wav2Vec2-1B-XLS-R
+# Wav2Vec2-XLS-R-300M
 
-[Facebook's Wav2Vec2 XLS-R](https://ai.facebook.com/blog/)
+[Facebook's Wav2Vec2 XLS-R](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/)
 
-XLS-R is Facebook AI's large-scale multilingual pretrained model for speech (the "XLM-R for Speech"), pretrained on 436k hours of unlabeled speech with the wav2vec 2.0 approach, in 128 languages. When using the model make sure that your speech input is sampled at 16Khz. Note that this model should be fine-tuned on a downstream task, like Automatic Speech Recognition, Translation or Classification. Check out [this blog](https://huggingface.co/blog/fine-tune-wav2vec2-english) for more information about ASR.
+XLS-R is Facebook AI's large-scale multilingual pretrained model for speech (the "XLM-R for Speech"). It is pretrained on 436k hours of unlabeled speech, including VoxPopuli, MLS, CommonVoice, BABEL and VoxLingua107. Is uses the wav2vec 2.0 objective, in 128 languages. When using the model make sure that your speech input is sampled at 16Khz. Note that this model should be fine-tuned on a downstream task, like Automatic Speech Recognition, Translation or Classification. Check out [this blog](https://huggingface.co/blog/fine-tune-wav2vec2-english) for more information about ASR.
 
 [XLS-R Paper](https://arxiv.org/abs/)
 
 Authors: ...
 
 **Abstract**
-This paper presents XLS-R, a large-scale study on learning cross-lingual speech representations with wav2vec 2.0.
-Models are pretrained on 436K hours of publicly available speech audio in 128 languages, an order of magnitude more public data than the largest known prior work.
-The best \xlsrp{} model sets a new state of the art on a wide range of tasks, domains, data regimes and languages, both high and low-resource. 
-On the CoVoST-2 speech translation benchmark, we improve the previous state of the art by an average of 5.5 BLEU over 21 translation directions into English.
-In speech recognition, we improve over the state of the art on five languages of BABEL by XX.X average WER. On MLS and VoxPopuli, we beat previous work by XX.X and YY.Y WER respectively on average. We also outperform the best prior work on VoxLingua107 language identification by 1% error rate.
-We hope \xlsrp{} can help improve speech processing tasks for many more languages of the world. 
-Models and code are available at [www.github.com/pytorch/fairseq/tree/master/examples/wav2vec](www.github.com/pytorch/fairseq/tree/master/examples/wav2vec).
+This paper presents XLS-R, a large-scale model for cross-lingual speech representation learning based on wav2vec 2.0. 
+Models are fitted with up to 2B parameters on 436K hours of publicly available speech audio in 128 languages, an order of magnitude more public data than the largest known prior work.
+Our evaluation covers a wide range of tasks, domains, data regimes and languages, both high and low-resource. 
+On the CoVoST-2 speech translation benchmark, we improve the previous state of the art by an average of 7.4 BLEU over 21 translation directions into English.
+For speech recognition, XLS-R improves over the best known prior work on BABEL, MLS, CommonVoice as well as VoxPopuli, lowering error rates by \ma{20\%-33\%} on average.
+XLS-R also sets a new state of the art on VoxLingua107 language identification.
+Moreover, we show that with sufficient model size, cross-lingual pretraining can outperform English-only pretraining when translating English speech into other languages, a setting which favors monolingual pretraining.
+We hope XLS-R can help to improve speech processing tasks for many more languages of the world. 
 
 The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20.
 
@@ -33,4 +34,6 @@ The original model can be found under https://github.com/pytorch/fairseq/tree/ma
 
 See [this notebook](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Fine_Tune_XLSR_Wav2Vec2_on_Turkish_ASR_with_%F0%9F%A4%97_Transformers.ipynb) for more information on how to fine-tune the model.
 
-![model image](https://raw.githubusercontent.com/patrickvonplaten/scientific_images/master/xlsr_wav2vec2.png)
+You can find the [100M parameters version](https://huggingface.co/facebook/wav2vec2-xls-r-100m), the [1B version version](https://huggingface.co/facebook/wav2vec2-xls-r-1b) and the [2B parameters version](h
+![model image](https://raw.githubusercontent.com/patrickvonplaten/scientific_images/master/xls-r.png)
+