Update readme

README.md

@@ -15,21 +15,25 @@ widget:
 
 # Wav2Vec2-Base-960h
 
-[Facebook's Wav2Vec2](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/)
-
-The base model pretrained and fine-tuned on 960 hours of Librispeech on 16kHz sampled speech audio. When using the model
-make sure that your speech input is also sampled at 16Khz.
-
-[Paper](https://arxiv.org/abs/2006.11477)
-
-Authors: Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli
+This repository is a reimplementation of [official Facebook’s wav2vec](https://huggingface.co/facebook/wav2vec2-base-960h).
+There is no description of converting the wav2vec [pretrain model](https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20) to a pytorch.bin file.
+We are rebuilding pytorch.bin from the pretrain model.
+Here is the conversion method.
 
-**Abstract**
+```bash
+pip install transformers[sentencepiece]
+pip install fairseq -U
 
-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.
+git clone https://github.com/huggingface/transformers.git
+cp transformers/src/transformers/models/wav2vec2/convert_wav2vec2_original_pytorch_checkpoint_to_pytorch.py .
 
-The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20.
+wget https://dl.fbaipublicfiles.com/fairseq/wav2vec/wav2vec_small_960h.pt -O ./finetuning/wav2vec_small_960h.pt
+mkdir dict
+wget https://dl.fbaipublicfiles.com/fairseq/wav2vec/dict.ltr.txt
 
+mkdir outputs
+python convert_wav2vec2_original_pytorch_checkpoint_to_pytorch.py --pytorch_dump_folder_path ./outputs --checkpoint_path ./wav2vec_small_960h.pt --dict_path ./dict
+```
 
 # Usage
 
@@ -109,4 +113,12 @@ print("WER:", wer(result["text"], result["transcription"]))
 
 | "clean" | "other" |
 |---|---|
-| 3.4 | 8.6 |
+| 3.4 | 8.6 |
+
+
+# Reference
+
+
+[Facebook's Wav2Vec2](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/)
+[Facebook's huggingface Wav2Vec2](https://huggingface.co/facebook/wav2vec2-base-960h)
+[Paper](https://arxiv.org/abs/2006.11477)