nithinraok commited on
Commit
46d69ed
1 Parent(s): 0b65545

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +133 -0
README.md CHANGED
@@ -168,3 +168,136 @@ metrics:
168
  - wer
169
  pipeline_tag: automatic-speech-recognition
170
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
168
  - wer
169
  pipeline_tag: automatic-speech-recognition
170
  ---
171
+
172
+
173
+ # Canary 1B
174
+
175
+ <style>
176
+ img {
177
+ display: inline;
178
+ }
179
+ </style>
180
+
181
+ [![Model architecture](https://img.shields.io/badge/Model_Arch-FastConformer--Transformer-lightgrey#model-badge)](#model-architecture)
182
+ | [![Model size](https://img.shields.io/badge/Params-1B-lightgrey#model-badge)](#model-architecture)
183
+ | [![Language](https://img.shields.io/badge/Language-en-lightgrey#model-badge)](#datasets)
184
+ | [![Language](https://img.shields.io/badge/Language-de-lightgrey#model-badge)](#datasets)
185
+ | [![Language](https://img.shields.io/badge/Language-es-lightgrey#model-badge)](#datasets)
186
+ | [![Language](https://img.shields.io/badge/Language-fr-lightgrey#model-badge)](#datasets)
187
+
188
+ CHANGE FROM HERE
189
+ `parakeet-rnnt-1.1b` is an ASR model that transcribes speech in lower case English alphabet. This model is jointly developed by [NVIDIA NeMo](https://github.com/NVIDIA/NeMo) and [Suno.ai](https://www.suno.ai/) teams.
190
+ It is an XXL version of FastConformer Transducer [1] (around 1.1B parameters) model.
191
+ See the [model architecture](#model-architecture) section and [NeMo documentation](https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/main/asr/models.html#fast-conformer) for complete architecture details.
192
+
193
+ ## NVIDIA NeMo: Training
194
+
195
+ To train, fine-tune or play with the model you will need to install [NVIDIA NeMo](https://github.com/NVIDIA/NeMo). We recommend you install it after you've installed latest PyTorch version.
196
+ ```
197
+ pip install nemo_toolkit['all']
198
+ ```
199
+
200
+ ## How to Use this Model
201
+
202
+ The model is available for use in the NeMo toolkit [3], and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.
203
+
204
+ ### Automatically instantiate the model
205
+
206
+ ```python
207
+ import nemo.collections.asr as nemo_asr
208
+ asr_model = nemo_asr.models.EncDecRNNTBPEModel.from_pretrained(model_name="nvidia/parakeet-rnnt-1.1b")
209
+ ```
210
+
211
+ ### Transcribing using Python
212
+ First, let's get a sample
213
+ ```
214
+ wget https://dldata-public.s3.us-east-2.amazonaws.com/2086-149220-0033.wav
215
+ ```
216
+ Then simply do:
217
+ ```
218
+ asr_model.transcribe(['2086-149220-0033.wav'])
219
+ ```
220
+
221
+ ### Transcribing many audio files
222
+
223
+ ```shell
224
+ python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py
225
+ pretrained_name="nvidia/parakeet-rnnt-1.1b"
226
+ audio_dir="<DIRECTORY CONTAINING AUDIO FILES>"
227
+ ```
228
+
229
+ ### Input
230
+
231
+ This model accepts 16000 Hz mono-channel audio (wav files) as input.
232
+
233
+ ### Output
234
+
235
+ This model provides transcribed speech as a string for a given audio sample.
236
+
237
+ ## Model Architecture
238
+
239
+ FastConformer [1] is an optimized version of the Conformer model with 8x depthwise-separable convolutional downsampling. The model is trained in a multitask setup with a Transducer decoder (RNNT) loss. You may find more information on the details of FastConformer here: [Fast-Conformer Model](https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/main/asr/models.html#fast-conformer).
240
+
241
+ ## Training
242
+
243
+ The NeMo toolkit [3] was used for training the models for over several hundred epochs. These model are trained with this [example script](https://github.com/NVIDIA/NeMo/blob/main/examples/asr/asr_transducer/speech_to_text_rnnt_bpe.py) and this [base config](https://github.com/NVIDIA/NeMo/blob/main/examples/asr/conf/fastconformer/fast-conformer_transducer_bpe.yaml).
244
+
245
+ The tokenizers for these models were built using the text transcripts of the train set with this [script](https://github.com/NVIDIA/NeMo/blob/main/scripts/tokenizers/process_asr_text_tokenizer.py).
246
+
247
+ ### Datasets
248
+
249
+ The model was trained on 64K hours of English speech collected and prepared by NVIDIA NeMo and Suno teams.
250
+
251
+ The training dataset consists of private subset with 40K hours of English speech plus 24K hours from the following public datasets:
252
+
253
+ - Librispeech 960 hours of English speech
254
+ - Fisher Corpus
255
+ - Switchboard-1 Dataset
256
+ - WSJ-0 and WSJ-1
257
+ - National Speech Corpus (Part 1, Part 6)
258
+ - VCTK
259
+ - VoxPopuli (EN)
260
+ - Europarl-ASR (EN)
261
+ - Multilingual Librispeech (MLS EN) - 2,000 hour subset
262
+ - Mozilla Common Voice (v7.0)
263
+ - People's Speech - 12,000 hour subset
264
+
265
+ ## Performance
266
+
267
+ The performance of Automatic Speech Recognition models is measuring using Word Error Rate. Since this dataset is trained on multiple domains and a much larger corpus, it will generally perform better at transcribing audio in general.
268
+
269
+ The following tables summarizes the performance of the available models in this collection with the Transducer decoder. Performances of the ASR models are reported in terms of Word Error Rate (WER%) with greedy decoding.
270
+
271
+ |**Version**|**Tokenizer**|**Vocabulary Size**|**AMI**|**Earnings-22**|**Giga Speech**|**LS test-clean**|**SPGI Speech**|**TEDLIUM-v3**|**Vox Populi**|**Common Voice**|
272
+ |---------|-----------------------|-----------------|---------------|---------------|------------|-----------|-----|-------|------|------|
273
+ | 1.22.0 | SentencePiece Unigram | 1024 | 17.10 | 14.11 | 9.96 | 1.46 | 2.47 | 3.11 | 3.92 | 5.39 | 5.79 |
274
+
275
+ These are greedy WER numbers without external LM. More details on evaluation can be found at [HuggingFace ASR Leaderboard](https://huggingface.co/spaces/hf-audio/open_asr_leaderboard)
276
+
277
+ ## NVIDIA Riva: Deployment
278
+
279
+ [NVIDIA Riva](https://developer.nvidia.com/riva), is an accelerated speech AI SDK deployable on-prem, in all clouds, multi-cloud, hybrid, on edge, and embedded.
280
+ Additionally, Riva provides:
281
+
282
+ * World-class out-of-the-box accuracy for the most common languages with model checkpoints trained on proprietary data with hundreds of thousands of GPU-compute hours
283
+ * Best in class accuracy with run-time word boosting (e.g., brand and product names) and customization of acoustic model, language model, and inverse text normalization
284
+ * Streaming speech recognition, Kubernetes compatible scaling, and enterprise-grade support
285
+
286
+ Although this model isn’t supported yet by Riva, the [list of supported models is here](https://huggingface.co/models?other=Riva).
287
+ Check out [Riva live demo](https://developer.nvidia.com/riva#demos).
288
+
289
+ ## References
290
+ [1] [Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognition](https://arxiv.org/abs/2305.05084)
291
+
292
+ [2] [Google Sentencepiece Tokenizer](https://github.com/google/sentencepiece)
293
+
294
+ [3] [NVIDIA NeMo Toolkit](https://github.com/NVIDIA/NeMo)
295
+
296
+ [4] [Suno.ai](https://suno.ai/)
297
+
298
+ [5] [HuggingFace ASR Leaderboard](https://huggingface.co/spaces/hf-audio/open_asr_leaderboard)
299
+
300
+
301
+ ## Licence
302
+
303
+ License to use this model is covered by the [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/). By downloading the public and release version of the model, you accept the terms and conditions of the [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/) license.