-
Tried using the container build for RNNT, as well as running within an existing RN50 container.
```
[2024-07-02 14:20:09,097 preprocess_data.py:59 INFO] Converting wav files to npy files for test …
-
Hey guys!
Thanks for the open-source project. It has been extremely helpful for my work.
I am currently in the process of installing the project and have encountered some issues regarding the co…
-
I trained two offline reworked conformer models on my own Chinese data using pruned_rnnt_loss and standard rnnt loss (warp-rnnt==0.7.0) respectively following pruned_transducer_stateless5. However, I …
-
I'm trying to run the rnnt inference on mlperf4.0 with the below commands and after a long time the whole terminal will disppear without any error prompt. I'm not sure whether is my basic environment …
-
will rnnt loss support fastemit?
Mddct updated
1 month ago
-
Appreciate for the job on supporting RNN-T training on CPU (models/language_modeling/pytorch/rnnt/training/cpu), just quick evaluated the training code and found that WER would keep in 1.00 after even…
-
First of all, thanks for this amazing work in benchmarking the several available RNNT implementations. This is more of a "discussion" rather than an issue.
I am sure you are aware about this, but t…
-
pip install warp_rnnt
Collecting warp_rnnt
Using cached warp_rnnt-0.6.0.tar.gz (10 kB)
Preparing metadata (setup.py) ... error
error: subprocess-exited-with-error
× python setup…
iflov updated
2 weeks ago
-
In conformer/convolution.py, line 183, the code
```
output_lengths = input_lengths >> 2
output_lengths -= 1
```
when the result of input_lengths >>2 is xx.75, the torchaudio.transforms.RNNTLoss w…
-
### 🚀 The feature
## Repro
```python
import torch
import torchaudio
from torchaudio.io import StreamReader
bundle = torchaudio.pipelines.EMFORMER_RNNT_BASE_LIBRISPEECH
feature_extractor…