-
I have 200 hours of training set and 20 hours of validation set. Loss is infinity even after 44 epochs.
I tried with different learning rates but no difference.
Here is the arch:
> SAUG 80 27 2 1…
-
I also faced same issue but it is not getting resolved by rebuilding libsndfiles with opus and rebuilding wav2letter again.
how do I access log to see which file is giving error?
terminate calle…
-
Could you please tell to how to config to save latest n checkpoints? The default recipe only save the best and last one checkpoint. Thanks.
-
### Question
Hi,
Other than the architecture what is the difference between [sota/2019/am_tds_ctc](https://github.com/facebookresearch/wav2letter/blob/master/recipes/sota/2019/librispeech/train_am…
-
I am using Flashlight latest cuda docker image for ASR training. However, when I run make Test command, I come across the following test errors.
```
98% tests passed, 10 tests failed out of 411
…
-
i ran the example provided in the examples section of pyython bindings i am getting error in loading lm
Loading the LM will be faster if you build a binary file.
Reading /home/vignesh/wav2letter/s…
-
Hi,
```
OS: Ubuntu 18.04
Flashlight and Wave2Letter : Compiled from sources (master branch, just pulled today)
Cuda: 10.2 (installed from cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_…
-
### Question
Hi, could you please provide a reference use of the online decoding API for speech to text?
Currently there is a short pseudocode at `lib/text/decoder/Decoder.h`:
```
* Online manner…
-
### Question
A previous issue #265 and my personal experience with the models in the ![RASR recipe](https://github.com/facebookresearch/wav2letter/tree/master/recipes/rasr) shows that the models' …
-
I'm trying to fine tune a pre-trained wav2vec 2.0 base model (no finetuning) on ~1h of data, with the following command:
python3 fairseq/train.py --distributed-world-size 1 \
"path/to/data" \
-…