-
### System Info
```shell
optimum 1.9.1
onnx 1.14.0
onnxconverter-common 1.13.0
onnxruntime 1.15.1
onnxruntime-tools …
-
I would like to ask everyone if they have tried converting short audio in this project and how effective it is. Has anyone compared the inference speed of this project with Faster-Whisper or WhisperX…
-
I encountered this error when applying this repo on my data. How to fix it?
-
Is there inference code? I could not find any. but I read through other issues and found this.
i'll write a inference script next so we can do some quick experiments.
_Originally p…
-
Hi, congrats on the project!
I am getting this error, could not figure out what's going on:
File "/media/user/hk-data/PycharmProjects/dnn_embeddings_pytorch/train_model.py", line 148, in train_m…
-
I'm trying to figure out if fast-wavenet-decoding (https://github.com/tomlepaine/fast-wavenet) is used when decoding, if not, it's quite simple to implement and would yield a big performance boost, of…
-
We need to convert keras.io examples to work with Keras 3.
This involves two stages:
## Stage 1: tf.keras backwards compatibility check
Keras 3 is intended as a drop-in replacement for tf.ker…
-
Hi @nyadla-sys ,
this is a very interesting work on OpenAI's Whisper 🙂👍.
I've built a multi engine, streaming server for STT ([SEPIA STT-Server](https://github.com/SEPIA-Framework/sepia-stt-serv…
-
Hi, can you explain the difference between the subband and duration experiments and share which you've found to perform better? Also, what have you discovered about the use of classifier free guidance…
-
Have you tried building the spectrogram and encoder output in smaller chunks and appending? I think the spectrogram should generate fairly easily with minimal noise depending on the size of the chunk,…