-
### What happened?
When flash attention is enabled, generating parallel sequences in a batch produces corrupt output on some GPUs. It seems to only happen on certain batch sizes. I've tested with lla…
-
**Describe the bug**
Sometimes, after asking a few questions to the AI, it gets stuck "thinking". I have to stop the LLM back-end manualy everytime to end the loop
**Expected behavior**
Should ne…
-
Hi,
how can we add a translation (language file) of the GUI? There is not a lot in English but our customer prefers German language.
If someone explains how to translate I will make the translation …
-
Hi
Once speech has been transcribed into text, it should be possible to edit this text directly in the window where it is displayed. After which there should be a button at the bottom of the window…
-
# Expected Behavior
I expected finetune to produce a usable LoRA adapter for all supported models.
# Current Behavior
For Mistral models (I tried both Mistral and Zephyr, Q8_0, Q5_K_M, Q5_0) …
maxxk updated
7 months ago
-
One issue to discuss everything that's difficult to understand in current version of documentation.
Click to expand
>Channels are present in the audio device(?)
From .txt: "list of channe…
-
At multiple instances across the web5-js docs and pages, I noticed that there is a missing full stop at the end of the message under the create(request) as done for the read(request) and other methods…
-
- Cum Laude candidate
[USA House of Representatives - INVESTIGATION OF COMPETITION IN DIGITAL MARKETS ](https://www.govinfo.gov/content/pkg/CPRT-117HPRT47832/pdf/CPRT-117HPRT47832.pdf) :
```
As p…
-
Hi!
Recently I launched transcription and received such error:
File "/usr/local/lib/python3.9/site-packages/whisper_timestamped/transcribe.py", line 259, in transcribe_timestamped
(transcriptio…
-
Doing this will allow the public talks coordinator to select according to category
Thank you