Herbert

Added required installation step, it worked for me in a python 3.5 virtual environment (no sudo).

Restart worker after idle period

2

**I'll work around this, so this is just a suggestion, to help the AI-comrads, as these often have long running tasks and hw/memory requirements.** - [x] I have checked the...

Llava becomes repetitive on OCR tasks

2

### Describe the issue OCR might not be the target task of Llava, but data is data and I still wanted to make a quick report on this. I tried...

Google Colab currently packs Cython 3, which doesn't compile the required `pandas` well: **ERROR:** ``` [16/40] Cythonizing pandas\_libs/reduction.pyx Error compiling Cython file: ------------------------------------------------------------ ... self.orig_data = self.buf.data self.orig_len = self.buf.shape[0]...

Fuzzy regular expressions in javascript with webasm

1

I made a crude version for javascript support with webasm, anyone interested in improving or desperate for this support can find [my repo here](https://github.com/prhbrt/trejs). I wrote [javascript-code](https://github.com/prhbrt/trejs/blob/main/src/tre.js#L19) to conveniently process...

`Write` seems to select one token rather than write them all

1

### Describe the issue as clearly as possible: [The Write instruction](https://github.com/dottxt-ai/outlines/blob/11143dfa2d603a1f98adbcdedd546ee3722f3ea5/outlines/fsm/guide.py#L22) suggests in the documentation that several tokens will be written at a certain stage, e.g. these tokens are forced...

All settings gone after update

2

### Description All my settings were gone on ubuntu 24.04, probably after an update and rebooted. I reverted ``` $ sudo snap revert brave brave reverted to 1.70.123 ``` And...

OCR source text extraction guide

9

I created a Guide to extract spans from source texts, as if they are fields from a registry. https://github.com/UG-Team-Data-Science/outlines-ocr-guidance Would this be something to integrate in outlines?