Michele Dolfi
Michele Dolfi
Python 3.9 seems easier than expected. It is coming soon.
Your point 1 is definitely the start. It will work with some models but not all of them yet. We are working on enabling explicit settings (and good defaults) to...
We have a follow up issue to allow proper device selection in all models. We will follow up there.
It would be possible to implement it, but be aware of 1. The OCR latency will for sure increase substantially 2. This should be clearly communicated to the users opting...
Pointing you to the new tech report. In Figure 1 we outline the steps in the PDF pipeline: https://github.com/DS4SD/docling/discussions/574 If you have a scanned document (I assume since you refer...
Do you get the error when installing Docling, i.e. `pip install docling` or when running it? The error you posted looks like a compilation issue which, in case, would happen...
We never encounter (yet) such an error. I'm a bit suspicious about the CUDA and Pytorch versions. I would recommend making sure torch and cuda are compatible, e.g. using the...
> * single A10 GPU > > * cpu 9 > > * ram 24Gb Note that these are just the reservation values. The actual usage is 1.5GB-2GB of RAM...
## Offline LLM ### vLLM Pros: - efficiently run vision models offline, see the [docs page](https://docs.vllm.ai/en/latest/getting_started/examples/offline_inference_vision_language.html). - supports different models without further specialization - already used by InstructLab and part...