HC Heine

One could use `get_pages_for_items()` for this right?

Could it be that the targetid is getting stuck during learning for the implementation given in the [3.6 fork](https://github.com/free-soellingeraj/open-sesame)? How many epochs till one can get improvements in avg. loss...

I also have this issue. In the [m2repo](https://cogcomp.seas.upenn.edu/m2repo/edu/illinois/cs/cogcomp/resources/gazetteers/) I cannot find gazetteers-list.txt, but the jar which has it inside.

What should the endpoint be? Or how can I find these: ``` readonly.org.cogcomp.mention 1.0\ACE_HEAD_TYPE.zip \1.0\ACE_HEAD_TYPE ``` And `ner-model-enron-conll-all-data.zip`

Same here

I also get it (with smaller Params) when running RTX 2080 TI OR GTX 1060 with codeqwen:chat and codegamma:instruct for Win10. It stays in RAM but will have to copy...

I could not get the .so building as well (@mengzhi7, could you even get g++ to build from the .cc operation file?) On windows, it seems like a complicated task,...

I think the setup_all.sh won't compile the needed coref_kernels.so on windows -> https://stackoverflow.com/questions/50967670/g-fails-to-build-custom-tensorflow-gpu-op-on-windows-from-installed-binaries

Qwen VL works much better than LLava 1.6, would be good to be able to use it with ollama. The OCR is also much better and supports much more languages.

with reranking:   without reranking