PaddlePaddle/PaddleOCR-VL-1.5, lightonai/LightOnOCR-2-1B, zai-org/GLM-OCR, nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-BF16, nvidia/Cosmos-Reason2-8B
I noticed these models have not been quanted by either you, unsloth or bartowski. It will be nice to have them, I think
that's a bit unconvenient way to receive links, but sure, let me try to do that for you =)
but knowing that none of them did quants, it probably means that the models are very unsupported by llama cpp that even custom forks dont exist.
(my very unexperience probably not true opinion) as far as I understood first 3 models could have failed because they are OCR, and we hate nvidia because they love to break configs and blame users because "works for me ^_^"
even my queue manager said that and refused to quant:
model broken, max arrogance. https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2/discussions/5
https://www.youtube.com/watch?v=OF_5EKNX0Eg
It's queued!
You can check for progress at http://hf.tst.eu/status.html or regularly check the model
summary page at https://hf.tst.eu/model#PaddleOCR-VL-1.5-GGUF
https://hf.tst.eu/model#LightOnOCR-2-1B-GGUF
https://hf.tst.eu/model#GLM-OCR-GGUF
https://hf.tst.eu/model#Cosmos-Reason2-8B-GGUF
for quants to appear.