https://huggingface.co/ConicCat/NemoPlotter-12B
https://huggingface.co/ConicCat/NemoPlotter-12B
Ye olde Nemo 12B, but with Hybrid thinking. Mostly made this b/c people wanted a nemo based reasoning model.
Still kinda working on the model card though. The qwen3 tokenizer / chat template isn't a mistake, this is based on the arcee distill.
Thanks in advance!
It's queued! :D
Nice I already almost forgot about the Mistral AI/NVIDIA collaboration. It was a great model at the time. I definitely should give it a try and to see how it performs 1 year later but with reasoning. I definitely see massive potential in adding reasoning to older models. Generally older models are such underappreciated and forgotten despite them often being very competitive even when compared to the latest releases. Which old I mean models released in 2023 and 2024 and not the Bert garbage that somehow everyone (probably students doing their assignments) seems to upload.
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#NemoPlotter-12B-GGUF for quants to appear.