trained on a mixture of synthetic and natural RP data, as well as a mixture of storywriting/novel data from various sources (sugarquill, SCP, and miscellaneous novels) for 17-ish hours on 2x3090 from runpod
quants: https://huggingface.co/Hasnonname/Qwen2.5-Monte-7B-v0.0-GGUF
it's either overcooked or undercooked, and I can't tell which. regardless, thanks for giving it a shot.