What is everyone's experiences so far with DPO trained versions of their favorite models? Been messing around with different models and my two new favorite models are actually just the DPO versions of my previous favorite models (causalLM 14b and openhermes 2.5 7b). Links below for the models in question.
CausalLM 14B-DPO-alpha - GGUF: https://huggingface.co/tastypear/CausalLM-14B-DPO-alpha-GGUF
NeuralHermes 2.5 Mistral 7B - GGUF: https://huggingface.co/TheBloke/NeuralHermes-2.5-Mistral-7B-GGUF
The former runs at 30 t/s for me with koboldcpp-rocm on a 6900 XT, and the latter at 15 t/s, both at Q6K. I don't have a favorite between these two models, they seem to be better at different things and trade blows in all the logic + creative writing tasks I've tested them in, despite causalLM being a larger model. I'm looking forward to seeing what nousresearch/teknium and CausalLM are bringing next.
I tried DPOpenHermes from TheBloke(Q6 GGUF version) and i love it but i think there's an issue with an EOS token as for some reason the model just keep generating text way past where it should logically stop. I see myself using it more but i hope there will be an update that adresses the EOS issue.
There is already an updated version that is supposed to fix that (with additional training on top which lowered it's overall capabilities apparently). I don't know if TheBloke has it already. But I see the first set of dpo models as test runs the next ones should fix the issues (except for NeuralHermes, maybe it's already good, I didn't hear much feedback about it).