successor of this model
#2
by
rinoa
- opened
Hi Openbuddy team, I found this deepseekcoder-based model has better performance on long context than mistral-based ones. Do you have any plans to update this model?
Yes, you can also take a look at our mixtral-18.1 model, which also has nice long text capabilities.
Yes, you can also take a look at our mixtral-18.1 model, which also has nice long text capabilities.
I have tried mixtral models and they also have good performance. The mixtral models, unfortunately, have a rather large size. On the other hand, the ~7b models are a much better match for my requirements.