7b model.
Any chance for a 7b model?.
Sure, I can make one. Probably release tomorrow.
Thank you, I'm rocking nous hermes 7b right now and it is crazy good at roleplay especially when you apply this LoRA https://huggingface.co/lemonilia/limarp-llama2
Really excited to see how this will perform.
Hi, just want to say sorry for the delay, I've been distracted with another project, but I am still working on this.
It's all good, don't worry about it. Whenever it happens it happens.
I finished the 7b model and it is uploading right now. Thanks for the patience. You actually inspired me to do a V2 with some modifications and longer context.
Awesome work. I just tried this for the first time last night (used TheBloke's GPTQ version), and this is insanely good. It has been generating easily the best RP content I've gotten short of GPT, and not short by much if any. Outperforms even the 30b and 70b models I've tested with so far. What is the EverythingLM? There's no model card for it. I was hoping for a v2 model with the larger context at 13b - is that what EverythingLM is? Thanks for all your hard work.
Glad you enjoyed the model. EverythingLM is different model trained on a brand new dataset (you can find it in my profile), and it will have 16k context. It is still in the works right now. I can also make AlpacaCielo2-13b-8k if there is demand for it.