How to quantize these models?
do quantized models exist for these models?
how indeed! I am ready to do GPTQ and GGUF, but I don't understand why Pickle?
is this repo dead? aside from us no one is talking here.
It's new and people are talking about it online. I expected something like this model: https://huggingface.co./RWKV/rwkv-5-world-7b, but maybe they are working on it.
And how good is world model? And is that a chat model as well?
The: https://huggingface.co./RWKV/rwkv-5-world-7b : Is a test model, for the use with huggingface transformer
This repo: https://huggingface.co./RWKV/v5-Eagle-7B/tree/main , is the same model without the conversion process
For quantization, you can use the existing library projects
- rwkv.cpp : https://github.com/RWKV/rwkv.cpp
- rwkv pip : https://pypi.org/project/rwkv/
"And is that a chat model as well?"
I guess not
Nope, its not a meant to be an instruct tune model - we are training one now
is this repo dead? aside from us no one is talking here.
Well, I'm just listening, does that count as activity? Love the model btw