3 ms·
30B would be interesting because that's the practical ceiling for local GPUs assuming 4-bit quantization. Is there some kind of dedicated fund for training har
by int_19h 3y ago
30B would be interesting because that's the practical ceiling for local GPUs assuming 4-bit quantization.
Is there some kind of dedicated fund for training hardware? Donating an A100 sounds unlikely, but surely they could be crowdfunded?
- pico_creator 3y agoweirdly enough, organisations are more willing to rent GPUs than money. If you want to help fund RWKV, the ko-fi link is - https://ko-fi.com/rwkv_lm https://ko-fi.com/rwkv_lm IMO: this needs way more funding, just to sustain blink leading this project, let alone GPUs for training. (Also - current tests shows this model doing really badly with 4bit quantized, but alright at Q5 and Q8)