Training on Cuda version 11.2, 11.3

BlinkDL / RWKV-LM

RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.

Apache License 2.0

12.48k stars 848 forks source link

Training on Cuda version 11.2, 11.3 #194

Open cuongnguyengit opened 11 months ago

cuongnguyengit commented 11 months ago

Hi,

Please support me how to train rwkv model (rwkv4neo) in cuda version 11.2 or 11.3. I can't install 11.7 or newer. Thanks.

BlinkDL commented 11 months ago

what's the error you get