Re:RedPajama: Reproduction of Llama with Friendly License (S...
by nl on Monday April 17, 2023 @07:31PM
(#35608041)
I'm not aware of any efficient transformer training code for AMD cards.
Also most training is done using bfloat, not single precision (which is usually only used for accumulators)
Re:RedPajama: Reproduction of Llama with Friendly License (S...
by Robotbeat on Monday April 17, 2023 @09:42PM
(#35608973)
Sure, you would need to rewrite the training code for AMD's ecosystem. If you're using mixed precision training, I suppose you're right about BF16. That puts the relative performance of A100 about 2.5x that of the Radeon RX 7900 XT. May be better to go with the Nvidia GeForce RTX 4090 with a $1600 retail.
Re:RedPajama: Reproduction of Llama with Friendly License (S...
by titaniumtown on Tuesday April 18, 2023 @10:58AM
(#35614854)
It all works with pytorch and huggingface's transformers library out of the box with Rocm.
An order of magnitude lower GPU-hour time, plus if you train it for 210 days instead of 21 days, means you could do a 7B model with 20 consumer GPUs which are $1000 apiece. $20k, not counting mainboard, etc. Really not bad. Might even be doable as a volunteer project.