Comment by skavi

can anyone comment on running the 2b quantized llama 70b on consumer cards like the 4090?