So can you just not run the f16 llama 3.2 3B llamafile on windows?
#2
by
bermancheese
- opened
My computer says i cant and it seems so does the instructions, but im an idiot and dont trust that im right.
If you don't have enough VRAM in your GPU, try one of the quantized version, as q8 or q6.
F16 is not the easy option.
I was also unable to start the BF16 model with the file name Llama-3.2-3B-Instruct.BF16.llamafile.exe under Windows 11. However, I was able to start the smaller model Q6_K with the file name Llama-3.2-3B-Instruct.Q6_K.llamafile.exe. The chat in the browser was also possible.