Only getting "G"
I setup the Conda environment and even reinstalled the model but all I get is "G" as an output?
Intresting, I've never seen that. What device are you working with ?
Intresting, I've never seen that. What device are you working with ?
Arch Linux/Zen kernal, no supported GPU (not that I think it works with one yet?) My CPU is an i5-2400
Do you need any other information?
I am also facing this issue in my ubuntu system, cpu i5 11 gen
Facing the same issue on a Debian 12.10 system (very old hp pavilion laptop, intel dual-core cpu, 4Gb ram ddr2, just testing how low-specs we can go).
I am attaching a script I used for setup.
P.S: I am not able to reproduce it on Debian under WSL on a modern system
It seems that the CPU should support avx2 instructions. My PC is I5-3570, which does not support avx2. I installed it several times and had the same problem. I created the corresponding docker image and installed it on my office's I5-10500, but GGGG was output in the same way. I recompiled bitnet.cpp and it worked fine. Note that lm studio will not work if avx2 is not supported either.