add pkgs

examples/gptneox/gptq_convert.sh

@@ -0,0 +1,9 @@
+git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa.git GPTQ-for-LLaMa
+
+pip install -r ./GPTQ-for-LLaMa/requirements.txt
+
+CUDA_VISIBLE_DEVICES=0 python3 GPTQ-for-LLaMa/neox.py ./gptneox_model \
+wikitext2 \
+--wbits 4 \
+--groupsize 128 \
+--save_safetensors ./gptneox_model/gptneox-20b-4bit-gs128.safetensors