Chenzhan Shang
Results
2
issues of
Chenzhan Shang
v13, page 48, table 10 caption > For quantization, we employ bitesandbytes to quantize the 16-bit models to 8/4 bits might be: > For quantization, we employ bitsandbytes to quantize...
page 46, bottom right part > For example, a 4-bit 60GB LLM is demonstrated to have better performance than a 8-bit 30GB LLM [422]. which might be?: > For example,...