A LLaMA2-7b chatbot with memory running on CPU, and optimized using smooth quantization, 4-bit quantization or Intel® Extension For PyTorch with bfloat16.
metacpuoptimizationchatbotintelllamanumaint8ipex4-bit-cpuhuggingfacestreamlitbfloat16neural-compressionchatgptlangchainllama2meta-aismooth-quantizationchatbot-memory
-
Updated
Feb 27, 2024 - Python