Use new LLaMA implementation (this will break stuff. I am sorry)

https://github.com/oobabooga/text-generation-webui/wiki/LLaMA-model

requirements.txt

@@ -6,4 +6,4 @@ numpy
 rwkv==0.1.0
 safetensors==0.2.8
 sentencepiece
-git+https://github.com/oobabooga/transformers@llama_push
+git+https://github.com/zphang/transformers@llama_push