Datasets:
Which LLaMA model (file) should I choose?
@andreaskoepf Maybe you can assist me with another question. I'm a bit lost here.
I'm installing OpenAssistant, and I found this info during the installation process:
[To use OpenAssistant LLaMA-Based Models, you should have a copy of the original LLaMA model weights and add them to a llama subdirectory here. If you cannot obtain the original LLaMA, see the note in italic below for a possible alternative.]
Where do I find a copy of the original LLaMA model weights? What is that exactly?
Plus, on this link https://huggingface.co./OpenAssistant/oasst-sft-6-llama-30b-xor/tree/main/oasst-sft-6-llama-30b-xor/oasst-sft-6-llama-30b-xor you can find many files:
pytorch_model-00001-of-00007.bin - 9.82 GB
pytorch_model-00002-of-00007.bin - 9.96 GB
pytorch_model-00003-of-00007.bin - 9.9 GB
pytorch_model-00004-of-00007.bin - 9.87 GB
pytorch_model-00005-of-00007.bin - 9.87 GB
pytorch_model-00006-of-00007.bin - 9.96 GB
pytorch_model-00007-of-00007.bin - 5.69 GB
And all of them are from about 1 month ago.
Which one should I download?
Thank you.
All of them