Is there a way to run llama2 in the new repo?

[dnatarajan00]

I understand this repo has been deprecated. I would like to run llama2 in the new location (https://github.com/meta-llama/llama-models) but am unable to get it working. Are there instructions for how to run llama2 in the new location?

Please see this associated issue here: meta-llama/llama-models#247

[DangNhutNguyen]

Running Llama 2 from meta-llama/llama-models

1. Clone the repository:

   git clone https://github.com/meta-llama/llama-models.git
   cd llama-models

2. Install dependencies:

   pip install -r requirements.txt

3. Download model weights (follow repo instructions).

4. Run the model with a script like this:

   from transformers import LlamaForCausalLM, LlamaTokenizer
   
   model_name = "llama-2"  # or local model path
   
   tokenizer = LlamaTokenizer.from_pretrained(model_name)
   model = LlamaForCausalLM.from_pretrained(model_name)
   
   inputs = tokenizer("Hello, world!", return_tensors="pt")
   outputs = model.generate(inputs['input_ids'])
   print(tokenizer.decode(outputs[0], skip_special_tokens=True))

5. Troubleshooting: Check the repo's [Issues](https://github.com/meta-llama/llama-models/issues) for solutions to common problems.