llama package issue, llama module not found

[Salman-Malik1]

when i run "torchrun --nproc_per_node 1 /opt/Meta-Llama-3-8B/example_text_completion.py --ckpt_dir /opt/Meta-Llama-3-8B/ --tokenizer_path /opt/Meta-Llama-3-8B/tokenizer.model" i got an error : Traceback (most recent call last): File "/opt/Meta-Llama-3-8B/example_text_completion.py", line 7, in from llama import Llama ModuleNotFoundError: No module named 'llama' but i have llama in my python-pip list

[mreso]

Hi, can you verify that you're using the same venv/conda env when running torchrun that you used to run pip install -e .?

[Salman-Malik1]

Hi, can you verify that you're using the same venv/conda env when running torchrun that you used to run pip install -e .?

i am using venv but i got same error, can you please help me how can i verify that i am using the same env ? so i can post my result here.

[Salman-Malik1]

Hi, can you verify that you're using the same venv/conda env when running torchrun that you used to run pip install -e .?

i tried in both venv python2.7 and 3.9 but same error.

[mreso]

Can you post the outputs of:

python -m pip list

and

torchrun --nnodes 1  -m pip list

[Salman-Malik1]

python -m pip list

torchrun --nnodes 1 -m pip list

[mreso]

What jumps out to me is that you've packages llama and llama3 installed. And llama points to a version 3 folder, though the name llama is connected to llama 2 for pip installations. Easiest would be to try a new env and only install llama once either using pip install . or pip install git+https://github.com/meta-llama/llama3.

[Salman-Malik1]

What jumps out to me is that you've packages llama and llama3 installed. And llama points to a version 3 folder, though the name llama is connected to llama 2 for pip installations. Easiest would be to try a new env and only install llama once either using pip install . or pip install git+https://github.com/meta-llama/llama3.

when i run this command it install both itself llama and llama3.

[Salman-Malik1]

What jumps out to me is that you've packages llama and llama3 installed. And llama points to a version 3 folder, though the name llama is connected to llama 2 for pip installations. Easiest would be to try a new env and only install llama once either using pip install . or pip install git+https://github.com/meta-llama/llama3.

getting same error :(

[Salman-Malik1]

What jumps out to me is that you've packages llama and llama3 installed. And llama points to a version 3 folder, though the name llama is connected to llama 2 for pip installations. Easiest would be to try a new env and only install llama once either using pip install . or pip install git+https://github.com/meta-llama/llama3.

I am currently facing a critical issue that requires immediate attention and I believe your expertise would be invaluable in resolving it. Given the urgency, I am prepared to provide you with all necessary access credentials securely and I'm willing to compensate you for your prompt and professional service. it is impacting our operations significantly.

Please let me know your availability at your earliest convenience, and your terms for such urgent tasks. You can reach me directly at this email (salman.malik@onboardsoft.com).

[magdalenapasternak]

I have the same issue as the author. Did anyone find a solution?