Fine tuning LLMs for generation of mojo code

Setting up the environment

Follow the next steps to install Mojo:

In the Ubuntu terminal, install the Modular CLI: curl https://get.modular.com | sh - && \ modular auth xxxxxxxxxxxxxxxxxxxxxxxxxxx
*It was necessary to install python venv: sudo apt install python3.10-venv
Install the Mojo SDK: modular install mojo

After installing Mojo CLI and SDK, define the necessary environment variables in the .bashrc file:

export MODULAR_HOME="$HOME/.modular"

export PATH="$MODULAR_HOME/pkg/packages.modular.com_mojo/bin:$PATH"

VS Code Extensions:

Name		Name	Last commit message	Last commit date
Latest commit History 39 Commits
chatgpt_generated_examples		chatgpt_generated_examples
data		data
fine_tuned_models_generated_code		fine_tuned_models_generated_code
FinetuneCodeLlama7b.ipynb		FinetuneCodeLlama7b.ipynb
FinetuneCodeLlamaInstruct7b.ipynb		FinetuneCodeLlamaInstruct7b.ipynb
FinetuneCodeLlamaPython7b.ipynb		FinetuneCodeLlamaPython7b.ipynb
MetricsPlots.ipynb		MetricsPlots.ipynb
README.md		README.md
generate_json_code_completion.py		generate_json_code_completion.py
generate_json_infilling.py		generate_json_infilling.py
generate_json_instruct.py		generate_json_instruct.py
prompt_length.py		prompt_length.py
testing_examples.txt		testing_examples.txt