Llama 3 Compatibility Support

[gauravsinghbitmesra-design]

Llama 3 Compatibility Support

Description

Refactors the model and tokenizer components to support Llama 3 specifications while maintaining backward compatibility with Llama 2.

Key Changes

• Configurable RoPE: Added rope_theta to ModelArgs to support Llama 3's 500k base.
• Tokenizer Support: Integrated tiktoken support in Tokenizer class to handle Llama 3 models, with automatic fallback to SentencePiece for Llama 2.
• Flexible Chat Templates: Refactored generation.py to use a ChatFormatter system, enabling support for Llama 3's special token structure (<|start_header_id|>, etc.).

Verification

• Added tests/test_llama3_compat.py to verify:
  • RoPE theta propagation.
  • Tokenizer backend selection (Tiktoken vs SentencePiece).
  • Correct chat template formatting for both Llama 2 and Llama 3 styles.
• Manual verification of instantiation and parameter loading.

[meta-cla]

Hi @gauravsinghbitmesra-design!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks!