QuantFactory

/

Ahma-3B-GGUF

like 2

Follow

Quant Factory 537

This PR improves the model card for the QuantFactory/Ahma-3B-GGUF model by:

• Adding library_name: llama.cpp to the metadata, which is appropriate for a GGUF artifact and enables the relevant "how to use" section.
• Removing the incorrect inference: false metadata tag.
• Adding gguf to the model tags for better discoverability.
• Prominently linking to the paper "Scaling Data-Constrained Language Models", which influenced the training of the base Ahma-3B model.
• Providing a clear usage example for the GGUF model using llama-cpp-python.
• Clarifying the existing transformers usage section to indicate it is for the original (non-GGUF) Finnish-NLP/Ahma-3B model.
• Adding a direct link to the llama.cpp GitHub repository in the introduction.

These changes help users understand how to best utilize this specific GGUF model and provides clearer context for its development.