Fine-tuned ALPaCA model checkpoints for slide-level whole-slide-image question answering in computational pathology.

Built with Llama.

Training, evaluation and demonstration code, GMM prototypes, example features and metadata are available in the separate ALPaCA GitHub repository. The repository README provides detailed installation, training, evaluation and single-WSI inference instructions.

The curated ALPaCA datasets are available from CNX-PathLLM on Hugging Face.

Users must separately obtain access to the meta-llama/Llama-3.1-8B-Instruct base model under Meta's applicable terms. ALPaCA and its checkpoints are intended for research use and are not approved for clinical diagnosis, clinical decision-making or use as a medical device.

These checkpoints are derivative Llama 3.1 materials and are distributed under the Llama 3.1 Community License. Access to, use of and redistribution of the checkpoints are subject to that licence and Meta's Acceptable Use Policy.

Llama 3.1 is licensed under the Llama 3.1 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.