galai
galai copied to clipboard
Fine-tuning
Any recommendations / advice on how we would go about fine-tuning this model?
Friends at HuggingFace working on an integration, think this will be the best route for this.
Hi @TheodoreGalanos, the checkpoints on HuggingFace hub are now fully compatible with the OPT architecture, so as mentioned by Ross you should be able to use them to fine-tune a model.