4 ms·
This appears to be a Llava model which was then fine-tuned using outputs from Llama 3. If I understand correctly, that would make it Llama-2-based.
by CGamesPlay 2y ago
This appears to be a Llava model which was then fine-tuned using outputs from Llama 3. If I understand correctly, that would make it Llama-2-based.
- GaggiX 2y ago>fine-tuned using outputs from Llama 3. Llama 3 outputs text and can only see text, this is a vision model. >that would make it Llama-2-based. It's based on Llama 3, Llama 2 has nothing to do with it. They took Llama 3 Instruct and CLIP-ViT-Large-patch14-336, train the projection layer first and then later finetuned the Llama 3 checkpoint and train a LoRA for the ViT.