3 ms·
I suppose siglipv2 wasn't out yet when they were training this - I wonder if there will be an update to the multimodal models or pali-gemma which utilizes Sigli
by jerrygenser 2y ago
I suppose siglipv2 wasn't out yet when they were training this - I wonder if there will be an update to the multimodal models or pali-gemma which utilizes Siglip2. Aya Vision from Cohere utilized siglip 2 to great effect
- changtimwu 2y agoFor llama.cpp, this would involve another handcrafted Siglip2 NN process. Interesting—they just figured out how to handle Siglip1. https://github.com/ggml-org/llama.cpp/pull/12344/commits/6313b97a1ed98c6e1cc9c8958e949fa7fe28b1cf#diff-d2bb0434579582f245f0130106c116dd642894fd4929d0c4775fefc4310cd729R642 https://github.com/ggml-org/llama.cpp/pull/12344/commits/631...