pepperfree@sh.itjust.works to LocalLLaMA@sh.itjust.worksEnglish · 3 months agoWhen DeepSeek V4 and R2?sh.itjust.worksexternal-linkmessage-square26fedilinkarrow-up1193arrow-down16
arrow-up1187arrow-down1external-linkWhen DeepSeek V4 and R2?sh.itjust.workspepperfree@sh.itjust.works to LocalLLaMA@sh.itjust.worksEnglish · 3 months agomessage-square26fedilink
minus-squarepepperfree@sh.itjust.worksOPlinkfedilinkEnglisharrow-up1arrow-down1·2 months agoLlama 3.3 was good, tho. For the multimodal, llama 4 also use llama3.2 approach where the image and text is made into single model instead using CLIP or siglip.
Llama 3.3 was good, tho. For the multimodal, llama 4 also use llama3.2 approach where the image and text is made into single model instead using CLIP or siglip.