Llava

Please keep all placeholder content unchanged, such as [IMAGE_0], [VIDEO_0], [IFRAME_0], , , etc. 2. Only translate the plain text content; do not translate or modify any placeholders. 3. Keep the original HTML formatting tags unchanged. 4. Return the translation result directly without adding any explanation. To be translated:

From models we deployed ourselves on cloud GPUs. Some models are open-source, while others are fine-tuned or developed by us.

Image recognition API, which identifies image content by uploading images, uses the llava-13b model. Source: https://github.com/haotian-liu/LLaVA

Average runtime: 5-10s

Price: 0.01 PTC/transaction