Generate multimodal answers from images using CLIP and Vicuna or LLaMA models.
npx skills add https://github.com/JKhyro/HERMES-AGENT --skill llava-jkhyro
Please help me install this Agent Skill. Skill: llava Source: https://github.com/JKhyro/HERMES-AGENT/tree/main/optional-skills/mlops/llava Command: npx skills add https://github.com/JKhyro/HERMES-AGENT --skill llava-jkhyro
LLaVA removes the gap between language models and visual input by letting you ask questions about images, generate descriptions, and hold multi-turn conversations grounded in what is seen.
Use the llava skill to analyze the attached image, describe what it shows, and answer any follow-up visual questions in plain language.
High-intent search queries and answers about installing and using this skill.
Visual question answering generates text answers by pairing visual inputs with language reasoning, while multimodal image chat maintains conversation context across iterative follow-up questions about the same image.
To build a visual question answering workflow, combine a CLIP vision encoder with a Vicuna or LLaMA-based language model to process image preprocessing inputs and generate grounded text descriptions for document understanding.
Yes, you can fine-tune a vision-language model on custom datasets using optional quantization and DeepSpeed support, enabling efficient instruction-tuning and adaptation for specialized multimodal tasks.
Multimodal image chat supports multi-turn conversations for scene interpretation by retaining context across follow-up questions, allowing iterative exploration of visual inputs and rich scene summaries.
Image captioning and scene understanding require a CLIP vision encoder, a Vicuna or LLaMA-based language model, and image preprocessing to translate visual inputs into concise captions or rich summaries.
ovachieverEncode images and texts to compute cosine similarity for zero-shot classification and cross-modal retrieval.
Devsoul2026Enable multimodal vision-language conversations using CLIP and Vicuna/LLaMA models.
sheawinklerEnable multimodal vision-language reasoning and conversation with images.
RheasilviaAnswer visual questions and caption images using vision-language models.
