Search Results for "qa"

...Trained on a large bilingual dataset — including 30 million high-quality Chinese image-text pairs from CogView and 300 million English pairs — VisualGLM-6B is designed for image understanding, description, and question answering. Fine-tuning on long visual QA datasets further aligns the model’s responses with human preferences. The repository provides inference APIs, command-line demos, web demos, and efficient fine-tuning options like LoRA, QLoRA, and P-tuning. It also supports quantization down to INT4, enabling local deployment on consumer GPUs with as little as 6.3 GB VRAM.

LongBench is a comprehensive benchmark designed to evaluate the ability of large language models to understand and reason over very long textual contexts. Traditional language model benchmarks typically evaluate tasks involving relatively short inputs, which does not reflect many real-world applications such as analyzing large documents or entire code repositories. LongBench addresses this gap by providing datasets that require models to process and reason over long sequences of text across...

Ferret is Apple’s end-to-end multimodal large language model designed specifically for flexible referring and grounding: it can understand references of any granularity (boxes, points, free-form regions) and then ground open-vocabulary descriptions back onto the image. The core idea is a hybrid region representation that mixes discrete coordinates with continuous visual features, so the model can fluidly handle “any-form” referring while maintaining precise spatial localization. The repo...