Florence 2 large

Name: Florence 2 large
Author: Microsoft

by MicrosoftFlorence

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released June 20244K context≈ 3,072 words

Florence 2 large is a focused, task-oriented vision-language model that handles a wide range of image understanding tasks — captioning, object detection, OCR, grounding, and segmentation — all from a single unified architecture. It takes plain text prompts to switch between tasks, making it straightforward to work with without task-specific fine-tuning. The 'large' variant trades some inference speed for stronger spatial reasoning and more detailed outputs compared to its smaller sibling.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Long Context

Moderate

Tool Use

Florence 2 large

by MicrosoftFlorence

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released June 20244K context≈ 3,072 words

Florence 2 large is a focused, task-oriented vision-language model that handles a wide range of image understanding tasks — captioning, object detection, OCR, grounding, and segmentation — all from a single unified architecture. It takes plain text prompts to switch between tasks, making it straightforward to work with without task-specific fine-tuning. The 'large' variant trades some inference speed for stronger spatial reasoning and more detailed outputs compared to its smaller sibling.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Long Context

Moderate

Tool Use

Glossary

ArchitectureThe underlying structural design of a neural network that defines how data flows through layers and components.Fine-TuningThe process of further training a pre-trained model on new data to adapt it for specific tasks or domains.GroundingThe practice of ensuring a model's responses are based on and supported by provided source documents rather than generated from general knowledge.InferenceThe process of running a trained model to generate predictions or outputs from new inputs.Inference SpeedHow quickly a model can generate predictions or outputs after being given an input, measured in time per token or tokens per second.Language ModelAn AI model trained to predict and generate text by learning patterns from large amounts of written data.Object DetectionA computer vision task that identifies and locates specific objects within an image by drawing boxes around them.ReasoningThe model's ability to work through multi-step logical problems and provide justified answers rather than just pattern-matching.SegmentationDividing an object or scene into distinct regions or parts.Spatial ReasoningThe ability to understand and reason about the location, size, and relationships between objects in an image.Task-Specific Fine-TuningAdapting a pre-trained model to excel at a particular task using domain-relevant training data.Unified ArchitectureA single model design that handles multiple different tasks without needing separate specialized models for each task.Vision-LanguageA model designed to understand and reason about both visual content (images) and natural language text together.Vision-Language ModelAn AI model that understands both images and text, allowing it to answer questions about images or describe what it sees.

Capabilities

Capabilities

Use Case Fit

Glossary