Florence 2 Models by Microsoft Research
About
The Florence-2 family, created by Microsoft, features advanced large language models designed specifically for vision and vision-language tasks. These models are known for their ability to effectively address a variety of assignments, such as captioning, object detection, and segmentation, by employing a prompt-based methodology 2. Their unified representation is a significant advantage, allowing seamless task execution within a single model framework 3. Leveraging the extensive FLD-5B dataset, which offers 5.4 billion annotations across 126 million images, these models excel in multitask learning 2. The Florence-2 suite includes the Florence-2-base and Florence-2-large models, featuring parameter counts of 0.23 billion and 0.77 billion, respectively. Additionally, fine-tuned iterations like Florence-2-base-ft and Florence-2-large-ft demonstrate enhanced performance across various downstream tasks, while their compact size ensures they are efficient and suitable for resource-limited environments 3.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Florence 2 Base | Use when the workload needs 230M parameters. | 2024-06 | 230M parameters | Current |
| Florence 2 Large | Use when the workload needs 770M parameters. | 2024-06 | 770M parameters | Current |
Release Timeline
1 release groupSpecifications(2 models)
| Model | Released | Parameters |
|---|---|---|
| Florence 2 Base | 2024-06 | 230M |
| Florence 2 Large | 2024-06 | 770M |
Frequently Asked Questions
- What is Florence 2 used for?
- Florence 2 is used for coding. The family description and listed model capabilities point to those workloads as the best fit.
- How does Florence 2 compare to Harrier?
- Florence 2 by Microsoft Research is strongest where you need coding, while Harrier by Microsoft Research is the closest related family to check for embedding. Florence 2 has 2 listed variants, while Harrier reaches up to 33k context, so compare the specs and pricing tables before choosing a production model.
- Which Florence 2 model should I use?
- If price is the main constraint, use the pricing table first because Florence 2 does not have complete provider pricing in the local data. For the most capable/latest local choice, evaluate Florence 2 Base.
