d1-2B
d1-2B is a 1.92B decision model for text, images and video.
What the release provides
It builds on LFM3-VL-2B. One forward pass supplies typed answers without text generation. Tasks include extracting spans, locating objects and estimating a quantity with an interval. The card lists a 128K context and the LFM Open License.d1-2B official model card
How to use it well
Practical guidance. Define the available outcomes before using the model. Validate decision thresholds on representative examples, including ambiguous inputs and missing fields.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | LiquidAI/d1-2B |