| DAD public-data experiment models | |
| These fine-tuned models are based on Qwen3-VL-2B-Instruct by the Qwen Team: | |
| https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct | |
| The base model is licensed under the Apache License, Version 2.0. | |
| A copy is included at licenses/Apache-2.0.txt. | |
| The released variants incorporate supervised fine-tuning and, for GRPO and | |
| EleRPO, reinforcement-learning updates. Adapter weights have been merged into | |
| the base model to provide standalone inference checkpoints. | |
| The license for these fine-tuned releases is CC BY-NC 4.0, included in LICENSE. | |
| This does not replace the license of the original, unmodified Qwen base model. | |