Brief
OFASys Enables Multitask Learning with One Line of Code
A new system called OFASys aims to simplify multimodal multitask learning by reducing setup to a single line of code, addressing difficulties the team faced with their previous model OFA.
OFASys enables multitask learning with one line of code. The team previously released OFA, a unified multimodal pretrained model, but encountered difficulties in implementation. They found it hard to set up multiple tasks concerning multiple modalities and to organize multitask learning, such as batching data and making training stable. The new tool aims to address these difficulties by reducing the setup to a single line of code.
Source details and supporting facts
Each line is stated by the page named above it.
Stated by qwenlm.github.io
- The team previously released an opensourced unified multimodal pretrained model OFA.
- They met difficulties in their implementation, such as hard to set up multiple tasks concerning multiple modalities.
- They also found it hard to organize multitask learning, e.g., how to batchify data and how to make training stable.
Sources
- Qwen blog (Alibaba)Text stored 15 September 2026
How this story was checked. Written from the 1 page listed above, stored 15 September 2026; claims checked against that stored text on 15 September 2026.
What that means
- 3 of 4 reported statements were confirmed against the page that carries them; the rest were removed rather than published.
- Figures in the text were required to appear in the stored source text: yes. Identifiers: yes.
- The check reads stored text only: no claim rests on a fresh look that did not happen.
- Where the reporting was silent, the text says so instead of filling the gap.