How to train if there are multiple sources?

pumpikano / tf-dann

Domain-Adversarial Neural Network in Tensorflow

MIT License

628 stars 224 forks source link

How to train if there are multiple sources? #25

Open kimheng8888 opened 6 years ago

kimheng8888 commented 6 years ago

Suppose we have 3 sources domain and 1 target domain. How should I train DANN in this case? Should I combine all the 3 sources into only 1 source or should I label those 3 domain separately during training?

erlendd commented 5 years ago

There's a paper for that: "Multi-adversarial domain adapation". https://pdfs.semanticscholar.org/ffda/dffca8bb57e59db10e5ee702b8f6ccab52cf.pdf