MADA
MADA copied to clipboard
Experiment related information
Hi,
- Could you please share the experiment details of table 3 in the paper( 31 classes--> 25 classes), which classes you used in the target domain and how did you choose them.
- In the office dataset, we do not have a validation set, so which accuracy you consider as final accuracy (i.e. do you consider the highest accuracy in one run or last accuracy of given the epoch). I will be thankful if you could share this information. Thanks
@zhuhan1236 @longmingsheng @caoyue10 Could you please share the information about your experiments in the paper?