total amount of training data

[yangcong1617]

The paper mentions the datasets used to train the model, including both synthetic and real data. Could you provide more details on the total amount of training data? For example, how much of the data is synthetic and how much is real?