Closed Radium98 closed 2 years ago
All models used in the paper were trained using --network_capacity 16 due to the hardware limitation. You can try increasing this number for a better quality if your dataset is large enough. The training was on a single GPU. Typically it takes 3-5 days to show reasonable results, but it varies based on the target domain and the dataset size.
Can you tell me your training configuration?Such how many GPUs you need for training and how long it takes for training?