@@ -79,7 +79,7 @@ If you want to use the pretrained models on APPLE_MOTS, download it [here][pretr
## Training
When using pretrained models, this step can be omitted.
To train a model, run `main.py` with a corresponding configuration file e.g. `python main.py configs/3dconv8_24` (`cd algorithms/TrackR-CNN`). Four models (3d_conv) use two 3D conv layers where the fifth model (lstm) uses two stacked lstm layers. The description behind the model corresponds to respectively the batch size and number of epochs. Note that the number of epochs in the configuration files are calculated by the total number of epochs / batch size. Because for every epoch all possible batches of e.g. 8 consecutive frames are processed.
To train a model, run `main.py` with a corresponding configuration file e.g. `python main.py configs/3dconv8_24` (`cd algorithms/TrackR-CNN`). Seven models (3d_conv) use two 3D conv layers where the model eight (lstm) uses two stacked lstm layers. The description behind the model corresponds to respectively the batch size and number of epochs. Note that the number of epochs in the configuration files are calculated by the total number of epochs / batch size. Because for every epoch all possible batches of e.g. 8 consecutive frames are processed.
