Anyone who should be attributed as part of the model. If multiple people or companies, use a comma seperated list
Example:
Firstname1 LastName1, Firstname2 Lastname2, Affiliation1
What tags describe the model and task performed? Use a comma seperated list
Example:
Segmentation, MR, Spleen, Self-Supervised
This section should describe general information about the model and task that it performs. Any high-level reference of what the model is doing, general architecture of the model, and brief mention of data used. A developer should be able to understand the purpose of the model from this section alone.
Example: This model is trained using the UNet architecture [1] and is used for volumetric 3D segmentation of the spleen from CT image. The segmentation of spleen region is formulated as the voxel-wise binary classification. Each voxel is predicted as either foreground (spleen) or background. And the model is optimized with gradient descent method minimizing soft dice loss between the predicted mask and ground truth segmentation.
This section should talk about the data. Where is the dataset from? How many images? What is the split for training, testing, and validation? Can the users get the data and where? Also, if you had to prepare the data in any particular way to make it work with the model, what steps or preprocessing steps (not transform steps) did you need to take to make it work.
Example:
The DETR model was trained on COCO 2017 panoptic, a dataset consisting of 118k/5k annotated images for training/validation respectively.
Images are resized/rescaled such that the shortest side is at least 800 pixels and the largest side at most 1333 pixels, and normalized across the RGB channels with the ImageNet mean (0.485, 0.456, 0.406) and standard deviation (0.229, 0.224, 0.225).
Any additional sections can be added with h4 (####) and bolded header to make sure it's easily navigable.
What sort of training or evaluation performance should people expect from this model? How long did it take you to train the model? If you have training/evaluation charts, you can include them in this section, but aren't required.
Example:
This model achieves the following results on COCO 2017 validation: a box AP (average precision) of 38.8, a segmentation AP (average precision) of 31.1 and a PQ (panoptic quality) of 43.4.
For more details regarding evaluation results, we refer to table 5 of the original paper.
If your bundle requires steps outside the normal flow of usage, describe those here in bash style commands.
Example:
My first special instruction for training:
$ python -m monai.bundle <my special instructions>
My special instruction for inference 2:
$ python -m monai.bundle <my special instructions 2>
What kind of system is required for training? General specs of CPU/GPU/RAM requirements to train and infer the model? How long did it take you to train the model?
Example:
The model was trained for 300 epochs on 16 V100 GPUs. This takes 3 days, with 4 images per GPU (hence a total batch size of 64).
Are there general limitations of what this model should be used for? Has this been approved for use in any clinicial systems? Are there any things to watch out for when using this model?
Example: This training and inference pipeline was developed by NVIDIA. It is based on a segmentation model created by NVIDIA researchers. This research use only software that has not been cleared or approved by FDA or any regulatory agency. Clara’s pre-trained models are for developmental purposes only and cannot be used directly for clinical procedures.
If people need to cite your model, how should they cite it?
Example:
@article{radford2019language,
title={Language Models are Unsupervised Multitask Learners},
author={Radford, Alec and Wu, Jeff and Child, Rewon and Luan, David and Amodei, Dario and Sutskever, Ilya},
year={2019}
}
This section should talk about the different references, links, papers, or anything else related to the model not covered above.
Example:
[1] Sakinis, Tomas, et al. "Interactive segmentation of medical images through fully convolutional neural networks." arXiv preprint arXiv:1903.08205 (2019).