5090 5161 minor doc/env fixes (#5162)

Fixes #5161 
part of #5090

### Types of changes
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [ ] New tests added to cover the changes.
- [ ] Integration tests passed locally by running `./runtests.sh -f -u
--net --coverage`.
- [x] Quick tests passed locally by running `./runtests.sh --quick
--unittests --disttests`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated, tested `make html` command in the `docs/`
folder.

Signed-off-by: Wenqi Li <wenqil@nvidia.com>

docs/requirements.txt

@@ -20,7 +20,7 @@ sphinxcontrib-serializinghtml
 sphinx-autodoc-typehints==1.11.1
 pandas
 einops
-transformers
+transformers<4.22  # https://github.com/Project-MONAI/MONAI/issues/5157
 mlflow
 tensorboardX
 imagecodecs; platform_system == "Linux"

docs/source/index.rst

@@ -23,8 +23,6 @@ Its ambitions are:
 Features
 --------
 
-*The codebase is currently under active development*
-
 - flexible pre-processing for multi-dimensional medical imaging data;
 - compositional & portable APIs for ease of integration in existing workflows;
 - domain-specific implementations for networks, losses, evaluation metrics and more;
@@ -72,6 +70,13 @@ Technical documentation is available at `docs.monai.io <https://docs.monai.io>`_
 
   bundle_intro
 
+Model Zoo
+---------
+
+`The MONAI Model Zoo <https://github.com/Project-MONAI/model-zoo>`_ is a place for researchers and data scientists to share the latest and great models from the community.
+Utilizing `the MONAI Bundle format <https://docs.monai.io/en/latest/bundle_intro.html>`_ makes it easy to get started building workflows with MONAI.
+
+
 Links
 -----
 

docs/source/modules.md

@@ -30,7 +30,7 @@ are handled with specific protocols, and the data arrays are often high-dimensio
 [`monai.data`](https://github.com/Project-MONAI/MONAI/tree/dev/monai/data) modules include a set of domain-specific APIs
 for various deep learning applications:
 
-### Transforms with data in 'array' and 'dictionary' styles
+### Transforms with data in array and dictionary styles
 
 ![3d transform examples](../images/affine.png)
 
@@ -83,7 +83,7 @@ domain-specific usability and pipeline performance.
 ### Cache IO and transforms data to accelerate training
 
 Data-driven methods require many (potentially thousands of) epochs of training data reading and preprocessing. MONAI
-provides multi-threaded cache-based datasets to accelerate the process. [[Datasets experiment]](https://github.com/Project-MONAI/tutorials/blob/master/acceleration/dataset_type_performance.ipynb). The
+provides multi-threaded cache-based datasets to accelerate the process [[Datasets experiment]](https://github.com/Project-MONAI/tutorials/blob/master/acceleration/dataset_type_performance.ipynb). The
 cache can be persistent and dynamic (`SmartCacheDataset`) and reused across different experiments [[SmartCache example]](https://github.com/Project-MONAI/tutorials/blob/master/acceleration/distributed_training/unet_training_smartcache.py).
 The following figure illustrates the training speedup compared with a regular PyTorch program.
 
@@ -102,7 +102,7 @@ a `ThreadDataLoader` example is within the [Spleen fast training tutorial](https
 ### Public datasets
 
 To quickly get started with popular training data, MONAI provides several ready-to-integrate Dataset classes
-(such as `MedNISTDataset`, `DecathlonDataset`), which include data downloading, and support training/evaluation splits generation with transforms.
+(such as `MedNISTDataset`, `DecathlonDataset`, [`TciaDataset`](https://github.com/Project-MONAI/tutorials/blob/main/modules/tcia_dataset.ipynb)), which include data downloading, and support training/evaluation splits generation with transforms.
 [[Public datasets tutorial]](https://github.com/Project-MONAI/tutorials/blob/master/modules/public_datasets.ipynb)
 The common workflow of predefined datasets:
 
@@ -231,8 +231,8 @@ A typical process of `decollate batch` is illustrated as follows (with a `batch_
 
 Except for the pytorch-ignite based `monai.engines`, most of the MONAI modules could be used independently or combined
 with other software packages. For example, MONAI can be easily integrated into popular frameworks such as
-PyTorch-Lightning and Catalyst: [Lightning segmentation](https://github.com/Project-MONAI/tutorials/blob/master/3d_segmentation/spleen_segmentation_3d_lightning.ipynb),
-[Catalyst segmentation](https://github.com/Project-MONAI/tutorials/blob/master/3d_segmentation/unet_segmentation_3d_catalyst.ipynb).
+PyTorch-Lightning and Catalyst. [[Lightning segmentation](https://github.com/Project-MONAI/tutorials/blob/master/3d_segmentation/spleen_segmentation_3d_lightning.ipynb),
+[Catalyst segmentation](https://github.com/Project-MONAI/tutorials/blob/master/3d_segmentation/unet_segmentation_3d_catalyst.ipynb)]
 
 ## Bundle
 
@@ -264,7 +264,7 @@ A typical bundle example can include:
      ┗━ *license.txt
 ```
 Details about the bundle config definition and syntax & examples are at [config syntax](https://docs.monai.io/en/latest/config_syntax.html).
-A step-by-step [get started](https://github.com/Project-MONAI/tutorials/blob/master/modules/bundles/get_started.ipynb) tutorial notebook can help users quickly set up a bundle. [[bundle examples]](https://github.com/Project-MONAI/tutorials/tree/main/modules/bundle)
+A step-by-step [get started](https://github.com/Project-MONAI/tutorials/blob/master/modules/bundles/get_started.ipynb) tutorial notebook can help users quickly set up a bundle. [[bundle examples](https://github.com/Project-MONAI/tutorials/tree/main/bundle), [model-zoo](https://github.com/Project-MONAI/model-zoo)]
 
 ## Federated Learning
 
@@ -288,7 +288,7 @@ with [`ClientAlgo`](https://docs.monai.io/en/latest/fl.html#clientalgo) to allow
 It leverages the latest advances in MONAI
 and GPUs to efficiently develop and deploy algorithms with state-of-the-art performance.
 It first analyzes the global information such as intensity, dimensionality, and resolution of the dataset,
-then generates algorithms in MONAI bundle format based on data statistics and algorithm templates.
+then generates algorithms in MONAI bundle format based on data statistics and [algorithm templates](https://github.com/Project-MONAI/research-contributions/tree/main/auto3dseg).
 Next, all algorithms initiate model training to obtain checkpoints with the best validation performance.
 Finally, the ensemble module selects the algorithms via ranking trained checkpoints and creates ensemble predictions.
 

docs/source/whatsnew_1_0.md

@@ -21,7 +21,7 @@ For more details about how to use the models, please see [the tutorials](https:/
 It leverages the latest advances in MONAI
 and GPUs to efficiently develop and deploy algorithms with state-of-the-art performance.
 It first analyzes the global information such as intensity, dimensionality, and resolution of the dataset,
-then generates algorithms in MONAI bundle format based on data statistics and algorithm templates.
+then generates algorithms in MONAI bundle format based on data statistics and [algorithm templates](https://github.com/Project-MONAI/research-contributions/tree/main/auto3dseg).
 Next, all algorithms initiate model training to obtain checkpoints with the best validation performance.
 Finally, the ensemble module selects the algorithms via ranking trained checkpoints and creates ensemble predictions.
 
@@ -42,15 +42,14 @@ collaborative learning in medical imaging.
 ## MetaTensor Support for Digital Pathology Workflows
 ![pathology](../images/pathology-meta.png)
 
-MetaTensor is officially released in MONAI v0.9, which is a simple yet elegant way to handle metadata along with the image
-in the same object. In this release (v1.0), we support MetaTensor in all digital pathology components, and make sure that
-the future development can benefit from them. With the help of MONAI Pathology Working Group, we have standardized a
-set of metadata attributes for patches of images extracted from WSI to ensure reproducibility and enhance functionality
-via relying on a standard set of attributes. The figure below shows all the pathology metadata attributes,
-their definition, and their relation to MetaTensors. Using `LoadImage` transform with WSIReader will output a
-MetaTensor with populated metadata inferred from the data file. Currently, WSIReader loads the patches into the CPU memory
-but they can be transferred to GPU via appropriate transforms. We are working with the cuCIM team to make sure that we can
-bring direct loading of images into the GPU in the future releases.
+In this release, we support MetaTensor in all digital pathology components, and
+make sure that the future development can benefit from them. With the help of
+MONAI Pathology Working Group, we have standardized a set of metadata
+attributes for patches of images extracted from WSI to ensure reproducibility
+and enhance functionality via relying on a standard set of attributes. The
+figure above shows all the pathology metadata attributes and their relation to
+MetaTensors. Please see [the tutorials and
+examples](https://github.com/Project-MONAI/tutorials/tree/main/pathology).
 
 ## Accelerated MRI Reconstruction
 ![MRI-reconstruction](../images/mri_recon.png)

environment-dev.yml

@@ -38,7 +38,7 @@ dependencies:
   - pandas
   - requests
   - einops
-  - transformers
+  - transformers<4.22  # https://github.com/Project-MONAI/MONAI/issues/5157
   - mlflow
   - tensorboardX
   - pyyaml
@@ -47,7 +47,6 @@ dependencies:
   - pynrrd
   - pydicom
   - h5py
-  - nni
   - optuna
   - pip
   - pip:
@@ -65,3 +64,4 @@ dependencies:
     - imagecodecs; platform_system == "Linux"
     - tifffile; platform_system == "Linux"
     - matplotlib!=3.5.0
+    - nni

monai/apps/auto3dseg/auto_runner.py

@@ -381,7 +381,7 @@ def set_prediction_params(self, params: Optional[Dict[str, Any]] = None):
     def set_hpo_params(self, params: Optional[Dict[str, Any]] = None):
         """
         Set parameters for the HPO module and the algos before the training. It will attempt to (1) override bundle
-        templates with the key-value pairs in ``params`` (2) chagne the config of the HPO module (e.g. NNI) if the
+        templates with the key-value pairs in ``params`` (2) change the config of the HPO module (e.g. NNI) if the
         key is found to be one of:
 
             - "trialCodeDirectory"
@@ -394,7 +394,7 @@ def set_hpo_params(self, params: Optional[Dict[str, Any]] = None):
 
         Args:
             params: a dict that defines the overriding key-value pairs during instantiation of the algo. For
-                BundleAlgo, it will overide the template config filling.
+                BundleAlgo, it will override the template config filling.
         """
         if params is None:
             self.hpo_params = self.train_params
@@ -406,7 +406,7 @@ def set_nni_search_space(self, search_space):
         Set the search space for NNI parameter search.
 
         Args:
-            search_space: hyper paramter search space in the form of dict. For more information, please check
+            search_space: hyper parameter search space in the form of dict. For more information, please check
                 NNI documentation: https://nni.readthedocs.io/en/v2.2/Tutorial/SearchSpaceSpec.html .
         """
         value_combinations = 1
@@ -471,7 +471,7 @@ def set_ensemble_method(self, ensemble_method_name: str = "AlgoEnsembleBestN", *
 
     def _train_algo_in_sequence(self, history: List[Dict[str, Any]]):
         """
-        Train the Algos in a seqential scheme. The order of training is randomized.
+        Train the Algos in a sequential scheme. The order of training is randomized.
 
         Args:
             history: the history of generated Algos. It is a list of dicts. Each element has the task name

monai/apps/auto3dseg/bundle_gen.py

@@ -75,7 +75,7 @@ def __init__(self, template_path: str):
 
     def set_data_stats(self, data_stats_files: str):  # type: ignore
         """
-        Set the data anlysis report (generated by DataAnalyzer).
+        Set the data analysis report (generated by DataAnalyzer).
 
         Args:
             data_stats_files: path to the datastats yaml file
@@ -253,8 +253,8 @@ def infer(self, image_file):
     def predict(self, predict_params=None):
         """
         Use the trained model to predict the outputs with a given input image. Path to input image is in the params
-        dict in a form of {"files", ["path_to_image_1", "path_to_image_2"]}. If it is not specified, then the pre-
-        diction will use the test images predefined in the bundle config.
+        dict in a form of {"files", ["path_to_image_1", "path_to_image_2"]}. If it is not specified, then the
+        prediction will use the test images predefined in the bundle config.
 
         Args:
             predict_params: a dict to override the parameters in the bundle config (including the files to predict).
@@ -295,7 +295,9 @@ class BundleGen(AlgoGen):
     Args:
         algo_path: the directory path to save the algorithm templates. Default is the current working dir.
         algos: if dictionary, it outlines the algorithm to use. if None, automatically download the zip file
-            from the defatult link. if string, it represents the download link.
+            from the default link. if string, it represents the download link.
+            The current default options are released at:
+            https://github.com/Project-MONAI/research-contributions/tree/main/auto3dseg
         data_stats_filename: the path to the data stats file (generated by DataAnalyzer)
         data_src_cfg_name: the path to the data source config YAML file. The config will be in a form of
             {"modality": "ct", "datalist": "path_to_json_datalist", "dataroot": "path_dir_data"}
@@ -332,7 +334,7 @@ def __init__(self, algo_path: str = ".", algos=None, data_stats_filename=None, d
                             ├── configs
                             │   ├── hyperparameters.yaml  # automatically generated yaml from a set of ``template_configs``
                             │   ├── network.yaml  # automatically generated network yaml from a set of ``template_configs``
-                            │   ├── transforms_train.yaml  # automatically generated yaml to define tranforms for training
+                            │   ├── transforms_train.yaml  # automatically generated yaml to define transforms for training
                             │   ├── transforms_validate.yaml  # automatically generated yaml to define transforms for validation
                             │   └── transforms_infer.yaml  # automatically generated yaml to define transforms for inference
                             └── scripts

monai/apps/auto3dseg/ensemble_builder.py

@@ -87,7 +87,7 @@ def ensemble_pred(self, preds, sigmoid=True):
         ensemble the results using either "mean" or "vote" method
 
         Args:
-            preds: a list of probablity prediction in Tensor-Like format.
+            preds: a list of probability prediction in Tensor-Like format.
             sigmoid: use the sigmoid function to threshold probability one-hot map.
 
         Returns:
@@ -205,7 +205,7 @@ class AlgoEnsembleBestByFold(AlgoEnsemble):
     Ensemble method that select the best models that are the tops in each fold.
 
     Args:
-        n_fold: number of cross-valiation folds used in training
+        n_fold: number of cross-validation folds used in training
     """
 
     def __init__(self, n_fold: int = 5):
@@ -291,7 +291,7 @@ def add_inferer(self, identifier: str, gen_algo: BundleAlgo, best_metric: Option
         """
 
         if best_metric is None:
-            raise ValueError("Feature to re-valiate is to be implemented")
+            raise ValueError("Feature to re-validate is to be implemented")
 
         algo = {AlgoEnsembleKeys.ID: identifier, AlgoEnsembleKeys.ALGO: gen_algo, AlgoEnsembleKeys.SCORE: best_metric}
         self.infer_algos.append(algo)

monai/apps/auto3dseg/utils.py

@@ -60,7 +60,7 @@ def export_bundle_algo_history(history: List[Dict[str, BundleAlgo]]):
     Save all the BundleAlgo in the history to algo_object.pkl in each individual folder
 
     Args:
-        history: a List of Bundle. Typicall the history can be obtained from BundleGen get_history method
+        history: a List of Bundle. Typically, the history can be obtained from BundleGen get_history method
     """
     for task in history:
         for _, algo in task.items():

setup.cfg

@@ -46,7 +46,7 @@ all =
     imagecodecs
     pandas
     einops
-    transformers
+    transformers<4.22
     mlflow
     matplotlib
     tensorboardX
@@ -93,7 +93,7 @@ pandas =
 einops =
     einops
 transformers =
-    transformers
+    transformers<4.22
 mlflow =
     mlflow
 matplotlib =