Add catalyst (#1)

* Start Catalyst branch

* Add Catalyst code

* Add Catalyst code

* minor fixes

* minor fixes

* minor fixes

* minor fixes

* Remove training from notebook

* Remove training from notebook

* license

* catalyst 19.12

* minor fixes

* minor

* train instruction

* minor proofreading

LICENSE

@@ -0,0 +1,15 @@
+Apache Software License 2.0
+
+Copyright (c) 2019, Neuromation.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

Makefile

@@ -1,7 +1,7 @@
 ##### PATHS #####
 
 DATA_DIR?=data
-CODE_DIR?=bone-age-code
+CODE_DIR?=src
 NOTEBOOKS_DIR?=notebooks
 RESULTS_DIR?=results
 

README.md

@@ -2,28 +2,14 @@
 
 # Description
 
-In this project, we introduce the problem of paediatric bone age assessment. During organism development, 
-the bones of the skeleton change in size and shape,
-and thus a difference between a child’s assigned bone and chronological ages
-might indicate a growth problem. Clinicians use bone 
-age assessment in order to estimate the maturity of a child’s skeletal system. Bone age assessment 
-usually starts with taking a single X-ray image of the left hand from the wrist to
-fingertips. Traditionally, bones in the radiograph are compared with images
-in a standardized atlas of bone development. This recipe represents a core approach 
-described in _"Paediatric Bone Age Assessment Using Deep Convolutional Neural Networks" by 
-V. Iglovikov, A. Rakhlin, A. Kalinin and A. Shvets_, [link 1](https://link.springer.com/chapter/10.1007%2F978-3-030-00889-5_34), 
-[2](https://www.biorxiv.org/content/biorxiv/early/2018/06/20/234120.full.pdf). 
-We validate the performance of the method using the data from
-the 2017 Pediatric Bone Age Challenge organized by the Radiological Society of
-North America (RSNA). The data set has been contributed by 3 medical centers at Stanford University, 
-the University of Colorado and the University of California - Los Angeles. Originally, the dataset was
-shared by the AIMI Center of Stanford University and now can be freely accessed 
-at [Kaggle platform](https://kaggle.com/kmader/rsna-bone-age). For the sake of simplicity we skip intense peprocessing
-steps as described in the original work and provide radiographs with already removed background 
-and uniformly registered hands.    
-
-![](./data/1381_original.png) 
-![](./data/1381_preprocessed.png)  
+In this project, we introduce the problem of pediatric bone age assessment. During organism development, the bones of the skeleton change in size and shape, and thus a difference between a child's assigned bone and chronological ages might indicate a growth problem. Clinicians use bone age assessment to estimate the maturity of a child's skeletal system. Bone age assessment usually starts with taking a single X-ray image of the left hand from the wrist to the fingertips. Traditionally, bones in the radiograph are compared with images in a standardized atlas of bone development. 
+
+This recipe represents a core approach described in _"Paediatric Bone Age Assessment Using Deep Convolutional Neural Networks" by V. Iglovikov, A. Rakhlin, A. Kalinin and A. Shvets_, [link 1](https://link.springer.com/chapter/10.1007%2F978-3-030-00889-5_34), [2](https://www.biorxiv.org/content/biorxiv/early/2018/06/20/234120.full.pdf). 
+
+We validate the performance of the method using the data from the 2017 Pediatric Bone Age Challenge organized by the Radiological Society of North America (RSNA). The data set has been contributed by 3 medical centers at Stanford University,  the University of Colorado and the University of California - Los Angeles. Initially, the dataset was shared by the AIMI Center of Stanford University and now can be freely accessed at [Kaggle platform](https://kaggle.com/kmader/rsna-bone-age). For the sake of simplicity, we skip intense preprocessing steps as described in the original work and provide radiographs with already removed background and uniformly registered hands.    
+
+![](./img/1381_original.png) 
+![](./img/1381_preprocessed.png)  
 *Original and preprocessed radiographs of a hand of 82 month old (approx. 7 y.o.) girl*
 
 # Quick Start
@@ -41,6 +27,11 @@ cd ml-recipe-bone-age
 make setup
 make jupyter
 ```
+##### 3. Train the model
+Download the dataset from within the demo notebook, then run:
+```shell
+make training
+```
 
 # Development Environment
 
@@ -51,7 +42,7 @@ This project is designed to run on [Neuro Platform](https://neu.ro), so you can
 | Local directory                      | Description       | Storage URI                                                                  | Environment mounting point |
 |:------------------------------------ |:----------------- |:---------------------------------------------------------------------------- |:-------------------------- | 
 | `data/`                              | Data              | `storage:ml-recipe-bone-age/data/`                              | `/ml-recipe-bone-age/data/` | 
-| `bone-age-code/` | Python modules    | `storage:ml-recipe-bone-age/bone-age-code/` | `/ml-recipe-bone-age/bone-age-code/` |
+| `src/` | Python modules    | `storage:ml-recipe-bone-age/src/` | `/ml-recipe-bone-age/src/` |
 | `notebooks/`                         | Jupyter notebooks | `storage:ml-recipe-bone-age/notebooks/`                         | `/ml-recipe-bone-age/notebooks/` |
 | No directory                         | Logs and results  | `storage:ml-recipe-bone-age/results/`                           | `/ml-recipe-bone-age/results/` |