Update README.md

README.md

@@ -2,37 +2,93 @@
 ![Lint & test](https://github.com/dblasko/low-light-event-img-enhancer/actions/workflows/continuous-integration.yml/badge.svg)
 [![HitCount](https://hits.dwyl.com/dblasko/low-light-event-img-enhancer.svg)](https://hits.dwyl.com/dblasko/low-light-event-img-enhancer)
 
-# low-light-event-img-enhancer
+# low-light-event-img-enhancer <!-- omit from toc -->
 Deep-learning-based low-light image enhancer specialized on restoring dark images from events (concerts, parties, clubs...).
 
+## Table of contents: <!-- omit from toc -->
+- [Project structure](#project-structure)
+  - [Structure of the codebase](#structure-of-the-codebase)
+  - [Structure of the MIRNet model](#structure-of-the-mirnet-model)
+- [General usage of the project](#general-usage-of-the-project)
+  - [Requirements](#requirements)
+  - [Training the model](#training-the-model)
+    - [Pre-training](#pre-training)
+    - [Fine-tuning](#fine-tuning)
+  - [Running the model for inference](#running-the-model-for-inference)
+    - [Inference on a directory of images](#inference-on-a-directory-of-images)
+    - [Inference on a single image](#inference-on-a-single-image)
+  - [Generating the datasets](#generating-the-datasets)
+    - [Pre-training dataset](#pre-training-dataset)
+    - [Fine-tuning dataset](#fine-tuning-dataset)
+  - [Running tests](#running-tests)
+- [Usage of the web-application based on the model](#usage-of-the-web-application-based-on-the-model)
+  - [Running the inference endpoint](#running-the-inference-endpoint)
+  - [Running the web application](#running-the-web-application)
+
+
+
 ## Applied Deep Learning course
-The deliverables for *Assignment 1 - Initiate* are located in `/initiate.md`. Those for *Assignment 2 - Hacking* are located in `/hacking.md`.
+The deliverables for *Assignment 1 - Initiate* are located in `milestones/initiate.md`. As for *Assignment 2 - Hacking*, the complementary information to this README that describes general usage is in `milestones/hacking.md`.
+
+# Project structure
+
+## Structure of the codebase
+***TODO: description + diagramme structure projet/code, avec "put the data/model weights here" sur le schéma structure. Description dire pour rajouter modèle ajouter ici, pour ajouter dataset ajouter ici + run script as described in (quote section). Ajouter vos configs ici...**
 
-# Usage
+## Structure of the MIRNet model
+The [model](https://github.com/dblasko/low-light-event-img-enhancer/blob/main/model/MIRNet/model.py) is composed of multiple successive [residual recurrent groups](https://github.com/dblasko/low-light-event-img-enhancer/blob/main/model/MIRNet/ResidualRecurrentGroup.py), preceded and followed by single convolutional layers and a residual connection between the input and the output. The model in-between the residual connection generates the values to be added to the original image to enlighten it. Each residual recurrent group has a similar structure: a convolutional layer, followed by multiple [multi-scale residual blocks](https://github.com/dblasko/low-light-event-img-enhancer/blob/main/model/MIRNet/MultiScaleResidualBlock.py), and another convolutional layer followed by a residual connection to the input of the block. The multi-scale residual blocks have the following structure:
+![Multi-scale residual block overview](utils/documentation_images/mirnet.jpg)
+
+# General usage of the project
 ## Requirements
-In a Python 3.10 environment, install the requirements with `pip install -r requirements.txt`.
+In a **Python 3.10** environment, install the requirements using `pip install -r requirements.txt`.
 
 ## Training the model
+
 Training is generally done with the `training/train.py` script. Before running it, the training run must be configured in `training/config.yaml`. The configuration file contains the following parameters:
 ***TODO***
+**TODO: specify also WANDB key, show example of tracking, describe config file necessary & optional params° etc.**
+
+**TODO: general usage of script, from scratch BUT ALSO explain supports checkpointing & resuming a training run!**
 
 Then, you can run the training script while pointing to your configuration file with `python training/train.py --config training/config.yaml`.
 ***TODO: document will be tracked in WANDB...***
 
 ### Pre-training
-***TODO: document pre-training pass & commands/configs used, alternatively download weights here***
+***TODO: Explain how to start from scratch, used LoL dataset generated in [refer to section described] but can use another as specified there. General form [describe]. document pre-training pass & commands/configs used, alternatively download weights here***
+
+**Remind model is available in releases**
 
 ### Fine-tuning 
-***TODO: document fine-tuning pass & commands/configs, alternatively download weights here***
+***TODO: Same steps as above. document fine-tuning pass & commands/configs, alternatively download weights here***
+
+***Doc how freeze/no freeze layers etc, cf train.py***
+
+**Remind model is available in releases**
 
 ## Running the model for inference 
+**Specify in which folder there should be a model, can be gotten from (link to releases).**
 
-## Running tests
+### Inference on a directory of images
 
+**TODO: explain, gives the grid + eval metrics, show example of output, explain how to use the script with example usage**
+
+### Inference on a single image
+*μTODO: implement & document**
 
 ## Generating the datasets
 ### Pre-training dataset
-
+**Explain how could also use Night2Day by changing the one line for example**
 
 ### Fine-tuning dataset
+**Explain choices, also rotations all horizontal & resizing, multiple iterations of adding image not exactly event too but photographic style improved performance further... show how to run the script on a raw dataset. Remind it is available in releases**
+
+## Running tests
+**Also specify all are run on commits on main through Github Actions**
+
+# Usage of the web-application based on the model
+*Coming soon.*
+
+## Running the inference endpoint
 
+## Running the web application

milestones/hacking.md

@@ -0,0 +1 @@
+***TODO: rappel complément au README qui document le projet et son usage**