13hannes11 commited on
Commit
361ab76
1 Parent(s): e2d20e2

add details to readme

Browse files
Files changed (1) hide show
  1. README.md +18 -26
README.md CHANGED
@@ -20,44 +20,32 @@ tags:
20
 
21
  ## Description
22
 
23
- TODO
24
 
25
- ## How to run
26
-
27
- TODO
28
 
29
- Install dependencies
30
 
31
- ```bash
32
- # clone project
33
- git clone https://github.com/YourGithubName/your-repo-name
34
- cd your-repo-name
35
-
36
- # [OPTIONAL] create conda environment
37
- conda create -n myenv python=3.8
38
- conda activate myenv
39
 
40
- # install pytorch according to instructions
41
- # https://pytorch.org/get-started/
42
 
43
- # install requirements
44
- pip install -r requirements.txt
45
- ```
46
 
47
- Train model with default configuration
48
 
49
  ```bash
50
- # train on CPU
51
- python train.py trainer.gpus=0
52
-
53
- # train on GPU
54
- python train.py trainer.gpus=1
55
  ```
56
 
57
- Train model with chosen experiment configuration from [configs/experiment/](configs/experiment/)
58
 
59
  ```bash
60
- python train.py experiment=experiment_name.yaml
61
  ```
62
 
63
  You can override any parameter from command line like this
@@ -65,3 +53,7 @@ You can override any parameter from command line like this
65
  ```bash
66
  python train.py trainer.max_epochs=20 datamodule.batch_size=64
67
  ```
 
 
 
 
 
20
 
21
  ## Description
22
 
23
+ This repository contains the models and training pipeline for my master thesis. The main repository is hosted on [GitHub](https://github.com/13hannes11/master_thesis_code).
24
 
25
+ The project structure is based on the template by [ashleve](https://github.com/ashleve/lightning-hydra-template).
 
 
26
 
27
+ The metadata is stored in `data/focus150/`. The relevant files are `test_metadata.csv`, `train_metadata.csv` and `validation_metadata.csv`. Image data (of 150 x 150 px images) is not published together with this repository therefore training runs are not possible to do without it. The layout of the metadata files is as follows
28
 
29
+ ```csv
30
+ ,image_path,scan_uuid,study_id,focus_height,original_filename,stack_id,obj_name
31
+ 0,31/b0d4005e-57d0-4516-a239-abe02a8d0a67/I02413_X009_Y014_Z5107_750_300.jpg,b0d4005e-57d0-4516-a239-abe02a8d0a67,31,-0.013672000000000017,I02413_X009_Y014_Z5107.jpg,1811661,schistosoma
32
+ 1,31/274d8969-aa7c-4ac0-be60-e753579393ad/I01981_X019_Y014_Z4931_450_0.jpg,274d8969-aa7c-4ac0-be60-e753579393ad,31,-0.029296999999999962,I01981_X019_Y014_Z4931.jpg,1661371,schistosoma
33
+ ...
34
+ ```
 
 
35
 
 
 
36
 
37
+ ## How to run
 
 
38
 
39
+ Train model with chosen experiment configuration from `configs/experiment/`
40
 
41
  ```bash
42
+ python train.py experiment=focusResNet_150
 
 
 
 
43
  ```
44
 
45
+ Train with hyperparameter search from `configs/hparams_search/`
46
 
47
  ```bash
48
+ python train.py -m hparams_search=focusResNetMSE_150
49
  ```
50
 
51
  You can override any parameter from command line like this
 
53
  ```bash
54
  python train.py trainer.max_epochs=20 datamodule.batch_size=64
55
  ```
56
+
57
+ ## Jupyter notebooks
58
+
59
+ Figures and other evaluation code was run in Jupyter notebooks. These are available at `notebooks/`