Update README

Author: ttanida

README.md

@@ -33,7 +33,8 @@ The automatic generation of radiology reports has the potential to assist radiol
 ## Setup
 
 1. Create conda environment with "**conda env create -f environment.yml**"
-2. In src/path_datasets_and_weights.py, specify the paths to the various datasets (Chest ImaGenome, MIMIC-CXR, MIMIC-CXR-JPG), CheXbert weights, and folders in which the runs are saved. Follow the instructions of the doc string of path_datasets_and_weights.py.
+2. Install Java 1.8.0 (required for pycocoevalcap library, see https://pypi.org/project/pycocoevalcap/). On Ubuntu, you can install Java 1.8.0 with "**sudo apt install openjdk-8-jdk**".
+3. In src/path_datasets_and_weights.py, specify the paths to the various datasets (Chest ImaGenome, MIMIC-CXR, MIMIC-CXR-JPG), CheXbert weights, and folders in which the runs are saved. Follow the instructions of the doc string of path_datasets_and_weights.py.
 
 ## Create train, val and test csv files
 

environment.yml

@@ -7,13 +7,8 @@ channels:
 dependencies:
   - spacy-model-en_core_web_trf
   - pillow=9.2.0
-  - pip=22.2.2=pyhd8ed1ab_0
-  - python=3.9.13=h9a8a25e_0_cpython
-  - python_abi=3.9=2_cp39
-  - pytorch=1.12.1=py3.9_cuda11.6_cudnn8.3.2_0
-  - pytorch-mutex=1.0=cuda
-  - torchaudio=0.12.1=py39_cu116
-  - torchvision=0.13.1=py39_cu116
+  - pip=22.2.2
+  - python=3.9.13
   - pip:
     - -e .
     - albumentations==1.1.0
@@ -45,7 +40,9 @@ dependencies:
     - tensorboard-plugin-wit==1.8.1
     - tensorboardx==2.5.1
     - tokenizers==0.12.1
+    - torch==1.12.1
     - torchinfo==1.7.0
     - torchmetrics==0.9.3
+    - torchvision==0.13.1
     - tqdm==4.64.0
     - transformers==4.19.2