Hackathon - TF-TG literature triage Baseline Files

<strong>Files needed to use the baseline system.</strong> Instructions to use the baseline system: Install DeLFT as explained in https://github.com/kermitt2/delft (<em>If you are not using GPU, it is </em><em>recomended</em><em> to change the requirements.txt to “tensorflow==1.8.0” instead of “tensorflow_gpu==1.8.0”)</em> Download the optimized embeddings (Model_FastText.zip) available in this repository and place it inside the main delft folder. Copy the following files available in this repository: Greekc.zip: Extract the contents to ./delft/data/models/textClassification greekClassifier.py: Copy to ./delft reader.py: Substitute in ./delft/delft/textClassification/reader.py Embedding-registry.json: Substitute in ./delft/embedding-registri.json Fasttext_300_opt.zip: Extract the contents to ./delft/data/db To make predictions, change the line 96 in greekClassifier to point to the text you want to predict. Each new line should be an abstract. Then run python3 greekClassifier.py classify &gt; path_to_predicted