fetch is Fast Extragalactic Transient Candidate Hunter. It has been detailed in the paper Towards deeper neural networks for Fast Radio Burst detection

You can use anaconda, and install fetch like this:

conda install -c conda-forge tensorflow-gpu keras scikit-learn pandas scipy numpy matplotlib scikit-image scikit-image scikit-image tqdm 
git clone https://github.com/devanshkv/fetch.git
cd fetch
python setup.py install

You would also require pysigproc to create the candidate files which can be found here.

The installation will put predict.py and train.py in your PYTHONPATH. To predict a bunch of candidate h5 files living in a directory /data/candidates/ use predict.py for model a as follows:

    predict.py --data_dir /data/candidates/ --model a

To fine-tune the model a, with a bunch of candidates, put them in a pandas readable csv, candidate.csv with headers 'h5' and 'label'. Use

    train.py --data_csv candidates.csv --model a --output_path ./

This would train the model a and save the training log, and model weights in the output path.

To generate the candidate h5 files containing DM-time and Frequency-time arrays, candmaker.py can be used. It can also be used to rebin the candidates on the fly. For example: Saving candidate h5s with their parameters in cands.csv to a directory /my/canddidates/ and rebinning the time and frequency axis to 256 bins using decimation can be done by:

    candmaker.py --frequency_size 256 --time_size 256 --cand_param_file cands.csv --plot --fout /my/candidates/

A typical example of cands.csv file would be (here boxcar_width is in units of log2(number of samples)):

    /path/to/filterbank/myfilterbank.fil,S/N,start_time,dm,boxcar_width,label