Open source implementation for NIM: Neural Input Modeling
- Install NIM:
pip install git+https://github.com/cenwangumass/nim.git
- In the
examples
directory, create a synthetic dataset with 1,000 length 100 sequences and values are sampled from i.i.d. Gamma:python create_data.py -n 1000 -t 10 data.npy
- Train the neural network:
python train.py nvm data.npy --transform transform:log_nvm
. For more options, runpython train.py --help
- Generate new samples and save in
samples.npy
:python generate.py nvm.json samples.npy -n 1000 -t 10 --transform=transform:exp