Raman spectra contain abundant information from molecules but are difficult to analyze, especially for the mixtures. Deep-Learning-Based Components Identification for Raman Spectroscopy (DeepCID) has been proposed for the problem of components identification. Convolution Neural Network (CNN) models have been established to predict the presence of the components in the mixtures.
conda create -n DPCID python=3.8
1.Numpy
pip install numpy==1.19.2
2.Tensorflow
pip install tensorflow-gpu==2.5.0
3.Scipy
pip install scipy==1.6.0
4.Matplotlib
pip install matplotlib==3.3.4
5.Cuda
conda install -c conda-forge cudatoolkit=11.2
conda install -c conda-forge cudnn=8.1.0
Since the model exceeded the limit, we have uploaded all the models and the information of mixtures to Release.
1.Training your model
Run the file 'one_component_model.py'.The corresponding example data have been uploaded to the folder named 'augmented data'.
Zhi-Min Zhang: [email protected]