The idea of this repository is to work on a given pre-trained model so that two main tasks can be accomplished:
- Extraction of the trained weights for the specified convolutional layers
- Extraction of the input and weight indexes of the specified convolutional layers which will be used by the multiplier of the XHEEP architecture during the convolutional operations
Note: to date [04/2024] an interpreter < "Python 3.12" must be used due to the lack of compatibility between Tensoflow module and the latest version of the interpreter
conda create -n <env_name>
conda init <your_shell_name>
, then restart the shell
conda activate <env_name>
- to deactivate an environment use:
conda deactivate <env_name>
- to list all your environments use:
conda info --envs
conda install <channel_name>::<pkg_name>