To run start the training, you can run the training script with
python src/pml_vqvae/train.py
On default (without extra parameters givin in the command), the training script will look up the config.yaml
-file and will train according those defined parameters.
If you want to change the default values you can either change them in the config.yaml
or you can pass the values yu want to change via the cli, e.g.
python src/pml_vqvae/train.py --learning_rate 0.1
This will take the config.yaml
as the foundation and overwrites all the given values passed in the command. See the python src/pml_vqvae/train.py -h
for more information.
ssh <username>@hydra.ml.tu-berlin.de
srun --partition=gpu-test --gpus=1 --pty bash
cd pml_vqvae
./run_script.sh
srun --partition=cpu-2h --pty bash
pml_vqvae
)apptainer build pml.sif pml.def
(can take up to 15min)export WANDB_API_KEY=<your key>
. If you run train the model within an apptainer make sure to add the environment variable to your pml.def
-fileconfig.yaml
-file. make sure log_wandb=True
Open the .tex file you want to edit, press Ctrl + Alt + B (Windows/Linux), Cmd + Option + B (macOS) to build the PDF and then Ctrl + Alt + V (Windows/Linux), Cmd + Option + V (macOS) to view the PDF. Then edit your file and you should get live updates in the PDF view.
Generally, you should be able to open the .tex document that you are interested in and use the following commands to compile, view, and navigate the document:
Compile the LaTeX document
View PDF Preview
Sync PDF with Source (Forward Sync)
Sync Source with PDF (Reverse Sync)