TornadoVM-MLOps-RENOPS

This repository shows how a ML model for TornadoVM can be trained using:

• MLOPs to track the training properties.
• RENOPS scheduler to schedule the training process when cost is lower.

1. Clone the repository

git clone https://github.com/stratika/tornadovm-mlops-renops.git

2. Move into the project directory

cd tornadovm-mlops-renops

3. Load environment variables

source load_variables.sh

4. Start the MLOps container

./mlops/start_mlops_container.sh

i. You can open the MLFlow window to see any ML training tasks

google-chrome http://127.0.0.1:5001/

5. Run the training of the TornadoVM model with RENOPS

i. The first time install the Python dependencies for the training

./tornadovm-training/install.sh

ii. Configure the script with the required RENOPS arguments and train the model

./tornadovm-training/run_renops_tornadovm.sh 
2025-04-10 15:33:20.444 | INFO     | renops.main:run:36 - RUNNING RENOPS SCHEDULER...
2025-04-10 15:33:20.444 | INFO     | renops.main:run:154 - Location specified, shifting in time...
2025-04-10 15:33:20.717 | WARNING  | renops.geolocation:_get_location_params:34 - Location is set to auto, IP will be used to detect location! found: Athens, GR
2025-04-10 15:33:21.545 | INFO     | renops.scheduler:run:193 - Task has to be finished by: 2025-11-04 15:33:21
2025-04-10 15:33:21.545 | INFO     | renops.scheduler:run:201 - No renewable window whitin a given deadline!
2025-04-10 15:33:21.545 | INFO     | renops.scheduler:_print_info_for_instant_execution:177 - Current energy price is: 9.64 EUR/MWh
2025-04-10 15:33:21.545 | INFO     | renops.scheduler:run:247 - Executing action now at 2025-04-10 15:33:21.545432
2025-04-10 15:33:21.545 | INFO     | renops.scheduler:run:249 - ----------------------------------------------------
2025/04/10 15:33:24 WARNING mlflow.models.model: Model logged without a signature and input example. Please set `input_example` parameter when logging the model to auto infer the model signature.
[Classifier 1] Dummy Accuracy: 0.0000
🏃 View run DummyClassifier_C1 at: http://127.0.0.1:5001/#/experiments/1/runs/0e2e67bfd0ce4299b415ee845d69bbf2
🧪 View experiment at: http://127.0.0.1:5001/#/experiments/1
2025/04/10 15:33:26 WARNING mlflow.models.model: Model logged without a signature and input example. Please set `input_example` parameter when logging the model to auto infer the model signature.
[Classifier 2] Dummy Accuracy: 0.0000
🏃 View run DummyClassifier_C2 at: http://127.0.0.1:5001/#/experiments/1/runs/29dd92dc14604a25bf2e68c3bdca9c12
🧪 View experiment at: http://127.0.0.1:5001/#/experiments/1
2025/04/10 15:33:28 WARNING mlflow.models.model: Model logged without a signature and input example. Please set `input_example` parameter when logging the model to auto infer the model signature.
[Classifier 3] Dummy Accuracy: 0.0000
🏃 View run DummyClassifier_C3 at: http://127.0.0.1:5001/#/experiments/1/runs/fd2fe292aa87462383a1d7aa326c6dd8
🧪 View experiment at: http://127.0.0.1:5001/#/experiments/1

iii. Observe the MLFlow window and see the tracked information for the trained model

6. Stop the MLOps container

./mlops/stop_mlops_container.sh

Acknowledgments

This work is funded by the European Union's Horizon Europe programme under grant agreement No 101070052 and the UK Reseach and Innovation Horizon Europe guarantee scheme grant No 10039107 (TANGO).

License

This project uses a two-licenses scheme. The mlops folder is licensed under the GNU Affero General Public License v3.0. Whereas, the tornadovm-training folder is licensed under Apache 2.0.

Module	License
mlops
tornadovm-training