TensorFlow Time-Series Model Training - $10/h (fixed)
Budget: $10 – $30 USD
I am working on a Time Series Analysis problem and need a concise TensorFlow/Keras training script that I can build on later. The data are already cleaned and shaped; what I’m missing is the initial model definition, compilation, and a quick training loop that demonstrates over-fitting versus under-fitting on a small validation split.
Deliverables:
• A Python (.py or .ipynb) file that imports TensorFlow, defines a straightforward recurrent or 1-D convolutional architecture suitable for univariate sequences, compiles it, and runs a first training pass.
• Brief inline comments explaining each key step so I can tweak hyper-parameters myself.
• A short note (markdown or plain text) describing any package versions or command-line arguments needed to reproduce your results.
Keep it light—no extensive hyper-parameter sweeps or elaborate dashboards are required this round; I simply want a clear, working starting point that trains without errors.
Deliverables:
• A Python (.py or .ipynb) file that imports TensorFlow, defines a straightforward recurrent or 1-D convolutional architecture suitable for univariate sequences, compiles it, and runs a first training pass.
• Brief inline comments explaining each key step so I can tweak hyper-parameters myself.
• A short note (markdown or plain text) describing any package versions or command-line arguments needed to reproduce your results.
Keep it light—no extensive hyper-parameter sweeps or elaborate dashboards are required this round; I simply want a clear, working starting point that trains without errors.