Deep Learning Model Debug & Tune
Budget: $25 – $50 USD
I need targeted debugging and tuning for a mixed-mode deep learning system. The model currently low accuracy.
Scope for this fixed-price, fixed-time task:
• Review the existing training pipeline and spot the key causes of overfitting or underfitting.
• Adjust hyper-parameters, introduce lightweight regularisation or data-augmentation steps, and refine any architectural bottlenecks to close the train-test performance gap.
• Provide a concise change log and before-versus-after metrics so I can see the impact of your updates.
Keep all edits self-contained within the current codebase; no major rewrites are expected. Once the revised model runs cleanly and shows improved generalisation, the job is done.
Scope for this fixed-price, fixed-time task:
• Review the existing training pipeline and spot the key causes of overfitting or underfitting.
• Adjust hyper-parameters, introduce lightweight regularisation or data-augmentation steps, and refine any architectural bottlenecks to close the train-test performance gap.
• Provide a concise change log and before-versus-after metrics so I can see the impact of your updates.
Keep all edits self-contained within the current codebase; no major rewrites are expected. Once the revised model runs cleanly and shows improved generalisation, the job is done.
Related categories:
Machine Learning (ML)
Data Science
Neural Networks
Tensorflow
Keras
Deep Learning