I will train a custom deep learning model for your image data
Python Data Scientist, AI Model Training, Data Analysis and Cleaning
About this Gig
I will build and train a deep learning model for your specific image classification or detection task using PyTorch.
What you get:
- Trained model file (.pt) ready to use
- Training and evaluation report with accuracy, precision, recall, F1-score
- Python inference script for running predictions on new data
- Full documentation
Types of projects I work on:
- Product image classification
- Defect and anomaly detection
- Medical image analysis
- Satellite and aerial imagery
- Object recognition and categorization
- Binary and multi-class classification
Tech: Python, PyTorch, NumPy, OpenCV, Matplotlib, scikit-learn
I have research experience in computer vision, with a deep learning project that improved over published baseline results by +10.2% to +17.6% F1-score using domain adaptation and uncertainty estimation.
I train models locally on my RTX 4050 GPU with CUDA acceleration, ensuring fast training and delivery.
Delivery: Trained model (.pt) + inference script + evaluation report
Programming language:
Python
•
SQL
•
Colab
•
Java
APIs:
Other
Tools:
Jupyter Notebook
•
OpenCV
•
CVAT
•
Colab
Frameworks:
Scikit-learn
•
Keras
•
PyTorch
•
Panda
FAQ
What format should my images be in?
JPG, PNG, or any standard image format. Organize them in folders by class/category if possible.
How much data do I need?
At least 50-100 images per class for decent results. More data means better accuracy.
Can I run the model on my own computer?
Yes. I deliver a Python script that works on any machine with Python installed. No GPU required for predictions.
Do you also do object detection (bounding boxes)?
Yes, contact me before ordering and I will set up the right package for your needs.

