I will train a custom image classification or object detection model in pytorch
Data Scientist, Machine Learning, Predictive Analytics
About this Gig
Have images you need to sort, count or inspect automatically? Send me 20 to 50 sample images and I will tell you, for free, whether a reliable model is possible before you order.
WHAT I BUILD
- Image classification with convolutional neural networks (CNN) and transfer learning in PyTorch
- Object detection with YOLO: find and count items in each image
- Training on your own images, on a local GPU
WHAT YOU GET
- A trained model and a clear report: accuracy, confusion matrix, examples of errors
- A Python script to run the model on new images
- On Premium: a simple web demo to upload an image and see the result
TYPICAL USES
Product or defect inspection, sorting photos into categories, counting items, document type classification.
Standard and Premium include image labeling. Your images stay confidential and are deleted after delivery.
Message me with sample images: free assessment, no obligation.
Expertise:
Image processing
•
Classification
•
Object detection
Programming language:
Python
Tools:
Jupyter Notebook
•
OpenCV
•
PyTorch
Frameworks:
Scikit-learn
•
PyTorch
Other Data Science & ML Services I Offer
FAQ
How many images do I need?
For a classifier, around 100 per class is a good start thanks to transfer learning. Send me 20 to 50 samples first and I will tell you if it is enough.
Can you label my images?
Yes. Standard includes labeling of up to 300 images and Premium up to 600.
Can I run the model myself?
Yes. You get the trained model and a Python script to run it on new images, with instructions.
Are my images kept confidential?
Yes. Training runs on my own GPU, your images are used only for your order and deleted after delivery.

