The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Dog Breed Classification System
Built a deep learning computer vision application capable of classifying 120 dog breeds from uploaded images.
Implemented transfer learning using ResNet and EfficientNet architectures with PyTorch and Torchvision.
Improved model performance from ~3% CNN baseline accuracy to 78.97% ensemble accuracy through optimization and fine-tuning.
Applied advanced deep learning techniques including label smoothing, gradual unfreezing, weight decay, learning rate scheduling, and Test Time Augmentation (TTA).
Deployed a production-ready Streamlit application with real-time image inference and probability predictions.
This is interesting! Which dog is the easiest to classify?
Hey Emman, it classified most dog breeds really well but I think a husky is the easiest to classify for me and its my favorite dog 😊, you can check out different dog breeds yourself and use the graph in the app to compare breedies prediction scores.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
A goal-tracking sub-system that grounds an AI assistant in the user's full life-state — target amount, hard deadline, days remaining, current balance, pace status (ahead / on-pace / behind), and reflection streak.
Core features:
Multi-entry timestamped reflection log — plain Enter submits, Shift+Enter newline, optimistic UI, all entries inside one unified container with centered timestamp dividers
Ephemeral Summarize chat modal — loads today's braindumps as live AI context, brother-tone system prompt, conversation NOT persisted (privacy by design)
Nightly auto-summary cron at 23:55 — pulls reflections, runs through Claude, creates a standalone daily summary that downstream AI engines pick up automatically
Pace analysis — surfaces required-vs-actual daily rate, projected gap at deadline, status flag
Each AI engine pulling this data adapts its suggestions based on whether the user is ahead or behind on their goal — recursive context that compounds over time.
Tech: Claude API, SQLite, Node.js cron, custom retrieval logic.
Open your homepage. Open your three closest competitors. Squint.
If you can't tell which one is yours from across the room, your buyer can't either. And they're trying to keep six of you straight in a Notion doc with seven other tabs open.
Here's what's happening in the AI category right now.
The dark mode. The lowercase wordmark. The orange-to-purple gradient. The one-word Greek-dictionary name. The abstract mark gesturing at "intelligence." Every AI startup pulled the same playbook from OpenAI and Anthropic's decks. Each decision is defensible alone. Together they're a uniform.
The uniform got the early winners funded, so every Series A since reverse-engineered it. The same fifteen VCs see the same thirty decks per quarter, and founders calibrate to what looks fundable, which means looks like the last company that got funded.
That trade made sense at seed. It stops making sense the moment you have to sell something.
Your buyer doesn't pattern-match on "this looks AI-native." They pattern-match on "is this the one Sarah recommended on Slack." Memorability is the moat. The category-standard look is a discount you apply to your own pipeline every visit.
Wrote about who broke rank, what it bought them, and the audit that tells you whether you're paying the tax. Link in comments.
Nice one
A goal-tracking sub-system that grounds an AI assistant in the user's full life-state — target amount, hard deadline, days remaining, current balance, pace status (ahead / on-pace / behind), and reflection streak.
Core features:
Multi-entry timestamped reflection log — plain Enter submits, Shift+Enter newline, optimistic UI, all entries inside one unified container with centered timestamp dividers
Ephemeral Summarize chat modal — loads today's braindumps as live AI context, brother-tone system prompt, conversation NOT persisted (privacy by design)
Nightly auto-summary cron at 23:55 — pulls reflections, runs through Claude, creates a standalone daily summary that downstream AI engines pick up automatically
Pace analysis — surfaces required-vs-actual daily rate, projected gap at deadline, status flag
Each AI engine pulling this data adapts its suggestions based on whether the user is ahead or behind on their goal — recursive context that compounds over time.
Tech: Claude API, SQLite, Node.js cron, custom retrieval logic.
Trending
Claude
Claude has entered the design space. How are you using Claude Design?
Contra University
Learn from expert creatives how to earn more using next-gen AI tools.
creativeaiflow
Creative AI workflows are evolving. What tools do you use, and what are their strengths and weaknesses?
portfolioreview
The best portfolios tell a story, not just show a grid. Share yours for feedback.
freelancerlife
Freelancer life is wins, pivots, and everything in between. What’s yours right now?