victorjanni/Ai-engineering-hub — GitHub trending stats & insights | Trendshift

Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

Terms and conditions
Privacy policy

More

onedump
gosnakego
Contact us

Social

X (Twitter)
GitHub

What they build speaks louder than words

© 2026 Trendshift, created by Julian Li with ❤️

Discover trends that matter

Daily Weekly Monthly Yearly

Live mentions Topics

Repositories Developers

victorjanni/Ai-engineering-hub — GitHub trending stats & insights | Trendshift

Featured

victorjanni/Ai-engineering-hub

⚡ AI Engineering Hub — Exploring modern AI engineering with PyTorch, Agentic AI, LangGraph, LangChain, RAG pipelines, multi-agent systems, LLMOps, workflow orchestration, and scalable AI applications.

Jupyter Notebook

1 contributors

Social mentions

Recent discussions about this repository across the web

the code is surprisingly clean: def objective(trial): lr = trial.suggest_float("lr", 1e-5, 1e-1, log=True) # ... more params # train model return accuracy study =…

@VictorJanni · x.com

results: before → train: 97.95% | test: 88.69% | gap: 9pts after → train: 93.25% | test: 88.32% | gap: ~5pts test accuracy barely moved. train accuracy dropped on purpose. that's the model…

@VictorJanni · x.com

evaluation: model.eval() + over test_loader torch.max(outputs, 1) → predicted class index accuracy: 83.25% across 10 classes 83% with a plain ANN and no convolutions. solid baseline before moving to…

@VictorJanni · x.com

evaluation: model.eval() + over test_loader threshold at 0.8 → binary prediction accuracy: 94.88% ✓ the jump from 60% wasn't a better model. it was proper batching + a real optimizer doing the job…

@VictorJanni · x.com

The same training loop works for 10 samples or 10 million. you don't change a single line of loop code. DataLoader is what makes that possible — it's not a convenience, it's the foundation.

@VictorJanni · x.com

loss: 3.93 → 0.72 over 35 epochs accuracy: ~60% not great — but that wasn't the point. i wanted to know what nn.Module and torch.optim actually do before using them. now i do.

@VictorJanni · x.com

3ways to stop gradient tracking when you don't need it (like inference): → requires_grad_(False) — disables it on the tensor permanently → .detach() — returns a new tensor with no grad history → —…

@VictorJanni · x.com

No trending activity

This repository has not yet been featured on GitHub Trending

Repository activities

repository's daily and monthly activities across stars, forks, merged PRs, issues, and closed issues

Data is not available yet

Recent activity data for stars, forks, merged PRs, issues, and closed issues will appear here once available

Featured

Bindu: The identity, communication, and payments layer for AI agents

Your Personal AI super intelligence. Private, Simple and extremely powerful

100K+ monthly visitors

High-intent: developers tracking what's rising in OSS to decide what to try next

No tracking pixels, no popups

Become a sponsor