Simplified PyTorch Implementation of AlphaFold 3

Hacker News

I created a simplified implementation of AlphaFold 3 (limited strictly to protein monomers) in PyTorch to see if it was possible to train on a single GPU. Rather than aiming to reproduce the accuracy of the original paper, the goal was only to generate plausible looking 3D protein structures. This turned out to be harder than expected - you can find a more in depth write up here - https://medium.com/@ogchen/the-worst-method-for-learning-ml-... . In summary, while most of the auxiliary losses did improve with training suggesting it was learning some structural information, the model's key loss defined on 3D coordinates never seemed to trend downwards (at the time of writing, the model is still training). Despite failing on that aspect, this project was initially a learning exercise (and one of my first machine learning related projects) and it was a success in that regard. While the outcome is perhaps not as exciting as I had initially hoped, I believe the project code stands on its own so decided to share it here!

Visit product View on Hacker NewsMay 20, 2024

Share card

Actual performance

2points

Did not reach leaderboard

Launch Intel predictions

Analyze your own launch →

Indie HackersFits the IH revenue-focused audience · Strong signals: created · Missing: supports, reddit linkedin, podcasting

68%68% predicted probability of success on Indie Hackers, based on ML models trained on real launch data.

best fitHighest predicted score across all platforms for this description.

Hacker NewsStrong engagement from HN community · Strong signals: ide, io · Missing: https docs, excited, just released

51%51% predicted probability of success on Hacker News, based on ML models trained on real launch data.

nativeThis product was originally launched on this platform.

TrustMRRLess likely to generate early MRR · Missing: mobile apps, ios, personal

45%45% predicted probability of success on TrustMRR, based on ML models trained on real launch data.

Product HuntUnlikely to reach the leaderboard · Strong signals: mac, model, single · Missing: agents, macos, agent

41%41% predicted probability of success on Product Hunt, based on ML models trained on real launch data.

AppSumoMay struggle as an AppSumo deal · Missing: plus, platform, intuitive

28%28% predicted probability of success on AppSumo, based on ML models trained on real launch data.

Acquire.comPre-revenue stage for this audience · Strong signals: training · Missing: arr, mrr, revenue

14%14% predicted probability of success on Acquire.com, based on ML models trained on real launch data.

BetaListMay not resonate with beta-testers · Missing: web3, chat, crypto

0%0% predicted probability of success on BetaList, based on ML models trained on real launch data.

Incorrect prediction on native model

Similar products

benchttp, simplified ApacheBench (ab) in Go43%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

benchttp, simplified ApacheBench (ab) in Go

Hacker News3

Okdb – OT Simplified43%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Okdb – OT Simplified

Hacker News37

Resticara – Simplified Restic Wrapper34%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Resticara – Simplified Restic Wrapper

Hacker News2

Simplified module dependencies for AngularJS34%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Simplified module dependencies for AngularJS

Hacker News4

Docket36%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Docket - Immigration simplified

Indie Hackers1b2b

Librin60%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Firestore, simplified.

Indie Hackers4productivity

Lightified45%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Lighting, simplified.

Indie Hackerscommitment-side-project

PagerTree20%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

On-Call. Simplified.

Indie Hackers

Chandrajit Ranawat20%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Hospitality Simplified

Indie Hackersemployees-200-plus

Ping16%Launch Intel prediction score: how likely this product is to succeed on its source platform, based on its name, tagline, and description.

Spontaneous hangouts, simplified

Product Hunt+6

Browse all products Analyze your launch