Show HN: I have trained a model to predict my blood sugar
1 comments
- Points5
- Comments2
1 comments
I have trained a custom 31,253 parameter encoder-only transformer on a pregenerated synthetic CGM dataset and fitted it to my own personal data using LoRA. It is competitive among other published models. I used 115 days of CGM data, of which 80 was used for training the adapter and 35 for testing. The model began with a pretraining stage where it consumed ~19,200 CGM-years worth of synthetic data, consisting of 1 million simulated type 1 diabetics at 7 days each. The simulator is designed to…
Story details from Hacker News ↗ · linked article on github.com.
Show HN: I have trained a model to predict my blood sugar was first seen on Hacker News · Show HN on Nov 14, 2026. Since then it has appeared in 25 checks on 1 different day. Its best position was #15 of 30, reached on Nov 14, 18:22 UTC. It dropped off the list after Nov 14, 20:52 UTC.
Position on Hacker News · Show HN at every check (lower is better).
Best rank per day (UTC).
Platform metric as reported by Hacker News.
A few top-level comments from Hacker News, shortened. Read the full thread on Hacker News.
I wrote a website that helped my wife's grandfather manage his diabetes. It was essentially captured that moment in time his blood sugar, blood pressure, and weight. This was 10 years ago, and his doctor loved it because he could generate a report and bring it to his visits. What you're doing here is far beyond what…
Archived daily, weekly and monthly lists that include this entry.
Not at the moment. It was last on the list on Nov 14, 20:52 UTC, after 1 day of trending.
#15 on Hacker News · Show HN, on Nov 14, 18:22 UTC. Trendoday checks this list every 30 minutes.
It was first seen on Nov 14, 2026 and has been on the list on 1 day (25 checks) since.
5 points and 2 comments; submitted by 0xdeadf1sh on Nov 13, 2026.