Calculating embeddings with gtr-t5-large in Python
I've long wanted to run some kind of large language model on my own computer. Now that I have a M2 MacBook Pro I'm even more keen to find interesting ways to keep all of those CPU cores busy.
Notes on software engineering
261–280 of 547 posts, newest first.
I've long wanted to run some kind of large language model on my own computer. Now that I have a M2 MacBook Pro I'm even more keen to find interesting ways to keep all of those CPU cores busy.
https://www.w3.org/History/1991-WWW-NeXT/Implementation/ It's served up as a browseable Apache directory listing, but that's not the most delightful way to browse code - clicking a link to a file…
I'm trying a new thing: a private daily planner, where each day I note down my goals for the day and make notes on my progress towards them as the day progresses.
Tiny TIL today: I learned how to make an HTTP call to an IPv6 address. The trick is to enclose the address in the URL in square braces: http:// 2a09:8280:1::1:2741 Here's that working as a request to…
I use ReadTheDocs for several of my projects. It's fantastic: among other things, it makes it easy to publish the documentation for my latest branch at and the documentation for my latest release at…
TIL that in the DOM has a size limit! I had a table cell element containing HTML-escaped JSON and I was doing this: This was breaking on larger JSON strings.
The latest version of SpatiaLite adds KNN support, which makes it easy to efficiently answer the question "what are the X closest records to this point".
I host the documentation for Datasette on Read the Docs. Until today it lived at https://datasette.readthedocs.io/ but today I moved it to a custom subdomain, https://docs.datasette.io/ Most of this…
uses a cache to avoid downloading packages again: The command can be used to find the location of that cache on your system: Wheels are cached in - in a nested set of folders based on a hash, for…
I managed to get my Homebrew installation back into shape today. The first problem I was having is that it complained that macOS Sequoia was unsupported: It turns out I was on an older Homebrew…
This TIL adapted from a Gist I put together in 2019, before I started tracking TILs here. My twitter-to-sqlite tool produced a SQLite table with an column that referenced another tweet ID, for…
I was putting together some notes for a talk I gave, and I wanted an efficient way to create screenshots of specific moments in a video of that talk.
This tweet from \@Mariatta tipped me off to the ability to measure "contexts" when running coverage - as a way to tell which tests exercise which specific lines of code.
I have my own private repository where I sometimes create research threads. Occasionally I want to transfer these to a public repository to publish their contents.
I use the MacWhisper macOS desktop app to run Whisper. It's a very pleasant GUI wrapper around the Whisper transcription model.
I was trying to run against a large SQLite database file ( 7GB) using and I got this error: sqlite3.OperationalError: database or disk is full The volume that the database lived in was 20GB in size…
The homebrew-core repository contains all of the default formulas for Homebrew. It's a huge repo, and if you browse it through the GitHub web interface you can run into errors like this one…
I tried to get jupyterlab working via today and ran into some sharp edges. You can start like this: That ran for a while and output: It also gave me a warning about my PATH.
The GitHub Actions default runner currently includes an installation of PostgreSQL 13. The server is not running by default but you can interact with it like this: You can install alternative…
I tried this for the first time today with my highly experimental datasette-table Web Component. Here's the source code for version 0.1.0.