jamez

7 ms·

31.
by jamez 4y ago
When you can't find a connection, always try the reverse! https://www.wikibinge.com/#Great_Pacific_garbage_patch/Cleop...
32.
by jamez 4y ago
Cool path! The last bit of the tour you described is fairly common - the algorithm is just doing its job, getting closer to the target, avoiding articles with larger PageRank. It's very uncommon to pass through "important" pa
33.
Show HN: WikiBinge – discover how all things are vaguely connected(wikibinge.com)
246 points by jamez 4y ago|61 comments
34.
by jamez 4y ago
I think channels like Primitive Technology[1] play with this sense of dizziness, the realization that crafting even basic tools of survival requires a chain of dependencies and non obvious skills to acquire. Separating from society, and thr
35.
by jamez 4y ago
This YouTube video (in Dutch, but with English subs) is also fascinating: https://www.youtube.com/watch?v=w1ODwgX_KkQ You can browse Tatjana van Vark's personal website here: http://www.tatjavanvark.nl/
36.
The machines of Tatjana van Vark(craftsmanshipmuseum.com)
224 points by jamez 4y ago|33 comments
37.
by jamez 4y ago
It was fine tuning, so the process was a lot faster than I originally anticipated. I'd say it was between 36 and 72 hours for each voice. I have been working on a gradient notebook provided by Paperspace, which guaranteed me A6000 inst
38.
by jamez 4y ago
At this point, zero. The framework I built automatically rejects certain patterns that are not conducive to an interesting conversation. The only thing I still do manually, and I think I will automate, is to decide when to stop a generated
39.
by jamez 4y ago
Ah! Welcome to my head.
40.
by jamez 4y ago
The TTS model is trained on two things: speech samples and their transcript. If you add enough sniffle-symbols every time a sniffle appears in the speech, I am confident the model would pick up on that. And then you would be able to replica
41.
by jamez 4y ago
I use Aeneas[1], a set of tools to do force alignment. I found it in equal measures an amazing and a hard to navigate resource. Took me a while to set up and configure everything to the point that it was usable. But when it works, it works
42.
by jamez 4y ago
Thanks for asking, I think I'll do a write up later this week. Let me know if you have any specific question.
43.
by jamez 4y ago
One of the guiding lights at the very beginning of this project was the question: "who would I never get tired of listening to?"
44.
by jamez 4y ago
The main issue is that there was no sniffling symbol in the transcript. And the generated text wouldn't contain it either, because (thankfully) they are pruned out of written interviews that I used to train the model.
45.
by jamez 4y ago
No need to be invited. Between their GitHub[1] page and the documentation[2], you'll find everything you need to get started. [1] https://github.com/coqui-ai/TTS [2] https://tts.readthedocs.io/en&#
46.
by jamez 4y ago
I haven't tried Tortoise, thanks for pointing me to it. The voices were cloned by fine tuning a VITS model with coqui.ai. I used about two hours of speech for each speaker. With more time and resources, I'm certain it's possi
47.
by jamez 4y ago
Hi there! "Author" here - glad to see this picking up. This was a fascinating project to work on and I learned a ton in the process. As it's often the case, I would do a lot of things differently if I were starting from scrat
48.
Herzog and ŽIžek endlessly talking philosophy and cinema, AI-generated(infiniteconversation.com)
4 points by jamez 4y ago|0 comments
49.
by jamez 4y ago
I suspect that the 5.65inch ACeP 7-Color E-Paper E-Ink Display Module (600×448) sold by Waveshare[1] is essentially the same screen, just with a different controller glued to the back. I use it with an ESP32, but their wiki[2] has a guide o
50.
The machines of polymath Tatjana van Vark(youtube.com)
5 points by jamez 4y ago|2 comments
51.
by jamez 6y ago
Hi Udik - thanks for your thoughts. To the best of my knowledge, it's impossible to mix languages at the word level without creating conflicts. In part, that's the challenge and the fun of a project like this. But mostly, it'
52.
Gradually morphing the tale of Pinocchio from English to Italian(jamez.it)
3 points by jamez 6y ago|2 comments
53.
by jamez 6y ago
I published this yesterday, in case you are interested: https://news.ycombinator.com/item?id=25461853
54.
by jamez 6y ago
The part about displaying a website every day is trivial, though that budget will not take you very far for a decently sized ePaper panel. Here's hoping that prices will go down in the future...
55.
by jamez 6y ago
I was not aware of that Google Calendar display. Very nice, thanks for sharing. Re: tinted glass on top of the border, it's an interesting angle, although I'm not sure I want to completely remove the look of a panel from the pictu
56.
by jamez 6y ago
Thanks for the suggestions! Thermoelectric energy harvesting never even occurred to me. Definitely looking into it.
57.
by jamez 6y ago
Clickable link to project description: https://www.jamez.it/blog/2020/12/17/made-epaper-solar-power...
58.
Show HN: I made a solar-powered, ePaper photo frame
27 points by jamez 6y ago|11 comments
59.
by jamez 6y ago
Good read. For anyone interested in seeing the human side of Dijkstra, I recommend this documentary: https://www.youtube.com/watch?v=RCCigccBzIU When I watched it I was both studying Computer Science and learning Dutch, so
60.
by jamez 6y ago
Yes, it was a great experience. I too have the impression that investors these days are looking for a more proven track record before buying in, especially for B2C products, but I think some markets are still open to "gambles". As

More ›