Телеграм канал «Morph’s Vintages»

As heartwrenching it is to watch, unfortunately there is little to no market for these watches. I tried to save a few of them, but it became a financial burden. If we were in new york, unless the watch was completely ruined, it would easily fetch 2x its weight value in gold. Sad. They don't even take out components like the movement and dial (which would fetch so much). They think the whole thing is gold. If you have a watch that you want to sell, old one, let me know. I can buy it, if I cant I know people that will. Don't sell it to a jewler, you'll get way less and the watch will be molten to make crap

A scene from a show called Cupertino, I was getting bored then they pulled out this curveball. Genius! 😭 it’s either a spoiler or a nice hook.

I went to Entoto after a very long while. I can actually feel the difference in air quality. It feels like I can take deeper, more satisfying breaths. Whoa

“If buying isn’t owning then piracy isn’t stealing”
- nautically inclined tiktoker

We open-sourced Addis Scribe Streaming, realtime speech recognition for Amharic.

It beats Google Chirp 3 on both accuracy and speed:

Word error rate: 28.4% vs 32.5%
First words on screen: 1.56s vs 6.67s
Final text after you stop talking: 30ms vs 1,411ms

On natural speech (WAXAL) we beat Google Chirp 3 in both streaming and batch mode. On FLEURS we went from 35.0% to 19.8% WER.

Hohe remains the most accurate offline Amharic model in our WAXAL test, and on FLEURS the two are within a point of each other (19.8% vs 20.6%). We built Addis Scribe for the live and edge inference side, where text has to appear while people are still talking.

Trained on 5,000+ hours of Amharic speech. Test sets were never used in training. Per-clip results and scripts are in the repo.

Try it😊: https://docs.addisassistant.com/docs/playground/speech-to-text

API and SDK docs: docs.addisassistant.com

HuggingFace🤗: huggingface.co/addisai/addis-scribe-streaming

@addisassistantai

The gift that keeps on giving

What’s happening?