# lstm

Published articles for lstm.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Continuous Thought Machine: Sakana AI's experimental architecture for time-based neural computation

DevFeed: [Continuous Thought Machine: Sakana AI's experimental architecture for time-based neural computation](<https://devfeed.tech/articles/continuous-thought-machine-sakana-ai-24020.md>)

Original publisher: [Read original article](<https://habr.com/ru/companies/redmadrobot/articles/916364/>)

Author: redmadrobot (red\_mad\_robot)

Published: 2025-06-06T16:44:09Z

Content type: article

Language: ru

Sources: [Redmadrobot EN](<https://devfeed.tech/sources/redmadrobot-en.md>), [Redmadrobot RU](<https://devfeed.tech/sources/redmadrobot-ru.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [cifar-10](<https://devfeed.tech/tags/cifar-10.md>), [ctm](<https://devfeed.tech/tags/ctm.md>), [cv](<https://devfeed.tech/tags/cv.md>), [lstm](<https://devfeed.tech/tags/lstm.md>), [machine](<https://devfeed.tech/tags/machine.md>), [ml](<https://devfeed.tech/tags/ml.md>), [nlm](<https://devfeed.tech/tags/nlm.md>), [rl](<https://devfeed.tech/tags/rl.md>), [sakana](<https://devfeed.tech/tags/sakana.md>), [sakanaai](<https://devfeed.tech/tags/sakanaai.md>), [tag-13e2af703774](<https://devfeed.tech/tags/tag-13e2af703774.md>), [tag-68e701e78517](<https://devfeed.tech/tags/tag-68e701e78517.md>)

### AI overview

The article examines Sakana AI's experimental Continuous Thought Machine (CTM), an architecture that makes internal temporal dynamics and neural synchronization central to computation. It describes iterative reasoning through internal states and ticks, while noting that the model is resource-intensive, unstable without careful tuning, and not intended to replace transformers.

### Source excerpt

Аналитический центр red_mad_robot продолжает следить за архитектурными прорывами в мире AI. В этот раз -- экспериментальная модель от команды Sakana AI, которая предлагает мыслить не в терминах слоёв, а в терминах времени. Их Continuous Thought Machine (CTM) -- попытка встроить в нейросеть внутреннюю динамику, вдохновлённую человеческим мозгом. Разбираем, как устроена архитектура, что такое "внутренние тики" и зачем нейросети синхронизировать собственные мысли -- на примерах из CV, сортировки, Q&A и RL. Читать далее

## Teaching an LSTM to Play Beethoven

DevFeed: [Teaching an LSTM to Play Beethoven](<https://devfeed.tech/articles/teaching-an-lstm-to-play-beethoven-40104.md>)

Original publisher: [Read original article](<https://korbonits.com/blog/2015-07-01-lstm-beethoven/>)

Published: 2015-07-01T00:00:00Z

Content type: tutorial

Language: en

Sources: [Alex Korbonits](<https://devfeed.tech/sources/alex-korbonits.md>)

Topics: [MIDI](<https://devfeed.tech/topics/midi.md>), [Code](<https://devfeed.tech/topics/code.md>), [Lua](<https://devfeed.tech/topics/lua.md>), [GPU](<https://devfeed.tech/topics/gpu.md>)

Tags: [audio](<https://devfeed.tech/tags/audio.md>), [channel](<https://devfeed.tech/tags/channel.md>), [code](<https://devfeed.tech/tags/code.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [format](<https://devfeed.tech/tags/format.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [lstm](<https://devfeed.tech/tags/lstm.md>), [lua](<https://devfeed.tech/tags/lua.md>), [music](<https://devfeed.tech/tags/music.md>), [structure](<https://devfeed.tech/tags/structure.md>), [timing](<https://devfeed.tech/tags/timing.md>), [torch](<https://devfeed.tech/tags/torch.md>), [training](<https://devfeed.tech/tags/training.md>), [types](<https://devfeed.tech/tags/types.md>), [velocity](<https://devfeed.tech/tags/velocity.md>)

### AI overview

An experiment trains Karpathy's char-rnn LSTM on Beethoven piano sonata MIDI files, treating the raw MIDI bytes as a sequence. The author describes prepending a shared MIDI header to generated output so players could interpret it, and reports imperfect results containing occasional phrase-like and melodic fragments.

### Source excerpt

In June 2015, I trained Karpathy's char-rnn on a corpus of Beethoven piano sonata MIDI files. Here's what it sounds like -- and how I got it to play.