Channel

Градиентное погружение

@gradientdip

On this record: Growth · Engagement · What this channel posts · Reactions · Posts · Citations · Cite this entry

4,248subscribers

-3 since we began measuring on 7 August 2026

Risers and fallers across the register · movement among entries of 3,162–10,000.

Register entry

Telegram ID	-1001754513217
Type	Channel
Username	@gradientdip
Description	Обсуждаем новости, рассказываем про ML с кодом и колабом, выигрываем соревы 🏆 Контакты: @Cene655, @Cucu_LaPraline
Created	26 January 2022 — measured — cross-checked against a third-party dataset (ext.tg_channel)
First recorded	7 August 2026
Last confirmed live	10 August 2026
Measurements held	3
Confirmed unchanged	2 times, most recently 10 August 2026
On Telegram	t.me/gradientdip

Growth

3 measurements taken within a single day, net -3. Dots are measurements; the straight line between them is drawn to join them, not to claim we know the path taken in between — snapshots are recorded only when a count changes, so gaps mean “no change observed”, never “interpolated”. The vertical axis spans 4,248–4,251 and does not start at zero.

Measurement log — every subscribers count we have recorded
Measured (UTC)	Subscribers	Change
8 Aug 2026, 05:02	4,248	-3
7 Aug 2026, 06:30	4,251	no change
7 Aug 2026, 06:21	4,251	first reading

Engagement

15 posts held, back to 12 July 2023 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 5 pagesof Telegram’s post history, 20 posts per page.

Nothing published in the last 30 days. ERR and ER are rolling 30-day measures, so there is nothing to compute — we hold 15 posts for this entry, the most recent from 3 February 2025. An engagement rate over an empty window would be a number about nothing.

What this channel posts

Photos: 179
Videos: 11
Links: 164

Lifetime counters from Telegram’s own channel header, read 8 August 2026 — not the date at the top of this page, which is when the subscriber count was last read. Below Telegram’s rounding threshold, so these counts are exact.

Video runtime: 3s
Average length: 3s

Measured directly from 1 video with a duration reading, out of the posts we hold for this channel — not this channel’s whole posting history, only the sample this register has actually read. An exact reading to the second, taken from the post itself rather than from Telegram’s own rounded chrome, so it carries no ≈ mark.

Reaction mix

515 reactions across 15 posts, in 11 distinct kinds. The most used accounts for 40.6% of them.

Every reaction kind recorded on the sample, most used first
Reaction	Count	Share
🔥	209	40.6%
👍	204	39.6%
❤‍🔥	35	6.80%
❤	21	4.08%
🤡	17	3.30%
👎	9	1.75%
😐	8	1.55%
⚡	6	1.17%
🤯	4	0.777%
🐳	1	0.194%
🥰	1	0.194%

No sentiment is inferred, and none should be read in. This table is ordered by count and by nothing else. Emoji do not carry stable meaning across languages or communities — 🙏 is thanks in one channel and mourning in another — so we publish which ones were pressed and how often, and pass no judgement on what an audience meant by them.

Precision. Telegram publishes reaction counts per emoji and short-forms each one — 4.34K, 1.2M — so any single kind at or above 1,000 reaches us at three significant figures, and only counts below 1,000 are exact. The shares above are ratios of those figures and carry the same error. This is also why the total here can differ slightly from a reaction total printed elsewhere on the page: both are sums of the same rounded parts, taken over samples with different edges.

Coverage. Reactions were read on 15 of the 15 sampled posts in this sample. Summed by Telegram’s own count on each post — not by adding up the per-emoji breakdown above — those same posts carry 515reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.

Measured over the 15 most recent posts we hold, published 12 July 2023 to 3 February 2025, using the newest reading held for each. Telegram Stars are excluded: they are a payment, not a reaction, and they have their own section.

Recent posts

3 Feb 2025, 18:01 UTC≈7,060 views24 reactionsread 8 August 2026

Forwarded from @lovedeathtransformers

Инженер LLM (Оптимизация и RL Alignment) Стартап в области безопасности ИИ Чем предстоит заниматься: Дообучение и Оценка Sota llm, аттаки на blackbox модели Улучшение RL для аттак на модели, настройки моделей (PPO, RLHF, стабильность обучения). Бенчмаркинг и оценка качества моделей (ELO-метрики, alignment). Оптимизация инференса (vLLM, SGLang, TRT). Требования: Опыт работы с LLM (архитектуры, RL, alignment). …

🔥16❤6👍2

Signed Maxim Gerasimov

24 Oct 2023, 18:58 UTC≈17,700 views26 reactionsread 8 August 2026

🔥TensorRT-LLM 👉Ваш любимый нейронный ускоритель получил расширение! TensorRT - движок, который применяет оптимизации для эффективного использования NVIDIA-GPU в Deep Learning. Что интересного? 💡Интеграция с NVIDIA Triton Inference Server 💡Поддержка Multiple GPU 💡Python API (очень напоминает API HF) 💡Есть поддержка 17 моделей (с примерами): Blib2, LLaMa 1/2, GPT-like, StarCoder, Bert, etc 💡Поддержка квантизации: F…

🔥21👍4❤1

Signed Максим Герасимов

15 Oct 2023, 14:54 UTC≈12,300 views30 reactionsread 8 August 2026

Forwarded from @data_morningVideo

🧹Scrapper Инструмент для парсинга страниц с JS. Работает в докере, можно обращаться по апи, сложнее детектируется как бот, может парсить в режиме просмотра (только текст), может использовать прокси и многое другое. @data_morning Github

👍24🔥5❤1

Signed Максим Герасимов

6 Oct 2023, 20:44 UTC≈11,400 views138 reactionsread 8 August 2026

Forwarded from @complete_aiPhoto

🏅А кто сегодня молодец и взял Trending paper на Hugging Face, обогнав Google DeepMind и Carnegie Mellon? Правильный ответ: статья про нашу модель Kandinsky, которую мы выложили на arxiv. ⚡А среди отметивших статью Julien Chaumond (CTO Hugging Face) и Patrick von Platten (Team lead Diffusers)! Со вторым мы активно сотрудничаем по внедрению модели в их самый крупный фреймворк генеративных моделей diffusers (18.2k⭐ н…

🔥104❤‍🔥10👍8🤡6⚡5🤯4❤1

Signed cene655

6 Oct 2023, 08:16 UTC≈8,400 views20 reactionsread 8 August 2026

Forwarded from @dealerAI

Всем привет, мы строили строили и наконец построили! Первая супер-библиотека по spelling corruption SAGE. Работа была не простой, вместе с ребятами из команды AGI NLP- мои герои снова: @alenusch, @qwertysobaka, @go_bobert. мы сделали оч крутой инструмент. Он позволяет атаковать тексты при помощи добавления опечаток, перестановок и пр., а также делать обратное - фиксить их. Интересно? Го читать на хабр! https://ha…

👍14🔥3❤2🥰1

Signed Максим Герасимов

3 Oct 2023, 14:07 UTC≈6,660 views6 reactionsread 8 August 2026

Forwarded from @complete_ai

⚡️Началось открытое голосование за номинантов премии HighLoad++. Среди прочих достойных номинантов представлена и модель Kandinsky! Верю, что команда заслуженно сможет получить эту награду😉 Спасибо всем за фидбэки, участие в развитии модели и её применении, ценные советы🙏 Все ваши комментарии не остаются не учтёнными, мы всё внимательно собираем и постепенно стараемся исправлять. Голосовать

👍6

Signed cene655

24 Sept 2023, 11:55 UTC≈7,240 views8 reactionsread 8 August 2026

Forwarded from @awesome_dl

Всем привет 👋 Продолжаем серию постов под авторством @TimeEscaper, посвященную применению ML в Robotics. В данном рассказе мы осветили вопрос того, как применять языковые модели для обучения роботов: - Как использовать LLM для генерации reward? - Как переводить языковые команды в действия робота? - Как использовать LLM для генерации алгоритмов роботов? Приятного чтения 🔥

🔥7❤1

Signed Максим Герасимов

23 Sept 2023, 17:23 UTC≈5,810 views16 reactionsread 8 August 2026

Forwarded from @data_morningVideo

Сморите что нашел 👉DragDiffusion - модель, способная редактировать изображения на основе точек, заданных юзером. Есть примеры использования, руководство по настройке и запуску модели. @data_morning github примеры

👍11🔥4❤1

Signed Максим Герасимов

14 Sept 2023, 14:44 UTC≈5,510 views32 reactionsread 8 August 2026

Forwarded from @abstractDLPhoto

Longformer-tiny для русского языка Друг скрестил Longformer и BERT-tiny, получив крошечную модель с контекстом в 16к токенов. Детали танцев с бубном и обучения есть на хабре. А картинку нарисовала моя сестра ☺️ Хабр, Huggingface

👍23🔥8⚡1

Signed Максим Герасимов

28 Aug 2023, 15:19 UTC≈6,190 views25 reactionsread 8 August 2026

Forwarded from @dreamforge_tgPhoto

Релиз новой версии AutoChar! Версия 0.9 стучится в ваши двери с новыми функциями, исправлениями ошибок и обновлённым интерфейсом. Спасибо большому сообществу за сообщения о багах и предложенные решения! https://civitai.com/models/95923 https://github.com/alexv0iceh/AutoChar ### Новые функции: • Добавлена полная поддержка Dynamic Prompts! Наслаждайтесь полностью автоматическим улучшением всех ваших рандомизированн…

👍19🔥5❤1

Signed cene655

12 Aug 2023, 07:59 UTC≈5,780 views29 reactionsread 8 August 2026

Умеете обучать модели и поднимать API, но не понимаете, что происходит за пределами этих процессов? Никто не знал о GPT, пока она была просто ML-моделью без публично доступного API и продакшена, отмасштабированного на миллионы пользователей, теперь известного всем как ChatGPT. Любой, даже самый минимальный жизнеспособный продукт при таких масштабах будет содержать как ML-компоненты, так и микросервисы, базы данных,…

👍11😐8👎6🤡3🐳1

Signed cene655

8 Aug 2023, 14:00 UTC≈5,390 views40 reactionsread 8 August 2026

Forwarded from @ppilif_chanel

Офигенно! Документирование чуть ли не построчно :3 This is a collection of simple PyTorch implementations of neural networks and related algorithms. These implementations are documented with explanations, and the website renders these as side-by-side formatted notes. We believe these would help you understand these algorithms better. https://nn.labml.ai/?_gl=1*1l3fv1n*_ga*NDc0MzgxMjA0LjE2OTE1MDI4ODU.*_ga_PDCL9PHMH…

❤‍🔥25👍11❤2👎1🤡1

Signed Максим Герасимов

Showing the 12 most recent of 15 posts we hold for @gradientdip. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked ≈ was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.

Citation-graph rank

Citation-graph rank — 670,374 of 1,160,990entries in the measured graph. A weighted position computed from the forward and mention edges below — republished posts weigh more than named mentions — and recomputed periodically, over the whole graph. Published only as this ordinal position, never as a score: a position is a fact, and a score printed beside one channel’s name would read as a verdict this register does not make. The two counts beneath stay separate for the same reason mentions are never summed with forwards anywhere else on this page — a named-by count costs nothing to manufacture. The top 100 by this measure, or how it is computed.

Forward network

Republishes

Channels on the register whose posts this channel has forwarded.

Дата-Утренник
@data_morning · 2713 posts Complete AI
@complete_ai · 7,7912 posts AbstractDL
@abstractDL · 18,2211 post Awesome DL
@awesome_dl · 8691 post Dealer.AI
@dealerAI · 16,4211 post Нейрокузня | DreamForge
@dreamforge_tg · 1,0271 post Love. Death. Transformers.
@lovedeathtransformers · 24,7081 post что-то на DL-ском
@nadlskom · 4,9221 post Записки Ппилифа
@ppilif_chanel · 4,4141 post

Built only from forwarded posts we have actually read, on both sides. Coverage is early and deliberately incomplete: a missing link means we have not read the post that would prove it, never that the relationship does not exist. Counts are distinct forwarded posts observed, so they only ever go up as we read more.

Mentions

Named by 2 registered channels — every channel on the register whose own posts have named this one, by its current username or any other username it currently holds, merged from two separately captured readings of the same fact so a namer caught by only one of them is not missed and a namer both caught is not counted twice. A username this channel has since dropped is not matched — that handle may belong to someone else now, and crediting today’s namer to yesterday’s owner would misattribute it.

Named by

Channels on the register whose posts name this channel's handle.

Дата-Утренник
@data_morning · 2711 post Кто-то идёт к дипворку
@go2dpwk · 841 post

Names

Channels on the register whose handles appear in this channel's posts.

Дата-Утренник
@data_morning · 2713 posts Complete AI
@complete_ai · 7,7911 post Dealer.AI
@dealerAI · 16,4211 post

A mention is a weaker signal than a forward and is counted separately for that reason — naming a channel is not republishing it, and a handle in a post body is easy to place deliberately. The post counts beside each row below are distinct posts in which the handle appeared, from posts we have read on both sides — the “Named by N registered channels” figure above is a different count, of distinct NAMING CHANNELS rather than posts, and is not the sum of the rows under it.

Cite this entry

A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 8 August 2026 — this entry's latest reading, not the date you are reading this.

“Градиентное погружение” (@gradientdip), 4,248 subscribers as measured 8 August 2026. Telegram Register, tgregister.com/channel/gradientdip.

Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.