Telegram RegisterThe public register of Telegram

Channel

КПД

@quant_prune_distill

On this record: Growth · Engagement · Reactions · Posts · Citations · Telegram's recommendations · Cite this entry

3,428subscribers

+18 since we began measuring on 6 August 2026

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

Register entry

Telegram ID-1001916228977
TypeChannel
Username@quant_prune_distill
CreatedBetween 1 April 2023 and 31 October 2023— estimated from Telegram’s id allocation, not measured. How this range is calculated.
First recorded6 August 2026
Last confirmed live12 August 2026
Measurements held5
Confirmed unchanged1 time, most recently 12 August 2026
On Telegramt.me/quant_prune_distill

Growth

3,4103,4283,4196 August 2026 — 3,410 subscribers6 August 2026 — 3,410 subscribers6 August 2026 — 3,414 subscribers9 August 2026 — 3,423 subscribers12 August 2026 — 3,428 subscribers6 August 202612 August 2026
5 measurements spanning 6 days, net +18. 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 3,407–3,431 and does not start at zero.
Measurement log — every subscribers count we have recorded
Measured (UTC)SubscribersChange
12 Aug 2026, 13:453,428+5
9 Aug 2026, 21:223,423+9
6 Aug 2026, 21:493,414+4
6 Aug 2026, 04:453,410no change
6 Aug 2026, 02:193,410first reading

Engagement

22 posts held, back to 29 June 2026the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 3 pagesof Telegram’s post history, 20 posts per page.

ERR · 30 days
73.9%
avg views ÷ 3,428 subscribers
Avg views / post
2,530
16 posts measured
Reaction rate
0.728%
reactions ÷ views · ER floor
Posts in window
16
of 22 held

ERR is average views per post over the last 30 days divided by subscribers, the definition TGStat uses, so this figure is comparable with the one you will see elsewhere. It falls structurally as a channel grows: a high ERR on a small channel and a low one on a large channel describe reach mathematics, not quality. We publish the figure and the sample it came from and pass no verdict on it.

ER is defined industry-wide as (forwards + reactions + comments) ÷ views— note the denominator is views, not subscribers. Telegram’s public web preview carries views and reactions but not forward or comment counts, so the reaction rate above is the reactions term only and is therefore a floor: the true ER for this channel is higher by an amount we have not measured and will not estimate.

What these figures were computed from
WindowRolling 30 days · latest post in window 7 August 2026
Posts held22 (29 June 20267 August 2026)
Views total40,508
Reactions total295
Forwards / commentsnot exposed by the public surface — not measured, not estimated
Readings taken7 Aug 2026, 23:15 UTC

Views are a single reading per post, taken at the time above. A post published in the last day or two is still accumulating views, which pulls the 30-day average down slightly. That is a property of the standard definition rather than a fault in it, so we keep the definition rather than “correcting” the number into something nobody can reproduce.

Precision. Telegram publishes view counts on its public widget in short form — 8.12K, 3.7M — so any reading at or above 1,000 reaches us rounded to three significant figures, and only counts below 1,000 are exact. Averages and rates derived from them are shown to the same precision rather than to the unit: a figure like 3,701,250 would assert digits nobody measured.

Reaction counts are published per emoji and rounded the same way, so a total below 1,000 is exact and a larger one is a sum that may carry a rounded component from each emoji above 1,000. Because it is a sum, it does not look rounded — read a large reaction total as three significant figures per contributing emoji rather than as the figure it prints.

Reaction mix

354 reactions across 20 posts, in 14 distinct kinds. The most used accounts for 30.2% of them.

Every reaction kind recorded on the sample, most used first
ReactionCountShareShare, drawn
😁10730.2%
8022.6%
🔥6819.2%
👍339.32%
😭164.52%
👏154.24%
🏆123.39%
🤣71.98%
👎51.41%
💯51.41%
😱20.565%
🥰20.565%
❤‍🔥10.282%
🥴10.282%

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 22 of the 22 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 423reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.

Measured over the 22 most recent posts we hold, published 29 June 2026 to 7 August 2026, 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

7 Aug 2026, 15:00 UTC643 views12 reactionsread 7 August 2026

99 usage limit resets in Codex, 99 usage limit resets Take one down and pass it around, 98 usage limit resets in Codex 98 usage limit resets in Codex, 98 usage limit resets Take one down and pass it around, 97 usage limit resets in Codex ... No usage limit resets in Codex, no more usage limit resets Go to chatgpt.com and buy a new subscription, 99 usage limit resets in Codex

👏12

7 Aug 2026, 09:25 UTC835 views52 reactionsread 7 August 2026
Photo

Лайфак по использованию агентов. Если не запускать GPT Sol и Claude Opus на любой чих, то подписка расходуется не так быстро.

😁48👎4

5 Aug 2026, 20:53 UTC≈1,040 views23 reactionsread 7 August 2026

Когда месяц еще только начался, а ты израсходовал на вайбкодинг почти всю месячную квоту. https://youtu.be/W7rnkwrgEm0?si=Ug5Ah5GYNy_Eqh5d

😁211🔥1

5 Aug 2026, 08:25 UTC≈1,010 views14 reactionsread 7 August 2026

🧪 Метод и эксперименты Основными задачи при разработке кернела были: • Уменьшить обьем CPU работы и CPU-GPU синхронизации • Максимально эффективно перекрывать вычисления и коммуникации Рассматриваются два типа dispatching Push-based dispatch (традиционный подход) 1. Router для каждого токена определяет его экспертов. 2. Для каждого токена выполняется: Отправить этот токен эксперту №7. То есть именно токен "толк

14

5 Aug 2026, 08:19 UTC≈1,030 views19 reactionsread 7 August 2026
Photo

Mixture-of-Kittens: our open-source MoE megakernel for NVL72s 📄 Блогпост 💻 Код Ребята из курсора реализовали MXFP8 мегакернел под названием Mixture-of-Kittens, где пофьюжены все операции для слоя смеси экспертов, специализированный под GB300 NVL72s. В отличие от MegaMoE от команды Дипсика 🐋, этот кернел не только под инференс, а еще и под обучение.

🏆12👍5🥰2

4 Aug 2026, 07:56 UTC≈1,150 views33 reactionsread 7 August 2026
File

📚 Презентация про квантизацию с моего выступления на нашей Hardware Efficiency Reading Group, где мы обсуждаем статьи, идеи и практические аспекты эффективного глубокого обучения. 🎥 Записи и 📋 программу прошлых семинаров можно найти здесь. 🗓️ Семинар проходит (почти) каждый понедельник с 13:00 до 14:00. 📢 Группа с анонсами и материалами — тут. Будем рады всем, кто интересуется эффективным глубоким обучением, GPU,

👍1711🔥5

3 Aug 2026, 14:47 UTC≈7,900 views17 reactionsread 7 August 2026

🚀 Лаба Song Han (одного из апостолов EfficientDL) из MIT выпустила репозиторий со скиллом для агентов, предназначенным для оптимизации кернелов под Hopper и Blackwell. 📚 Репозиторий содержит набор скриптов, которые позволяют агенту обращаться к различным базам знаний, блогам, PR'ам и другим артефактам, посвященным тензорным ядрам и архитектурным особенностям Hopper и Blackwell. 🛠️ В частности, в него заложены знани

🔥114👍2

31 Jul 2026, 20:54 UTC≈1,800 views29 reactionsread 7 August 2026

step: 11 loss: 9.8672 grad_norm: 2.8441 tps: 5,471 time/step: 2.99s tflops: 198,319.19 mfu: 63563.84% memory: 243.85GiB MFU > 60000% Такое даже Уроборосу не снилось)

😁17🤣7🔥5

30 Jul 2026, 13:27 UTC≈2,030 views21 reactionsread 7 August 2026

Серёжа сьел токены сына Зачем? Не объяснив причину Серёжа запускает клод, рой агентов И сьедает квоту его Зачем так ждать чего-то? Так сильно хотеть чего-то? Так сильно обливать сердце кровью? И в снег, и в метель, и в грозы Писать Деду Морозу Чтоб выслал новые токены А не какой-нибудь сраный пенал…

😭16👏3👎1🥴1

23 Jul 2026, 15:03 UTC≈2,780 views19 reactionsread 7 August 2026

Бывает и такое, что Reviewer #2 ставит тебе Accept. Но ревью настолько короткое и несодержательное, что Area Chair не примет во внимание...

😁14💯5

22 Jul 2026, 20:01 UTC≈2,750 views10 reactionsread 7 August 2026

Интересная по описанию либа Humming от InclusionAI. Позиционируется как легковесный, высокопроизводительный фреймворк с JIT-компилированными GEMM-операциями (под NVIDIA GPU). ⚙️ Он предлагает широкий ассортимент кернелов под разные конфигурации квантованных весов и активаций: • 🧮 Веса можно квантовать почти в любую целочисленную битность от 1 до 8, а также в разные варианты FP. • ⚡ Активации можно квантовать в FP1

10

21 Jul 2026, 20:52 UTC≈1,890 views7 reactionsread 7 August 2026

🧪 Метод и эксперименты В данной работе фокусируются на 2-битном weight-only-сжатии ризонящих моделей. В аппендиксе есть эксперименты с FP4, со сжатием активаций и KV-кэшей. Квантизуют модели Qwen3-8B / Qwen3-32B через GPTQ. Оказывается, что квантизация сильно меняет поведение трейсов ризонинга: 🔄 Число циклов резко возрастает, особенно для меньшей модели. 📏 Многие трейсы не вписываются в заданный лимит токенов. ⚠️

7

Showing the 12 most recent of 22 posts we hold for @quant_prune_distill. 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 — 103,011 of 1,176,251entries 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

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.

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.

Appears in Telegram’s recommendations for other channels

The reverse of the list above, and a different kind of signal. This does not require this channel to have ever been asked about directly — each row below is a channel we DID ask Telegram about, whose Telegram-generated list happened to include this one. A channel can appear here with an empty list above it, because being named by someone else’s query is independent of having been queried itself.

Machinelearning
@ai_machinelearning_big_data · 285,981
Telegram ranks this channel #43 of 95 here — alongside 94 others — read 10 August 2026

This channel appears in 1 seed channel's Telegram-generated recommendation list in total. Each is Telegram’s list for THAT channel, not this one — see how this is measured.

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 12 August 2026 — this entry's latest reading, not the date you are reading this.

“КПД” (@quant_prune_distill), 3,428 subscribers as measured 12 August 2026. Telegram Register, tgregister.com/channel/quant_prune_distill.

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.