/ Модель

Характеристики

Лицензия
Apache 2.0
Библиотека
nanochat
Опубликована
6 окт 2026

queen-v3

Генерация текстаТекст → текст

Небольшая GPT-модель Crawlnet для генерации текста и ответов на вопросы о криптовалютах, блокчейнах и DeFi.

339скачиваний за 30 дней1лайквеса обновлены6 окт 2026

Бенчмарки

Как модель справилась с задачами пользователей ailist и что о ней публикуют открытые лидерборды

Публичные оценки

Из открытых лидербордов. Сравнивайте результаты одного теста при одинаковых условиях — отсутствие оценки не значит, что модель хуже.

В подключённых источниках результатов для этой модели пока нет.

Что умеет модель

Описание составлено по карточке модели и метаданным Hugging Face

Queen v3 «weaver» — небольшая языковая модель Crawlnet, обученная с нуля на архитектуре GPT из nanochat. В ней 122,6 млн параметров, контекст — 2048 токенов. Для предобучения использовались страницы криптовалютной тематики, собранные краулерами, и текстовые записи о рынках, финансировании и ончейн-данных. Модель также дообучена на парах вопросов и ответов, созданных из этих материалов.

Модель предназначена для генерации текста и ответов на вопросы о криптовалютах, блокчейнах, токенах и DeFi. В карточке приведены примеры ответов на английском и других языках, но основной язык данных и тег языка модели — английский. Лицензия — Apache-2.0.

Модель мала и обучалась на узком корпусе; ответы могут быть повторяющимися, бессвязными или уверенно ошибочными. Она может выдумывать цены, даты, адреса и ссылки, поэтому её нельзя использовать как источник фактов или финансовых рекомендаций.

Сценарии применения

  • черновые ответы на вопросы о криптовалютах
  • объяснение базовых понятий блокчейна
  • подготовка текстов о токенах и криптопроектах
  • ответы на вопросы о DeFi и ончейн-данных

Ограничения

Модель мала и обучалась на узком корпусе, поэтому ответы могут быть повторяющимися, бессвязными и фактически неверными. Она может выдумывать цены, даты, адреса и ссылки; не используйте её как источник фактов или финансовых рекомендаций.

Карточка модели

README от автора на Hugging Face, как он его опубликовал

Открыть на Hugging Face

Queen v3 “weaver” — Crawlnet

A small language model trained from scratch. Pretraining data and tokenizer: only pages the Crawlnet crawlers read on the crypto web (real browsers), plus public market, funding and on-chain records written out as plain-text pages (6% of the tokens, see below). No pretrained weights, no FineWeb, no other corpus. Chat tuning: question/answer pairs generated by deepseek-flash from those pages (details below). Expect a dumb, funny, confidently-wrong model; that is the point.

version3 (queen-v3)
parameters122.6M total, 59.6M non-embedding
architecturenanochat GPT, depth 10, width 640, 5 heads, context 2048, vocab 16384
codekarpathy/nanochat @ 92d63d4e8bb4 (MIT) + crawlnet training/
dataset135,077 pages exported 2026-10-06T11:30:31+00:00; after cleaning 134,994 pages: 89,367 crawler pages from 1936 domains + 45,627 record pages
dataset manifest sha25600a69452525b0533fdddde5ff0f345ac19d451cd5cc842b4711a0cfd833edea3
unique training tokens191,488,840 (+1,996,669 held out); the record pages are about 10,950,432 tokens of it (6%)
tokens trained625,999,872 = 3.27 epochs, 10.5 tokens per non-embedding param
tokenizerBPE, 16384 tokens, trained on this dataset's pages only (crawler pages and record pages)
validation loss0.8174 bits/byte on held-out pages (pretraining)
chat tuning (SFT)synthetic+openbook:deepseek-flash: 247,648 question/answer pairs generated by deepseek-flash (DeepSeek-V4.1-Flash (open weights, MIT license; DeepSeek API terms 4.2(3) allow training other models on outputs), via api.deepseek.com) from 170,994 page chunks (crawler pages and record pages), plus the pairs cut from the pages. The generator was told to use only facts on each page, but its wording is in this model's chat behaviour. headings: 487,456 question/answer pairs cut out of the crawler pages themselves (FAQ questions found on pages, page headings, page titles; answers are verbatim page text). No external model was used. The only human-written words are six fixed question templates such as "Tell me about {heading}." Open book: the generated pairs were also trained once more with the page text they were written from in front of the question (183,204 conversations, 46% with a second page the search ranked high for the same question), so she can answer from the pages retrieved for a question.
compute1 GPU (our GPU server), 1.771 h wall, 1.771 GPU-hours
cost$0.60 GPU + ~$36.33 deepseek-flash API for the chat-tuning pairs (list price)
weightshttps://huggingface.co/Crawlnet/queen-v3

Contributors

Queen v3 was trained on pages found by these crawlers. Owners spawned theirs by burning $CRAWLNET; the wallets are the ones that made the burn, public on-chain.

crawlerownerpagestokens
pp3JVm…Fsm4 (burn)6551,649,028
spiderman5WKN…Meic (burn)7691,634,545
smokie8kwY…JY7G (burn)7211,631,346
crawlzilla3dmq…dLBC (burn)7561,627,782
ghero5FcEE…Cin4 (burn)6901,622,127
pp3JVm…Fsm4 (burn)6061,586,243
azeBX6f…WpEf (burn)7241,541,422
spiderman5WKN…Meic (burn)7351,521,654
ch0pperH8iW…VE3N (burn)6901,514,921
aze2BX6f…WpEf (burn)7341,514,315
werkle-weaver3api…NA9T (burn)6761,508,869
mrnakamoto2eek…pBky (burn)7441,502,714
testzero…PPUT (burn)7741,498,672
ghero4FcEE…Cin4 (burn)7261,489,409
spiderman5WKN…Meic (burn)7491,477,210
cashH2Lt…LbYp (burn)5801,475,211
onepellegrino2eek…pBky (burn)7291,471,774
jdubs3v2K…GTgA (burn)6071,466,629
charlotte2eek…pBky (burn)7361,454,450
fern2eek…pBky (burn)7061,446,090
v4GZUQ…F59x (burn)7221,443,046
spiderman5WKN…Meic (burn)7241,441,441
blackandwhitedogs2eek…pBky (burn)7391,425,614
ghero10FcEE…Cin4 (burn)6821,396,780
ruso4rif…jxnZ (burn)7311,396,359
littledog2eek…pBky (burn)7371,395,448
ghero6FcEE…Cin4 (burn)6361,392,838
redacted002BvVr…URcY (burn)6351,386,664
near-weaverBr2o…hHFG (burn)7061,386,197
sparsityGZUQ…F59x (burn)7211,380,285
yancyDPXE…XNqJ (burn)6601,372,115
ghero3FcEE…Cin4 (burn)6951,367,591
manylegsETgB…wfpa (burn)6281,362,227
seig4gtD…rJMz (burn)6281,355,402
ghero8FcEE…Cin4 (burn)6761,352,758
wilbur2eek…pBky (burn)7011,348,588
swolecrawlsF3QF…Dfrs (burn)6971,342,354
benny36ny…e6gG (burn)6221,342,326
ch0pper2H8iW…VE3N (burn)6251,331,559
dashH2Lt…LbYp (burn)6341,323,652
warm6TAx…SE4b (burn)6751,320,192
jackDiu8…atMh (burn)6711,310,227
bigdog2eek…pBky (burn)7001,306,680
lmaocryH6GW…Hmdr (burn)6471,302,177
babushkaCj9h…Cf4C (burn)6731,280,835
redacted003BvVr…URcY (burn)6511,278,545
gigachadfuruAbMD…52T4 (burn)6551,269,907
weaverF5Vd…3hVn (burn)6321,255,923
ghero9FcEE…Cin4 (burn)6731,253,976
deezCDYb…2Xxv (burn)7431,252,961
spidermanH8iW…VE3N (burn)7521,250,842
smokiev28kwY…JY7G (burn)6561,244,408
spiderman5WKN…Meic (burn)6191,243,629
spy8zQb…eogQ (burn)5791,242,732
ghero7FcEE…Cin4 (burn)6711,237,701
spidermanFGyn…UuQt (burn)6151,228,981
spiderman5WKN…Meic (burn)6911,222,845
ghero2FcEE…Cin4 (burn)6311,214,378
redacted001BvVr…URcY (burn)6421,208,395
e02AcUm…vLRa (burn)5741,203,136
turtletheoryipMd…U4xb (burn)6001,193,175
morelegsETgB…wfpa (burn)6311,171,977
ghero1FcEE…Cin4 (burn)6801,170,013
weenis36ny…e6gG (burn)5651,156,575
octangone01G64q…5PiU (burn)5351,146,451
vinland3cSP…oYeU (burn)6051,137,549
sneaky6yNt…iikJ (burn)5231,128,930
orb-weaver3Zgo…wfpF (burn)4451,121,301
web-reaperKrrc…fSaT (burn)6571,117,543
aze4BX6f…WpEf (burn)4861,107,142
peterparkerAbMD…52T4 (burn)6421,091,238
aze5BX6f…WpEf (burn)4841,088,067
oreocakester36ny…e6gG (burn)6221,082,914
e05AcUm…vLRa (burn)5651,081,442
e04AcUm…vLRa (burn)5501,076,001
jeff81Y8…GXSt (burn)6571,062,912
007ipMd…U4xb (burn)4881,053,236
1gjbAyXY…HcDw (burn)4941,044,150
canisFUy5…cjoG (burn)4691,037,454
page-pounderExtb…VrFG (burn)5501,025,958
e01AcUm…vLRa (burn)5221,023,918
kk1438PRb…2Ufe (burn)5461,010,842
41nEXTV…831R (burn)512997,888
e03AcUm…vLRa (burn)518995,799
ruso-24rif…jxnZ (burn)495986,696
needmorecrawls4rif…jxnZ (burn)518973,364
elevatd3MRp…31K7 (burn)571965,189
aze3BX6f…WpEf (burn)494950,331
aze6BX6f…WpEf (burn)497942,707
snoopdogDRmh…ow3V (burn)488939,193
matrix-multiplier6s1H…36Jg (burn)324926,865
aspen41CL…Q9fU (burn)465923,557
crawlzilla-23dmq…dLBC (burn)485920,025
funnelweb1DmsF…uvrF (burn)482894,026
onaizahparisnajd4KZT…HerD (burn)315879,329
41nEXTV…831R (burn)480877,520
chungus-spawnH7oG…CsLj (burn)474873,686
redactedBvVr…URcY (burn)304869,415
tarantulaBkPa…YiMW (burn)508856,717
smokiev38kwY…JY7G (burn)462843,648
onaizah4KZT…HerD (burn)428832,589
grand-couFGcD…k1Dc (burn)563832,183
son-of-chungusH7oG…CsLj (burn)458814,445
onaizahparisnajd4KZT…HerD (burn)300811,509
onaizahparisnajd4KZT…HerD (burn)295787,528
creepycrawly3v2K…GTgA (burn)241786,130
9ballshooterAPM9…89Nm (burn)240777,586
onaizahparisnajd4KZT…HerD (burn)318753,269
redactedBvVr…URcY (burn)296743,433
onaizahparisnajd4KZT…HerD (burn)309739,697
onaizahparisnajd4KZT…HerD (burn)287719,159
redactedBvVr…URcY (burn)319708,897
goatishere258pZ…waHQ (burn)307704,407
redactedBvVr…URcY (burn)307703,071
redactedBvVr…URcY (burn)286700,234
redactedBvVr…URcY (burn)307689,288
onaizahparisnajd4KZT…HerD (burn)276682,222
redactedBvVr…URcY (burn)286660,308
onaizahparisnajd4KZT…HerD (burn)315654,837
redactedBvVr…URcY (burn)299650,036
redactedBvVr…URcY (burn)293648,020
onaizahparisnajd4KZT…HerD (burn)301647,437
redactedBvVr…URcY (burn)294642,916
onaizahparisnajd4KZT…HerD (burn)291638,567
onaizahparisnajd4KZT…HerD (burn)298632,122
addu4gff…ksBn (burn)308628,534
onaizahs-kids4KZT…HerD (burn)271617,561
onaizahskdis4KZT…HerD (burn)301613,835
onaizahparisnajd4KZT…HerD (burn)308613,439
wei0019ptc…3LbM (burn)198607,746
callmeH5y3…BR5q (burn)305601,084
onaizahparisnajd4KZT…HerD (burn)305593,401
onaizahparisnajd4KZT…HerD (burn)299590,157
9ballshooterAPM9…89Nm (burn)241589,726
onaizahparisnajd4KZT…HerD (burn)296586,169
onaizahparisnajd4KZT…HerD (burn)295584,413
gummy8zqX…Gv6M (burn)308578,724
gg4gff…ksBn (burn)300578,301
redactedBvVr…URcY (burn)297571,411
goatishere58pZ…waHQ (burn)304571,280
redactedBvVr…URcY (burn)323570,936
redactedBvVr…URcY (burn)301565,677
bimboboy58pZ…waHQ (burn)298562,195
gg4gff…ksBn (burn)306560,534
wei0089ptc…3LbM (burn)229556,394
dennCQFL…PBVr (burn)318554,954
gg4gff…ksBn (burn)297547,969
onaizahparisnajd4KZT…HerD (burn)279546,142
onaizahparisnajd4KZT…HerD (burn)245540,650
redactedBvVr…URcY (burn)302538,080
onaizahparisnajd4KZT…HerD (burn)306537,684
kaiyo4gff…ksBn (burn)295529,359
onaizahparisnajd4KZT…HerD (burn)287519,828
redactedBvVr…URcY (burn)300517,834
onaizahparisnajd4KZT…HerD (burn)287499,311
denn2CQFL…PBVr (burn)223484,754
redactedBvVr…URcY (burn)209482,906
historyG8zB…gepY (burn)234473,438
wei0099ptc…3LbM (burn)214467,048
widowmakerGRjD…spY8 (burn)237466,377
neet18X78…cmmd (burn)203448,475
wakemeH5y3…BR5q (burn)240440,659
wei0019ptc…3LbM (burn)226417,907
wei0049ptc…3LbM (burn)200414,214
redactedBvVr…URcY (burn)223398,578
wei0029ptc…3LbM (burn)192395,936
laquisha3v2K…GTgA (burn)241387,955
wei0059ptc…3LbM (burn)207384,339
charl0tteFtgb…9DWW (burn)182367,684
wei0109ptc…3LbM (burn)207350,115
wei0069ptc…3LbM (burn)178309,394
wei0079ptc…3LbM (burn)174278,748
neet28X78…cmmd (burn)185244,963
snaiduwa4KZT…HerD (burn)3059,816
snaiduwa4KZT…HerD (burn)117,868
the queen's own crawlers (10)—7,36415,115,933

Pages and tokens count what made it into this version's dataset after cleaning and dedupe. Machine-readable: contributors.json. The 45,627 record pages (see below) were not found by a crawler and are not in this table.

What went in besides crawler pages

  • The nanochat training code, architecture and hyperparameter rules (no weights, no data).
  • Six hand-written question templates used to turn page headings into questions (see SFT above).
  • Record pages: 45,627 pages, about 10,950,432 tokens, of public market, funding and on-chain records (projects and their funding rounds, investors, token unlocks and token facts, DeFi protocols, chains and stablecoins, DEX and on-chain token statistics), written out as plain text from public data. Only factual fields were kept: no third-party labels, ratings or risk scores, and no social-media posts. They are in the pretraining data, the tokenizer and the generated chat pairs; the pairs cut from page headings use crawler pages only.
  • Chat-tuning question/answer pairs generated by deepseek-flash from crawler pages and record pages (see SFT above).
  • Nothing else. nanochat's default SFT mixture (SmolTalk, MMLU, GSM8K) and its CORE eval were switched off.

What it says

Fixed prompts, seeded sampling (temperature 0.7, top-k 50), unedited:

What is Bitcoin?

Bitcoin (BTC) je skratka računalnom šifrev in spletne i zaščitne svoje denarja, ki so se na ključe strežnim denarjem. Bitcoin je zaščit neboli računalnom šifrev in spletne i zaščitne svoje denarja (a nekom jer javne stroške vrstniški) in poslovnega svo

Who created Bitcoin?

Bitcoin was created by Satoshi Nakamoto, a pseudonym used by the pseudonymous creator of the cryptocurrency. Nakamoto also published the Bitcoin whitepaper in 2008.

What is a blockchain?

A blockchain is a distributed database of all digital assets that can be accessed and verified by anyone in the world to be globally distributed. It is a secure, transparent, and decentralized database that stores all digital assets in a decentralized database.

Explain proof of work.

↑ "What are "algorithmic" cryptocurrencies?". Investopedia. Retrieved 2025-08-12. ↑ "Bitcoin - Economics - Bitcoin". bitcoin.org (in Engels).

Explain proof of stake.

In proof of stake, the validators that take turns producing blocks are called validators. In short, the validators that take turns producing blocks are called validators. In some PoS systems, validators are selected based on the amount of cryptocurrency they have staked. In the example given above, the validators can be selected based on the amount of cryptocurrency they are staking. For example, if there are 500,000 validators, they are selected based on the amount of cryptocurrency they stake.

What is Ethereum?

Ethereum – kurz ETH (jenis 100.000.000 ETH) – kurz ETH (jenis 100.000.000) – kurz ETH (jenis 100.000.000 ETH) – kurz ETH (jenis 100.000.000) – kurz ETH (jenis 100.000.000) – kurz ETH (jenis 100.000.000) – kurz ETH (jenis

What is a smart contract?

A smart contract is a program stored on the blockchain that executes when certain predefined conditions are met. It is the backbone of any self-executing computer program, and it is self-executing.

What is gas on Ethereum?

Gas is the unit that measures the computational cost of performing operations on Ethereum, including computational resources. Gas is a unit of measurement for the computational effort required to execute a transaction or smart contract.

What is Solana?

Solana is a decentralized computing platform that aims to improve blockchain scalability by using a consensus mechanism called proof of stake (PoS). It allows for low transaction costs and fast speeds. Solana is different from many other blockchains that the network has its own governance token, SOL, which can be bought and traded on various cryptocurrency exchanges.

Why is Solana fast?

Solana is fast. Solana is one of the fastest blockchain networks in crypto, offering speeds up transaction times and low fees. It is also scalable, with the ability to process around 65,000 transactions per second (TPS). It is also known for its low fees. Solana achieves this by using Proof of History (PoH) in its consensus mechanism. PoH is a cryptographically secure clock that timestamps transactions. Solana is fast with a block time of 400 milliseconds (i.e., a few seconds) and confirmation times of 300 milliseconds (i.e., a few

Use it

chat.py in this folder needs only karpathy/nanochat (MIT) at the commit above:

pip install torch tiktoken safetensors
git clone https://github.com/karpathy/nanochat && git -C nanochat checkout 92d63d4e8bb4
PYTHONPATH=nanochat python chat.py "What is staking?"

Or load it yourself: config.json has the nanochat GPTConfig; build nanochat.gpt.GPT, load model.safetensors; the tokenizer is tokenizer.tiktoken (tiktoken rank file) + the pattern and special tokens in tokenizer.json. Chat format: <|bos|><|user_start|>...<|user_end|><|assistant_start|>.

Limitations

Tiny model, narrow data. It makes things up, including prices, dates, addresses and links. It is not financial advice and must not be used as a source of facts.

Как запустить

На своей GPU в ailist. Цена — по текущим предложениям провайдеров

GPU в ailist

Универсальный рантайм · nanochat

Нельзя запустить

Архитектура ? неизвестна transformers в образе рантайма.

Ревизия весов
ceed523

0 моделей