Instructions to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M # Run inference directly in the terminal: llama cli -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M # Run inference directly in the terminal: llama cli -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
Use Docker
docker model run hf.co/dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
- LM Studio
- Jan
- vLLM
How to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "dikiyplayerpig/dpp-gpt-V2.1-Pro-260m" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "dikiyplayerpig/dpp-gpt-V2.1-Pro-260m", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
- Ollama
How to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with Ollama:
ollama run hf.co/dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
- Unsloth Studio
How to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for dikiyplayerpig/dpp-gpt-V2.1-Pro-260m to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for dikiyplayerpig/dpp-gpt-V2.1-Pro-260m to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for dikiyplayerpig/dpp-gpt-V2.1-Pro-260m to start chatting
- Docker Model Runner
How to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with Docker Model Runner:
docker model run hf.co/dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
- Lemonade
How to use dikiyplayerpig/dpp-gpt-V2.1-Pro-260m with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull dikiyplayerpig/dpp-gpt-V2.1-Pro-260m:Q4_K_M
Run and chat with the model
lemonade run user.dpp-gpt-V2.1-Pro-260m-Q4_K_M
List all available models
lemonade list
- Atomic Chat
🧠 dpp-gpt v2.1 Pro (260M)
(🇺🇸 English / 🇷🇺 Русский)
This is a major upgrade in the dpp-gpt family. Version 2.1 Pro is a 259M parameter language model trained entirely from scratch. It features significant improvements in logic, multi-lingual translation, and utilizes the [THINK] token for Chain-of-Thought (CoT) reasoning.
⚙️ Model Details
- Parameters: 259M
- Layers / Hidden Size / Heads: 20 / 1024 / 16
- Context Length: 4096 tokens
- Vocabulary Size: 16,384
- Format: GGUF / PyTorch (.pth)
- License: Apache 2.0
📊 Training Data
- Pre-training: 11.8 Billion tokens (~45.5 tokens/parameter) with a batch size of 512k.
- Fine-Tuning (SFT): >16.5M high-quality tokens generated primarily by Gemma 4 (26b/12b/4b), Qwen 3.5 (35b/4b), and complex code from DeepSeek v4 Flash.
🚀 Capabilities & Advantages
- Languages & Translation: Excellent comprehension of Russian, English, and French. Capable of translating simple phrases between these languages seamlessly.
- Text Processing: Strong text manipulation skills. It can spell words letter-by-letter, assemble words from spelled-out letters, count total letters in a word, and count specific letters.
- Math & Logic: Solves arithmetic operations (
a + bup to hundreds of thousands,a + b + c,a + b + c + dfor addition/subtraction), simple linear equations, and basic math word problems using step-by-step reasoning. - Creative Writing & Chat: Consistently generates structured essays, writes poems, and maintains natural dialogue.
- Coding: Generates basic functional Python code (significantly improved over v2.0).
💡 Prompting & System Prompt
The model uses a strict ChatML format. (Note: The 4-bit quantized version of this model understands and follows System Prompts noticeably better than other versions).
Standard Mode (No thinking):
<|im_start|>user
[NOTHINK] {prompt}<|im_end|>
<|im_start|>assistant
Reasoning Mode ([THINK] token):
To force the model to "think" and use logic before answering, modify the prompt template. If you are using LM Studio, simply type . or [THINK] right before your prompt (without a space).
<|im_start|>user
[THINK] {prompt}<|im_end|>
<|im_start|>assistant
🇷🇺 Описание на русском
Это масштабное обновление линейки dpp-gpt. Версия 2.1 Pro — это модель на 259М параметров, обученная полностью с нуля. Версия отличается значительным улучшением логики, качественным мультиязычным переводом и использует токен [THINK] для пошаговых рассуждений.
⚙️ Детали модели
- Параметры: 259M
- Слои / Размерность / Головы: 20 / 1024 / 16
- Контекст: 4096 токенов
- Словарь: 16,384 токена
- Формат весов: GGUF / PyTorch (.pth)
- Лицензия: Apache 2.0
📊 Данные для обучения
- Pre-training: 11.8 млрд токенов (~45.5 токенов/параметр, батч 512k).
- Fine-Tuning (SFT): >16.5 млн высококачественных токенов, сгенерированных в основном Gemma 4 (26b/12b/4b), немного Qwen 3.5 (35b/4b) и сложным кодом от DeepSeek v4 Flash.
🚀 Особенности и навыки
- Языки и Перевод: Отличное понимание русского, английского и французского языков. Уверенный перевод простых предложений между этими языками.
- Работа с текстом: Отличная работа со структурой слов. Разбор слов побуквенно, сборка слов из побуквенного написания, подсчет всех букв в слове, подсчет конкретной буквы в слове.
- Математика и Логика: Решение примеров вида
a + b(до сотен тысяч),a + b + c,a + b + c + d(только сложение и вычитание). Решение простых линейных уравнений и простых текстовых задач с использованием логики (Chain-of-Thought). - Творчество и диалог: Написание структурированных сочинений, стихов, поддержание адекватного диалога.
- Код: Написание базового функционального кода на Python (существенный шаг вперед по сравнению с 2.0).
💡 Шаблоны промпта и Системный промпт
Модель использует формат ChatML. (Примечание: 4-битная версия модели (Q4) справляется с пониманием системного промпта заметно лучше остальных квантований).
Стандартный шаблон (Без размышления):
<|im_start|>user
[NOTHINK] {запрос}<|im_end|>
<|im_start|>assistant
Режим размышления (Токен [THINK]):
Для включения пошагового обдумывания нужно использовать соответствующий тег. При запуске через LM Studio достаточно просто написать . или [THINK] прямо перед началом вашего запроса (без пробела).
<|im_start|>user
[THINK] {запрос}<|im_end|>
<|im_start|>assistant
Hotfix
The GGUF files in this repository were re-uploaded as a hotfix.
An LM Studio update changed the way special tokens are handled, which made the previously published GGUF files generate broken output. The conversion has been fixed and the quantizations here were rebuilt from the corrected model. The weights are unchanged — only the token metadata inside the GGUF files.
If you downloaded a GGUF from this repository before this commit, please download it again.
- Downloads last month
- 154
2-bit
4-bit
6-bit
8-bit
16-bit