🧠 dpp-gpt v2.0 Base (135M)
(🇺🇸 English / 🇷🇺 Русский)
⚠️
Note: This is a base foundation model. It has not been fine-tuned for chat. For the chat-ready version, download dpp-gpt-v2.0-flash-135m.
This is a compact foundation language model trained from scratch on the Llama 3 architecture. It is designed for raw text completion.
⚙️ Model Details
- Architecture: Llama 3 (136.9M Parameters)
- Layers / Hidden Size / Heads: 16 / 768 / 12
- Type: Base (Pre-trained only)
- Format: GGUF
- License: Apache 2.0
💡 Fun Fact (Math in ChatML)
Although this is a Base model for raw text completion, the mathematical portion of the pre-training dataset was explicitly formatted using ChatML. Because of this, it can successfully solve basic math equations if prompted with standard <|im_start|>user tags!
For example:
<|im_start|>user
20 + 656<|im_end|>
<|im_start|>assistant
🇷🇺 Описание на русском
⚠️
Внимание: Это базовая (foundation) модель. Она не обучалась формату диалога. Если вам нужна чат-версия, скачайте dpp-gpt-v2.0-flash-135m.
Это базовая компактная языковая модель, обученная с нуля (from scratch) на архитектуре Llama 3. Предназначена для классического продолжения текста (text completion).
⚙️ Детали модели
- Архитектура: Llama 3 (136.9M параметров)
- Слои / Размерность / Головы: 16 / 768 / 12
- Тип: Base (Только Pre-training)
- Формат весов: GGUF
- Лицензия: Apache 2.0
💡 Интересный факт (Математика)
Хотя это базовая модель, математическая часть датасета при претрейне была размечена в формате ChatML. Из-за этого модель умеет решать базовые примеры, если обратиться к ней через теги <|im_start|>user, хотя формально она не проходила стадию дообучения инструкциям (SFT)!
Пример:
<|im_start|>user
20 + 656<|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.