ruGPT training and fine-tuning examples
Inspect Russian-model fine-tuning and pretraining scripts, with explicit warnings about the legacy environment.
The ruGPT repository pairs fine-tuning notebooks with pretraining scripts and the SberDevices team’s training notes. Trace how text files, context length and evaluation enter the workflow. This is a legacy implementation reference, not a current turnkey recipe: dependency pins, external datasets and GPU assumptions must be rechecked. The main README is English despite the Russian-language models.
- Level
- Advanced
- Models
- rugpt
- Category
- Model Training
- Tags
- practical-guide
- Source date
- date not stated
- Added
- 2026-10-18
- Last verified
- 2026-10-18
- Status
- ok
- Prerequisites
- Requires experience with transformer training, GPU environments and dependency compatibility to assess and adapt legacy scripts.
- Subcategory
- fine tuning
- Region
- RU
- Original language
- en
- Access
- free
- Resource reuse
- open: Repository code carries Apache-2.0. Check individual model and dataset terms separately. Facet definitions
Summary in our own words. Rights remain with the original creator. Credit and correction policy.