Blind — Russian-language LLMs, trained from scratch
A series of models from 50M to 750M parameters, created entirely from scratch and trained on Russian-language data.
A project for enthusiasts, students, and developers who want to experiment with LLMs without a huge investment.
I will continue training for 2 languages RU and EN and also increase the model if there are 100 downloads
Models
| File | Size | Parameters |
|---|---|---|
blind-0-ultrasmall.pt |
173 MB | ~50M |
blind-0-small.pt |
356 MB | ~85M |
blind-0-medium.pt |
407 MB | ~100M |
blind-0-large.pt |
445 MB | ~110M |
blind-1.pt |
1 GB | ~740M |
blind-1.1.pt |
1 GB | ~750M |
Install from pip and serve model
# Install vLLM from pip: pip install vllm# Start the vLLM server: vllm serve "debugdll/Blind"# Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "debugdll/Blind", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'