metadata
license: mit
language:
- ru
base_model:
- VGribAI/Tensor-28m
Tensor-28m
Our first model using RLHF—Tensor-28m!!! The Tensor-28m-sft model, unlike the base version, has a better understanding of context and is trained to respond to certain standard phrases.
About Us
Tensor-28m is our second successfully trained model, built on the Russian 8 MB dataset.
You are free to use, refine, and experiment with the model.