CoolFace
Modelpublic

RTU-MIREA/Ancia

sourceHugging Faceupdated 10mo agoView on Hugging Face
0likes
Model Card

[Will be released soon...]

Ultra small GPT model, trained on different sources from scratch.

254m parameters, 71% Russian datasets.

Trained on:

  • —AMD Ryzen Threadripper PRO 5955WX
  • —64GB RAM
  • —Nvidia RTX 4090 (24GB) x3