# Openai-whisper failing on larger models (NVIDIA / CUDA issues)

**URL:** https://discourse.nixos.org/t/openai-whisper-failing-on-larger-models-nvidia-cuda-issues/58074
**Category:** Help
**Created:** [December 30, 2024, 7:36pm UTC](https://discourse.nixos.org/t/openai-whisper-failing-on-larger-models-nvidia-cuda-issues/58074 "2024-12-30T19:36:18Z")
**Posts on this page:** 1
**Showing post:** 14

<div class="post-metadata">

### Author: ![NestorLiao](https://discourse.nixos.org/user_avatar/discourse.nixos.org/nestorliao/32/16720_2.png) [@NestorLiao](https://discourse.nixos.org/u/NestorLiao)
#### Post date: [January 3, 2025, 4:37pm UTC](https://discourse.nixos.org/t/openai-whisper-failing-on-larger-models-nvidia-cuda-issues/58074/14 "2025-01-03T16:37:15Z")

</div>

I don’t know why I didn’t the same thing, it error me:  
` error: triton-2.1.0 not supported for interpreter python3.12`  
what’s your default python3 version.  
I set up [nix-community.cachix.org](http://nix-community.cachix.org) according to [CUDA Cache for Nix Community](https://discourse.nixos.org/t/cuda-cache-for-nix-community/56038) , and use  
python311  
python311Packages.torch-bin  
python311Packages.openai-whisper  
and **cuda still compile for hours**.

---

_[View the full topic](https://discourse.nixos.org/t/openai-whisper-failing-on-larger-models-nvidia-cuda-issues/58074)._