Skip to content

Conversation

dsikka
Copy link
Collaborator

@dsikka dsikka commented Sep 8, 2025

Summary

  • More generic handling when running fp4 compression
  1. If group / tensor_group / channel - nvfp4 - packed quantized
  2. If not and running w4a4 - float-quantized
  3. Otherwise, naive-quantized

@dsikka dsikka changed the title [FP4] Update to make compression handling to be more generic for fp4 [FP4] Update to make compression handling more generic for fp4 Sep 8, 2025
Base automatically changed from update_format to main September 8, 2025 22:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant