Skip to content

Main#227

Closed
Tervezoo wants to merge 2 commits into
TheTom:feature/turboquant-kv-cachefrom
Tervezoo:main
Closed

Main#227
Tervezoo wants to merge 2 commits into
TheTom:feature/turboquant-kv-cachefrom
Tervezoo:main

Conversation

@Tervezoo

Copy link
Copy Markdown

Overview

Additional information

Requirements

Tervezoo added 2 commits July 18, 2026 16:21
- Added GGML_TYPE_Q4_K_XL = 47 type definition
- Added mmq.cu switch case for Q4_K_XL mul_mat
- Added convert.cu dequantize_row_q4_K_cuda mapping for Q4_K_XL
- Enables GPU acceleration for Q4_K_XL quantized models

Assisted-by: Hermes Agent
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant