[PR #30] [MERGED] [feature] support q2_k & q3_k dequantize on gpu #1323

New issue

Closed

opened 2026-10-29 14:51:26 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-29 14:51:26 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/kvcache-ai/ktransformers/pull/30
Author: @BITcyman
Created: 8/12/2024
Status: ✅ Merged
Merged: 8/12/2024
Merged by: @UnicornChan

Base: develop-0.1.2 ← Head: feature-lc-dequantize_q2k_q3k_gpu

📝 Commits (1)

7c4cb52 [feature] support q2_k & q3_k dequantize on gpu

📊 Changes

5 files changed (+160 additions, -11 deletions)

View changed files

📝 ktransformers/ktransformers_ext/cuda/binding.cpp (+5 -1)
📝 ktransformers/ktransformers_ext/cuda/custom_gguf/binding.cpp (+5 -0)
📝 ktransformers/ktransformers_ext/cuda/custom_gguf/dequant.cu (+129 -3)
📝 ktransformers/ktransformers_ext/cuda/custom_gguf/ops.h (+4 -2)
📝 ktransformers/util/custom_gguf.py (+17 -5)

📄 Description

Support dequantize q2_k & q3_k on gpu.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/kvcache-ai/ktransformers/pull/30 **Author:** [@BITcyman](https://github.com/BITcyman) **Created:** 8/12/2024 **Status:** ✅ Merged **Merged:** 8/12/2024 **Merged by:** [@UnicornChan](https://github.com/UnicornChan) **Base:** `develop-0.1.2` ← **Head:** `feature-lc-dequantize_q2k_q3k_gpu` --- ### 📝 Commits (1) - [`7c4cb52`](https://github.com/kvcache-ai/ktransformers/commit/7c4cb520bd935f944312604b4e6a79380da6a3ae) [feature] support q2_k & q3_k dequantize on gpu ### 📊 Changes **5 files changed** (+160 additions, -11 deletions) <details> <summary>View changed files</summary> 📝 `ktransformers/ktransformers_ext/cuda/binding.cpp` (+5 -1) 📝 `ktransformers/ktransformers_ext/cuda/custom_gguf/binding.cpp` (+5 -0) 📝 `ktransformers/ktransformers_ext/cuda/custom_gguf/dequant.cu` (+129 -3) 📝 `ktransformers/ktransformers_ext/cuda/custom_gguf/ops.h` (+4 -2) 📝 `ktransformers/util/custom_gguf.py` (+17 -5) </details> ### 📄 Description Support dequantize q2_k & q3_k on gpu. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-29 14:51:26 +00:00