Article
Harim Choi PRO
HarimxChoi
AI & ML interests
None yet
Recent Activity
published an article about 22 hours ago
WarpQuant: Dual-Domain LLM Quantization via Hadamard Rotation and Output-Fisher Sensitivity updated a collection about 22 hours ago
WarpQuant: Dual-Domain INT3 LLM Quantization updated a collection about 22 hours ago
WarpQuant: Dual-Domain INT3 LLM QuantizationOrganizations
None yet
Hadamard-rotated 3-bit PTQ with Output-Fisher weak-column recovery for Qwen3.8-27B, Qwen3.5-4B, and Llama 3 8B.
-
HarimxChoi/WarpQuant-Qwen3.8-27B-R16E4H4-Text
Text Generation • 27B • Updated • 209 -
HarimxChoi/WarpQuant-Qwen3.8-27B-R16E4H4
Image-Text-to-Text • 28B • Updated • 12 -
HarimxChoi/WarpQuant-Qwen3.5-4B-R16E4-Text
Text Generation • 4B • Updated • 125 -
HarimxChoi/WarpQuant-Qwen3.5-4B-R16E4V4
Image-Text-to-Text • 5B • Updated • 11
Hadamard-rotated 3-bit PTQ with Output-Fisher weak-column recovery for Qwen3.8-27B, Qwen3.5-4B, and Llama 3 8B.
-
HarimxChoi/WarpQuant-Qwen3.8-27B-R16E4H4-Text
Text Generation • 27B • Updated • 209 -
HarimxChoi/WarpQuant-Qwen3.8-27B-R16E4H4
Image-Text-to-Text • 28B • Updated • 12 -
HarimxChoi/WarpQuant-Qwen3.5-4B-R16E4-Text
Text Generation • 4B • Updated • 125 -
HarimxChoi/WarpQuant-Qwen3.5-4B-R16E4V4
Image-Text-to-Text • 5B • Updated • 11
models 5
HarimxChoi/WarpQuant-Llama-3-8B-R16E4H4
Text Generation • 8B • Updated • 100
HarimxChoi/WarpQuant-Qwen3.5-4B-R16E4V4
Image-Text-to-Text • 5B • Updated • 11
HarimxChoi/WarpQuant-Qwen3.5-4B-R16E4-Text
Text Generation • 4B • Updated • 125
HarimxChoi/WarpQuant-Qwen3.8-27B-R16E4H4
Image-Text-to-Text • 28B • Updated • 12
HarimxChoi/WarpQuant-Qwen3.8-27B-R16E4H4-Text
Text Generation • 27B • Updated • 209
datasets 0
None public yet