Optimal Null-Space Rotations for Grouped Quantizers
Yanlong Chen PRO
ForeverBlue
AI & ML interests
None yet
Recent Activity
upvoted a collection 1 day ago
PrsimQuant commentedon a paper 2 days ago
PrismQuant: Optimal Null-Space Rotations for Grouped Quantizers commentedon a paper 2 days ago
PrismQuant: Optimal Null-Space Rotations for Grouped QuantizersOrganizations
None yet
ICML 2026 · Qwen3-VL-2B distilled from 8B. Start with the real AWQ-packed INT4 checkpoint, retaining 98% of GRACE BF16.
-
ForeverBlue/Qwen3-VL-2B-GRACE-W4G128-AWQ
Image-Text-to-Text • 0.9B • Updated • 23 • 1 -
GRACE-VLM
🦢1Try GRACE-VLM on an image, then deploy the INT4 build.
-
ForeverBlue/Qwen3-VL-2B-GRACE-BF16
Image-Text-to-Text • 2B • Updated • 42 • 1 -
ForeverBlue/Qwen3-VL-2B-GRACE-W4G128
Image-Text-to-Text • 2B • Updated • 28 • 1
Optimal Null-Space Rotations for Grouped Quantizers
ICML 2026 · Qwen3-VL-2B distilled from 8B. Start with the real AWQ-packed INT4 checkpoint, retaining 98% of GRACE BF16.
-
ForeverBlue/Qwen3-VL-2B-GRACE-W4G128-AWQ
Image-Text-to-Text • 0.9B • Updated • 23 • 1 -
GRACE-VLM
🦢1Try GRACE-VLM on an image, then deploy the INT4 build.
-
ForeverBlue/Qwen3-VL-2B-GRACE-BF16
Image-Text-to-Text • 2B • Updated • 42 • 1 -
ForeverBlue/Qwen3-VL-2B-GRACE-W4G128
Image-Text-to-Text • 2B • Updated • 28 • 1
models 11
ForeverBlue/PrismQuant-Llama-3.2-3B
Text Generation • Updated • 377 • 1
ForeverBlue/PrismQuant-Llama-3.1-8B
Text Generation • Updated • 372 • 1
ForeverBlue/PrismQuant-Qwen3-Base
Text Generation • Updated • 375 • 1
ForeverBlue/PrismQuant-Qwen3-30B-A3B-Base
Text Generation • Updated • 375 • 1
ForeverBlue/PrismQuant-Llama-3.1-70B
Text Generation • Updated • 372 • 1
ForeverBlue/Qwen3-VL-2B-GRACE-W4G128-AWQ
Image-Text-to-Text • 0.9B • Updated • 23 • 1
ForeverBlue/Qwen3-VL-2B-GRACE-W8G128
Image-Text-to-Text • 2B • Updated • 26 • 1
ForeverBlue/Qwen3-VL-2B-GRACE-W4G128
Image-Text-to-Text • 2B • Updated • 28 • 1
ForeverBlue/Qwen3-VL-2B-GRACE-BF16
Image-Text-to-Text • 2B • Updated • 42 • 1
ForeverBlue/LLaVA-1.5-7B-GRACE-W4G128-AWQ
Image-Text-to-Text • 1B • Updated • 15 • 1
datasets 0
None public yet