hatchetProject / QuEST

QuEST: Efficient Finetuning for Low-bit Diffusion Models

hatchetProject/QuEST Issues

Simulated quantization or actual quantization?
Closed a month ago3
About w4a4 calibration model
Closed a month ago21
Reason for eta=0.0 in the imagenet script?
Closed a month ago2
Discrepancy in the order of conditioning
Closed 2 months ago3
Code looks like QKMatmul and SMVMatmul calculates with full precision matrix multiplication.
Closed 2 months ago3
[BUG] The time-step seems not set in the quantization init.
Closed 2 months ago8
Some questions about the code
Closed 2 months ago1
Can you provide inference time data for Stable Diffusion (w4a4 vs full precision fp32) on GPU/CPU?
Updated 3 months ago
Resume calibrated models
Closed 3 months ago4
Strange Results Produced on W4A4 LSUN-Bedrooms Dataset Calibration
Closed 4 months ago5
Model Initialization Time Inquiry
Closed 4 months ago11
layer_reconstruction() got an unexpected keyword argument 'timesteps'
Closed 4 months ago3
How do you measure FID, sFID, IS and CLIP Score?
Closed 5 months ago1
Can you share your calibration script and evaluation script?
Closed 5 months ago10
KeyError in cross_attn_forward
Closed 5 months ago2