issues
search
aredden
/
flux-fp8-api
Flux diffusion model implementation using quantized fp8 matmul & remaining layers use faster half precision accumulate, which is ~2x faster on consumer devices.
Apache License 2.0
209
stars
22
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Certain lora not applied correctly.
#36
fyepi
opened
1 week ago
1
Acceleration not as expected
#35
alecyan1993
opened
1 week ago
1
when load certain lora, AttributeError: 'Flux' object has no attribute 'diffusion_model' happened.
#34
fyepi
opened
1 week ago
1
The possibility of supporting GPUs with other architectures
#33
ziyaxuanyi
opened
1 week ago
1
Compatibility Inquiry: Using flux-fp8 with OpenFLUX.1
#32
veyorokon
opened
2 weeks ago
1
After changing lora many times, the pictures are getting weirder and weirder
#31
81549361
opened
3 weeks ago
2
Issue: torch._scaled_mm RuntimeError on RTX 6000 (with runpod/pytorch:2.4.0-py3.11-cuda12.4.1-devel-ubuntu22.04)
#30
veyorokon
closed
3 weeks ago
2
A question regarding whether the LoRA has been successfully applied to the inference process？
#29
zhangqi420
closed
1 month ago
0
LoRA loaded successfully but the effect wasn't applied
#28
EntroSanity
opened
1 month ago
2
Why is vae decoder so slow? Can you help me?
#27
radish0926
opened
1 month ago
4
The speed of drawing is not satisfactory
#26
lvjin521
opened
1 month ago
4
TypeError: NoneType takes no arguments
#25
lvjin521
closed
1 month ago
4
Initial Delay in Image Generation with Flux Schnell on H100
#24
uayodev
opened
1 month ago
4
[bug]UnboundLocalError: cannot access local variable 'temp_77_token_ids' where it is not associated with a value
#23
81549361
closed
1 month ago
2
fix unloading bug
#22
aredden
closed
1 month ago
0
Removable lora
#21
aredden
closed
1 month ago
0
Load a LORA using the API
#20
acaladolopes
closed
1 month ago
3
PuLID support
#19
81549361
opened
2 months ago
0
Hot Lora Replacement
#18
Lantianyou
closed
1 month ago
17
Docker image support.
#17
ShivamB25
opened
2 months ago
5
How to save a "prequantized_flow" safetensor?
#16
smuelpeng
opened
2 months ago
5
Any plans for controlnet + inpainting support?
#15
0xtempest
opened
2 months ago
2
add benchmarks numbers for rtx4000ada (non-sff)
#14
flowpoint
closed
2 months ago
1
LoRA loading fails if only trained on specific blocks
#13
fblissjr
opened
2 months ago
21
Consider adding a license to the code
#12
flowpoint
closed
2 months ago
4
add h100
#11
ClashLuke
closed
2 months ago
1
Where is the code about "remaining layers use faster half precision accumulate"?
#10
goldhuang
opened
2 months ago
5
Potential LoRA performance issue
#9
ashakoen
closed
1 month ago
7
`NotImplementedError: Cannot copy out of meta tensor; no data!`
#8
montyanderson
opened
2 months ago
3
WHL files for torch-cublas-hgemm.git and ao?
#7
SoftologyPro
closed
2 months ago
2
[feature] Support for ControlNet from x-lab
#6
George0726
closed
2 months ago
3
Error No module named 'cublas_ops'
#5
ankitsiliconithub
closed
2 months ago
2
No issue - just a thank you!
#4
ashakoen
closed
2 months ago
2
Improved but configurable precision
#3
aredden
closed
3 months ago
0
python 3.10 compatibility
#2
XmYx
closed
3 months ago
1
update README
#1
dsingal0
closed
3 months ago
1