issues
search
sayakpaul
/
diffusers-torchao
End-to-end recipes for optimizing diffusion models with torchao and diffusers (inference and FP8 training).
Apache License 2.0
271
stars
8
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
'FakeTensor' object has no attribute '_quantized_linear_op'
#40
Revive624
closed
4 days ago
8
add a guide on ahead-of-time compilation
#39
sayakpaul
closed
2 weeks ago
1
Inference is fast but saving is slow.
#38
Leommm-byte
closed
1 month ago
4
qkv_fuse_projections() fails with torchao quantized Flux2DTransformerModel
#37
ngaloppo
opened
1 month ago
5
pipe.enable_sequential_cpu_offload()
#36
al-swaiti
closed
1 month ago
4
higher batch size details.
#35
sayakpaul
closed
2 months ago
1
gradio error- for inference
#34
jadechoghari
closed
2 months ago
13
fpx_weight_only quantization + torch.compile on CogVideoX
#33
a-r-r-o-w
closed
2 months ago
3
How should a quantized + compiled model be serialized?
#32
sayakpaul
closed
1 month ago
29
cleanup utility
#31
sayakpaul
closed
2 months ago
2
Use with lora
#30
akedia
closed
1 month ago
2
Keep getting `failed to autoquant AQInt8DynamicallyQuantizedLinearWeight`
#29
kilimchoi
closed
2 months ago
3
run_inference function taking too long time
#28
vatsal2473
closed
2 months ago
1
Update benchmark_image.py
#27
bonlime
closed
2 months ago
0
ComfyUI implementation
#26
pandayummy
closed
2 months ago
1
Attribute "weight" already exists
#25
nachoal
closed
1 month ago
4
Share requirements.txt file for this repo
#24
Tushar-ml
closed
2 months ago
1
fpx_weight_only quantization
#23
a-r-r-o-w
closed
2 months ago
2
Fix CogVideoX speedup numbers
#22
a-r-r-o-w
closed
2 months ago
3
CogVideoX benchmarks with fp8dqrow
#21
a-r-r-o-w
closed
2 months ago
4
test fp8dqrow
#20
sayakpaul
closed
2 months ago
0
Issue with torchao version maybe?
#19
gradjitta
closed
2 months ago
10
Add a note on sequential cpu offload for Cog
#18
a-r-r-o-w
closed
2 months ago
0
Smol analysis
#17
sayakpaul
closed
2 months ago
0
add flux table and smol analysis
#16
sayakpaul
closed
2 months ago
0
refactor readme iii to include flux.1 dev results
#15
sayakpaul
closed
2 months ago
0
simplify serialization and loading.
#14
sayakpaul
closed
2 months ago
3
README refactor II
#13
sayakpaul
closed
2 months ago
0
[FP8] Add fp8 configs
#12
drisspg
closed
2 months ago
2
README part I
#11
sayakpaul
closed
2 months ago
0
torchao-diffusers image & video model benchmarks
#10
a-r-r-o-w
closed
2 months ago
0
Using quantize_to_float8 results in "Only support dtype kwarg for autocast"
#9
a-r-r-o-w
closed
3 months ago
3
`torchao.autoquant` does not work with CogVideox
#8
a-r-r-o-w
closed
3 months ago
3
Initial results with CogVideoX
#7
a-r-r-o-w
closed
2 months ago
3
Error with FP8 compilation
#6
sayakpaul
closed
3 months ago
6
Training example with FP8?
#5
sayakpaul
closed
3 months ago
14
Try out FP6
#4
sayakpaul
closed
3 months ago
6
End-to-end example on serializing and loading
#3
sayakpaul
closed
3 months ago
5
Initial results
#2
sayakpaul
closed
2 months ago
24
saprsification + dynamic int8 doesn't work
#1
sayakpaul
closed
2 months ago
14