issues
search
pytorch
/
opacus
Training PyTorch models with differential privacy
https://opacus.ai
Apache License 2.0
1.65k
stars
328
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Add Clipping schedulers
#556
Darktex
closed
1 year ago
8
Update PyTorch requirement to >=1.13
#555
ffuuugor
closed
1 year ago
31
Fix website deployment via github actions
#554
ffuuugor
closed
1 year ago
11
Using 'from opacus import PrivacyEngine' and 'import logging', the formatter of print Information was set automatically.
#553
JiafengtTang
closed
1 year ago
4
Github actions for docusaurus deployment
#552
karthikprasad
closed
1 year ago
12
cannot install opacus with torch(+gpu)
#551
giladcohen
closed
1 year ago
5
Add option to specify number of steps in UniformSampler
#550
alexandresablayrolles
closed
1 year ago
3
Update FAQ to meet current api
#549
MarcinMisiurewicz
closed
1 year ago
5
AttributeError: 'Parameter' object has no attribute 'grad_sample' in Projected GAN
#548
sword-king1
opened
1 year ago
1
Test PR 2
#547
ashkan-software
closed
1 year ago
2
Test PR 3
#546
ashkan-software
closed
1 year ago
2
test PR 2
#545
ashkan-software
closed
1 year ago
1
test PR
#544
ashkan-software
closed
1 year ago
2
default prefetch factor is None
#543
ffuuugor
closed
1 year ago
4
Opacus report an error when freezing weights but preserving bias gradients. It is recommended to add the weight freezing function.
#542
xtyXpastor
closed
1 year ago
3
change how we access accountant history in GDP
#541
ashkan-software
closed
1 year ago
1
release v1.3
#540
pierrestock
closed
1 year ago
2
When using adaclip , how to calculate eps or delta
#539
Adsensess
closed
1 year ago
0
Floating-point Vulnerability Protection
#538
gonzalo-munillag
closed
1 year ago
1
Add third blog series to the website
#537
ashkan-software
closed
1 year ago
3
test pr
#536
ashkan-software
closed
1 year ago
1
GDP Accountant get_epsilon method should not pop from history
#535
kiddyboots216
closed
1 year ago
1
Support empty batches for arbitrary dataset structures
#534
ffuuugor
opened
1 year ago
0
relax benchmarking thresholds for new PyTorch
#533
ffuuugor
closed
1 year ago
1
Per sample grad correctness util
#532
psolikov
closed
1 year ago
20
alternative functorch fix
#531
ffuuugor
closed
1 year ago
0
Support empty batches
#530
ffuuugor
closed
1 year ago
13
Make DPMultiheadAttention drop-in compatible with nn.MultiheadAttention
#529
Wei-1
closed
1 year ago
17
Make DPMultiheadAttention drop-in compatible with nn.MultiheadAttention
#528
Wei-1
closed
1 year ago
1
Conv3d test fix
#527
ffuuugor
closed
1 year ago
3
align docstring of make_private_with_epsilon with args
#526
Solosneros
closed
1 year ago
4
Docstring of make_private_with_epsilon is outdated
#525
Solosneros
closed
1 year ago
1
Opacus: I made a model-agnostic callback for PyTorch Lightning
#524
gianmarcoaversanoenx
opened
1 year ago
0
Opacus does not work on provided example for GAN and is broken on GANs in general
#523
gianmarcoaversanoenx
opened
1 year ago
1
The `BatchSplittingSampler` cannot handle empty batches
#522
xichens
opened
1 year ago
1
functorch mode computes the forward twice
#521
kshitij12345
opened
1 year ago
3
Fixed issue with missing argument in MNIST example
#520
gouthamuoc
closed
1 year ago
4
Support nn.EmbeddingBag
#519
alexandresablayrolles
closed
1 year ago
10
EmbeddingBag with offsets is not supported on Opacus
#518
karthikprasad
closed
1 year ago
1
Missing argument in MNIST example
#517
danperazzo
closed
1 year ago
3
Length of BatchSplittingSampler with Poisson sampling
#516
s-zanella
opened
1 year ago
6
coverage via coveralls in CircleCI
#515
arn7av
closed
1 year ago
4
Add an FAQ about variable batch size
#514
karthikprasad
opened
1 year ago
0
increase deadline for test
#513
karthikprasad
closed
1 year ago
2
AttributeError: 'DPMultiheadAttention' object has no attribute 'batch_first'
#512
Simha55
closed
7 months ago
4
functorch-based per sample gradients don't match with vanilla
#511
ffuuugor
opened
1 year ago
0
Functorch gradients: investigation and fix
#510
ffuuugor
closed
1 year ago
12
Align make_private_with_epsilon with make_private
#509
alexandresablayrolles
closed
1 year ago
2
restore benchmarking for dp layers and fix nightly tests
#508
ffuuugor
closed
1 year ago
6
Fix benchmarking CI
#507
ffuuugor
closed
1 year ago
1
Previous
Next