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
Allowing users to use large `target_epsilon` for debugging and research
#604
jeandut
closed
7 months ago
8
feat: update lightning example to lightning 2.0
#603
lsc64
opened
10 months ago
7
The privacy engine make_private_with_epsilon function leads to memory error even for a very small network.
#602
telegraphroad
closed
7 months ago
7
PRV Accountant fails for specific input values for its args
#601
gauriprdhn
closed
7 months ago
2
WeightedRandomSampler is invalid
#600
LanpingTech
closed
11 months ago
1
Type check lint fixes
#599
karthikprasad
closed
11 months ago
2
Fixing bugs for DP MultiheadAttention
#598
HuanyuZhang
closed
7 months ago
9
ImportError: cannot import name 'DifferentiallyPrivateDistributedDataParallel' from 'opacus.layers'
#597
Ethan-Chen-plus
closed
7 months ago
6
DPMultiheadedAttention is not a full drop in replacement for nn.MultiheadAttention
#596
jfb54
closed
7 months ago
3
Backward pass with Opacus and ViT model. Batching rule not implemented for aten::_chunk_grad_outputs_efficient_attention.
#595
sebasrb09
closed
1 month ago
1
Real-time change of clipping thresholds
#594
Jenson66
closed
8 months ago
1
How to integrate Opacus with Nvidia NeMo
#593
yangjundy
closed
11 months ago
1
Got error when finetuning a gpt-like model with DPSGD
#592
xiehuanyi
closed
8 months ago
3
Cannot compute epsilon
#591
01manoj20
closed
7 months ago
2
AttributeError: The following layers do not have gradients
#590
Camelliavin
closed
8 months ago
4
Disable memory benchmarking
#589
ffuuugor
closed
11 months ago
10
Opacus with torch_geometric.nn and GCN's
#588
sagerkudrick
opened
1 year ago
5
More less the same value over different EPS values
#587
zedoul
opened
1 year ago
0
Opacus differential privacy support graph convolutional layers
#586
sachinkumarUI
opened
1 year ago
0
Negative Epsilons at high deltas
#585
tudorcebere
opened
1 year ago
2
RuntimeError: Expanded Weights encountered but cannot handle function mul
#584
heilrahc
closed
7 months ago
2
Different models and different image sizes can get the same epsilon
#583
oneheartforone
closed
1 year ago
1
How to fetch the original model again?
#582
TsingZ0
closed
1 year ago
1
Clean-up unnecessary warnings (including update to PyTorch 2.0)
#581
ffuuugor
closed
1 year ago
22
Pass `num_groups` arg to ModelValidator.fix
#580
ffuuugor
closed
1 year ago
8
Supporting the case torch sparse coo tensor as neural network input
#579
elichienxD
opened
1 year ago
3
per_sample gradient is None but grad is populated
#578
anirban-nath
opened
1 year ago
4
TypeError: 'type' object is not subscriptable
#577
GuYueJiahui
closed
1 year ago
2
DP with NAS and meet a ‘Model contains a trainable layer that Opacus doesn't currently support’
#576
Drxdx
opened
1 year ago
4
Support for torch compiled models
#575
tudorcebere
closed
8 months ago
2
Change z in PRV accountant to remove Userwarning for log(x<0)
#574
alexandresablayrolles
closed
1 year ago
4
Difference in Accuracy between setting epsilon to 10^6 and model wiithout opacus to train.
#573
Simha55
closed
2 months ago
1
added definitions for some required parameters
#572
sagerkudrick
closed
1 year ago
10
Privacy Leakage at low sample size
#571
tudorcebere
closed
2 months ago
6
Differential Privacy with Varying Epsilon Values in parameter: No Impact on Output
#570
Hafizamariaiqbal
opened
1 year ago
4
RuntimeError: The size of tensor a (32) must match the size of tensor b (21) at non-singleton dimension 0
#569
reanatom
closed
1 year ago
2
up Pytorch version for PyPI
#568
ffuuugor
closed
1 year ago
2
Same uncertainty value for every input
#567
zedoul
closed
1 year ago
3
Improve noise multiplier finding
#566
kiddyboots216
opened
1 year ago
6
Update compute_dp_sgd_privacy.py
#565
alexandresablayrolles
closed
1 year ago
7
Privacy computation script throws error due to lacking keyword arguments
#564
gkaissis
closed
1 year ago
1
release v1.4
#563
ffuuugor
closed
1 year ago
11
'PRVAccountant' object has no attribute 'get_privacy_spent' in Opacus code. Any solution please
#562
Hafizamariaiqbal
closed
1 year ago
3
ValueError: Per sample gradient is not initialized. Not updated in backward pass?Need solution
#561
Hafizamariaiqbal
closed
1 year ago
1
Per sample gradient is not initialized. Not updated in backward pass? Any Solution??Can anyone post a link to the Colab notebook that has the solution implemented?
#560
Hafizamariaiqbal
closed
1 year ago
5
make_private with clipping parameter
#559
lucacorbucci
closed
3 weeks ago
6
Install dependencies for website build
#558
ffuuugor
closed
1 year ago
11
No documentation at all
#557
zedoul
closed
1 year ago
3
Add Clipping schedulers
#556
Darktex
closed
1 year ago
8
Update PyTorch requirement to >=1.13
#555
ffuuugor
closed
1 year ago
31
Previous
Next