Opacus

Topic	Replies	Views	Activity
Re-directing Questions to Github Issues Dear Opacus users, We kindly request that you redirect your questions to our Github issue page (Issues · pytorch/opacus · GitHub) if you would like attention from the Opacus team. Due to limited bandwidth and a desire t…	1	191	October 1, 2024
About the Opacus category	0	1275	December 9, 2020
Why are the scales of max_grad_norm different between loss_reduction='mean' and 'sum'	0	40	March 18, 2026
Opacus: training MONAI U-Net on a MSD dataset	0	31	December 10, 2025
Attempted to use an uninitialized parameter in <built-in method empty_like of type object at 0x00007FFF5B826810>. This error happens when you are using a `LazyModule` or explicitly manipulating `torch.nn.parameter.UninitializedParameter` objects	0	41	October 13, 2025
OPACUS unexpected behaviour	0	33	September 10, 2025
Training Loss and Accuracy both increasing while training DP-SGD	0	71	April 25, 2025
BatchMemoryManager with Opacus in Lightning	4	938	April 4, 2025
Loss calculation error	0	112	January 29, 2025
IndexError: pop from empty list2	0	132	January 8, 2025
LoRA + DP-SGD tutorial	0	523	December 18, 2024
How to handle .detach() in opacus enabled model	0	95	November 29, 2024
Dependencies for `csprng` becoming untenable?	1	572	November 28, 2024
Using Opacus in federated learning	5	2397	November 24, 2024
IterableDataset support in Opacus	1	159	October 1, 2024
How to calculate epsilon for DP-SGD while applying using Func torch	0	201	September 27, 2024
Problem when finetuning LLM using opacus	0	143	September 20, 2024
Privacy accounting based on DPDataLoader	1	89	September 12, 2024
Query on ExpandedWeights	1	176	September 12, 2024
Opacus for 3D Segmentation	0	159	September 9, 2024
Source code for gradient clipping and noise addition in Opacus	4	216	September 3, 2024
Profiling time spent during DP training	2	207	September 2, 2024
IndexError: pop from empty list	0	172	August 14, 2024
What is the best practise to use FSDP with opacus?	4	1245	July 3, 2024
Opacus ValueError: Per sample gradient is not initialized. Not updated in backward pass?	0	169	June 27, 2024
Error: No activations detected in Lightning module with Opacus	0	239	June 2, 2024
How to use make_private() with a model that registers a buffer?	1	254	May 15, 2024
Influence of `max_grad_norm` on DP parameter `epsilon`	2	550	May 15, 2024
BatchMemoryManager with gans	0	220	April 26, 2024
Relation between Batch_size and Gradients	0	273	April 22, 2024

Re-directing Questions to Github Issues

Dear Opacus users, We kindly request that you redirect your questions to our Github issue page (Issues · pytorch/opacus · GitHub) if you would like attention from the Opacus team. Due to limited bandwidth and a desire t…