dotnet/TorchSharp

Issues

• Optim.Rprop issue?

  #1417 opened a month ago by gavolio
  2

• Memory Allocation Performance

  #1416 opened a month ago by mdcarr941
  4

• tensor.print() - Missing default "newLine" Parameter

  #1414 opened 23 days ago by ozanMSFT
  1

• New Error with devices

  #1418 opened 23 days ago by Sprinzl
  2

• draw_bounding_boxes

  #1406 opened 2 months ago by BehalDavid
  1

• Is there a plan to package pytorch3d into torchsharp?

  #1412 opened a month ago by williamlzw
  0

• `torch.nn.functional.l1_loss` computes a criterion with the MSE, not the MAE

  #1359 opened 5 months ago by RensOliemans
  6

• Look into whether parameter creation from a tensor leads to incorrect dispose scope statistics.

  #1397 opened a month ago by NiklasGustafsson
  14

• Add support for `torch.nn.functional.normalize`

  #1382 opened 2 months ago by MiWeiss
  0

• slice bug

  #1399 opened 2 months ago by whuanle
  3

• diagonal() has incorrect default

  #1402 opened 2 months ago by LeeReid1
  1

• There may be an error in torchvision.transforms.GaussianBlur

  #1400 opened 2 months ago by williamlzw
  6

• No way to copy a tensor from gpu to cpu to pre allocated array.

  #1388 opened 2 months ago by LukePoga
  6

• Is TorchText implemented?

  #1381 opened 2 months ago by przemyslawbak
  3

• New contributor didn't link tests in WithCudaBinaries project

  #1394 opened 2 months ago by mvphelps
  0

• ToTensor extension method creates an orphaned tensor (Memory Leak)

  #1392 opened 2 months ago by mvphelps
  0

• Add accumulate to index_put_

  #1374 opened 4 months ago by BalashovK
  1

• Provide better metrics to enable memory leak troubleshooting

  #1390 opened 2 months ago by mvphelps
  0

• Attaching unowned tensor to a DisposeScope makes Statistics.DetachedFromScopeCount go negative.

  #1387 opened 2 months ago by mvphelps
  0

• PackedSequences not aware of DisposeScope system

  #1385 opened 2 months ago by mvphelps
  0

• Support use_deterministic_algorithms()

  #1384 opened 2 months ago by medovina
  0

• Torch was not compiled with flash attention warning

  #1375 opened 3 months ago by lostmsu
  1

• torchaudio support

  #1377 opened 3 months ago by williamlzw
  7

• Using TorchSharp In Unity

  #1372 opened 4 months ago by StefanWRL
  2

• How to use it in maui

  #1370 opened 4 months ago by Barshan-Mandal
  7

• multi-gpu training

  #1369 opened 4 months ago by lindadamama
  2

• Why this dotnet pack error?

  #1366 opened 4 months ago by GeorgeS2019
  0

• GELU does not appear to support approximate tanh

  #1368 opened 5 months ago by travisjj
  4

• Move to libtorch 2.4.0

  #1330 opened 5 months ago by NiklasGustafsson
  1

• MissingMethodException: Method not found: '!!0 TorchSharp.ModuleExtensionMethods.cuda(!!0, Int32)'

  #1362 opened 5 months ago by dansimon363
  8

• Suppress the API breakchange introduced by PR 1354 and bump up minor version

  #1356 opened 5 months ago by LittleLittleCloud
  3

• Build failing after merge of PR 1354

  #1355 opened 5 months ago by NiklasGustafsson
  0

• IterableDataset

  #1353 opened 5 months ago by yueyinqiu
  0

• Arrays of Arrays

  #1349 opened 6 months ago by travisjj
  4

• Jagged Array to Tensor

  #1347 opened 6 months ago by gktval
  5

• Modernize torchsharp infrastructure

  #1346 opened 6 months ago by ericstj
  0

• Builds fail after package validation PR merged

  #1342 opened 6 months ago by NiklasGustafsson
  2

• expose lr scheduler fields

  #1320 opened 7 months ago by yueyinqiu
  4

• MultivariateNormal.log_prob() exception in TorchSharp but works in pytorch

  #1334 opened 6 months ago by Kermalis
  16

• Unexpected parameters

  #1332 opened 6 months ago by lintao185
  11

• How do I turn off these warnings?

  #1322 opened 6 months ago by lintao185
  17

• `Module.to(Device)` will cause warnings because non-leaf tensors `.grad` is accessed (Also for optimizers)

  #1323 opened 6 months ago by yueyinqiu
  6

• Request tensor factory methods accepting Span<T>

  #1326 opened 6 months ago by dje-dev
  9

• Add register_pre_hook and register_hook for generic module class?

  #1321 opened 6 months ago by LittleLittleCloud
  3

• The function torch.cuda.empty_cache() is missing.

  #1327 opened 7 months ago by lintao185
  6

• Linear module creates parameter on cpu even after I set default device to meta

  #1324 opened 7 months ago by LittleLittleCloud
  1

• How do I get the value of such an attribute, just for display purposes.

  #1314 opened 7 months ago by lintao185
  4

• How to set default device type in torchsharp.

  #1317 opened 7 months ago by LittleLittleCloud
  0

• In PyTorch, the `non_blocking` parameter of the `torch.to` method is not present in TorchSharp.

  #1313 opened 7 months ago by lintao185
  0

• Calls to NativeLibrary should not assume explicit paths

  #1311 opened 7 months ago by glopesdev
  0