Issues

• _select_initial_step crashes on complex y0 with RuntimeError: maximum not implemented for complex tensors

  rtqichen/torchdiffeq#294 · Filter by author mitchdz opened on Oct 8, 2026

• gradient computed through _select_initial_step

  rtqichen/torchdiffeq#284 · Filter by author LeonardoPascotto opened on Sep 13, 2026

• torch.vmap (or other form of parallelism) Support for odeint_event

  rtqichen/torchdiffeq#282 · Filter by author LucasGrasso opened on Jun 16, 2026
  9

• solver unwraps batch and runs a python loop

  rtqichen/torchdiffeq#278 · Filter by author TobiasPfeifer opened on Feb 1, 2026
  1

• Why is the number of timesteps smaller than the number of forward calls?

  rtqichen/torchdiffeq#276 · Filter by author sambaPython24 opened on Oct 19, 2025
  1

• Can Torchdiffeq add support for sparse tensors for y0

  rtqichen/torchdiffeq#275 · Filter by author Zhiyong-Xu2 opened on Jul 14, 2025
  1

• TypeError: 'file' is an invalid keyword argument for warn()

  rtqichen/torchdiffeq#274 · Filter by author pgzmnk opened on Jul 11, 2025

• New release?

  rtqichen/torchdiffeq#272 · Filter by author hchau630 opened on Jun 30, 2025
  2

• About the nonlinear function to fit

  rtqichen/torchdiffeq#270 · Filter by author MADONOKOUKI opened on Jun 11, 2025

• 0-dim t issue on multi gpus

  rtqichen/torchdiffeq#268 · Filter by author cb6e310 opened on Apr 16, 2025

• Use Sherman-Morrison formula with implicit solvers

  rtqichen/torchdiffeq#267 · Filter by author psv4 opened on Apr 4, 2025

• Request Example for odeint_dense with Interpolation Support

  rtqichen/torchdiffeq#266 · Filter by author lixfrank opened on Mar 31, 2025

• Adjoint support for complex dtypes.

  rtqichen/torchdiffeq#264 · Filter by author gitvicky opened on Feb 25, 2025

• Help needed. Is there a way to call the function inside a numba @cuda.jit kernel function? if not how should i use the package on a GPU?

  rtqichen/torchdiffeq#262 · Filter by author anbonimus opened on Feb 11, 2025

• Why is the integration upper bound t=1 in the MNIST example?

  rtqichen/torchdiffeq#260 · Filter by author xliu99 opened on Jan 7, 2025
  1

• Can't replicate O(1) memory with Adjoint method

  rtqichen/torchdiffeq#259 · Filter by author petercmh01 opened on Dec 23, 2024

• MPS does not support float64

  rtqichen/torchdiffeq#257 · Filter by author bkrenusz opened on Dec 4, 2024
  1

• "nextafter_cuda" not implemented for 'Half'

  rtqichen/torchdiffeq#255 · Filter by author maxin-cn opened on Oct 1, 2024

• Enabling Mixed Precision Training for Your Model

  rtqichen/torchdiffeq#252 · Filter by author xwqtju opened on Jun 26, 2024

• A warning message when there is only 1 evaluation point provided would be helpful

  rtqichen/torchdiffeq#251 · Filter by author lcdavis13 opened on Jun 25, 2024

• Question about the gradient of odeint

  rtqichen/torchdiffeq#250 · Filter by author Hank0626 opened on May 31, 2024

• Scipy LSODA for stiff ODE

  rtqichen/torchdiffeq#249 · Filter by author janrohleff opened on Apr 3, 2024
  2

• Perform one integration step

  rtqichen/torchdiffeq#248 · Filter by author JackCaster opened on Mar 10, 2024
  2

• Bug: Memory Leaky with from torchdiffeq import odeint

  rtqichen/torchdiffeq#247 · Filter by author wangmiaowei opened on Feb 26, 2024
  3

• How to pass extra paramaters of func to odeint？

  rtqichen/torchdiffeq#246 · Filter by author shifttttttt opened on Jan 23, 2024
  3