Issues

• Support of AdamW

  facebookresearch/higher#140 · Filter by author ridiculouz opened on May 6, 2023

• Does higher work with huggingface (hugging face, HF) models? e.g. ViT?

  facebookresearch/higher#139 · Filter by author brando90 opened on Mar 17, 2023
  2

• CUDA out of memory

  facebookresearch/higher#138 · Filter by author aooating opened on Mar 1, 2023

• once_differentiable

  facebookresearch/higher#137 · Filter by author aooating opened on Mar 1, 2023
  1

• grad clip correctness

  facebookresearch/higher#136 · Filter by author Whalefishin opened on Jan 6, 2023

• Using higher for hyperparameter optimization

  facebookresearch/higher#135 · Filter by author aruniyer opened on Dec 21, 2022
  1

• Non scalar loss

  facebookresearch/higher#134 · Filter by author janglinko-dac opened on Dec 14, 2022
  1

• Intialize Differentiable Optimizer with non-leaf tensort

  facebookresearch/higher#133 · Filter by author andrearosasco opened on Aug 8, 2022
  1

• higher for dpt architectures?

  facebookresearch/higher#132 · Filter by author Ainaz99 opened on Jul 22, 2022
  2

• How to Update/Optimize subset of params

  facebookresearch/higher#130 · Filter by author blake-camp opened on Mar 11, 2022
  2

• Potential bug with first order MAML using only higher, setting track_higher_grads = False leads to .grad field to not be populated and be None, is that a bug?

  facebookresearch/higher#129 · Filter by author brando90 opened on Feb 2, 2022
  4

• Would making the gradient "data" by detaching them implement first order MAML?

  facebookresearch/higher#128 · Filter by author brando90 opened on Feb 1, 2022

• Mixed precision training

  facebookresearch/higher#127 · Filter by author zhiqihuang opened on Jan 31, 2022

• In DifferentiableAdam, sqrt() is non-differentiable at zero

  facebookresearch/higher#125 · Filter by author rickyloynd-microsoft opened on Dec 10, 2021
  6

• Does higher work with hugging face Adafactor?

  facebookresearch/higher#124 · Filter by author brando90 opened on Nov 30, 2021
  2

• second-order deravative

  facebookresearch/higher#123 · Filter by author xugy16 opened on Nov 14, 2021

• Setting up a templates got gitissues that invite users to post to pytorch forum

  facebookresearch/higher#121 · Filter by author brando90 opened on Nov 4, 2021

• What does the warning flag in testing for maml mean? what does it mean to "fine tune for testing"?

  facebookresearch/higher#120 · Filter by author brando90 opened on Nov 4, 2021

• How does one return an adapted model without using the context manager?

  facebookresearch/higher#119 · Filter by author brando90 opened on Nov 2, 2021
  8

• Does higher work with DDP only test time evaluation (no training just testing)?

  facebookresearch/higher#118 · Filter by author brando90 opened on Oct 26, 2021

• When will higher allow the use of DDP (distributed data parallel)?

  facebookresearch/higher#116 · Filter by author brando90 opened on Sep 28, 2021
  4

• implementing MAML with MiniImageNet

  facebookresearch/higher#115 · Filter by author ligeng0197 opened on Sep 23, 2021

• metaclass issue with fastai

  facebookresearch/higher#113 · Filter by author shayanfazeli opened on Aug 3, 2021
  4

• maml omniglot - finetuning / test

  facebookresearch/higher#111 · Filter by author shayanfazeli opened on Jul 29, 2021
  5

• variables work outside of context manager scope

  facebookresearch/higher#110 · Filter by author hughperkins opened on Jun 22, 2021
  2