issues
search
pytorch
/
functorch
functorch is JAX-like composable function transforms for PyTorch.
https://pytorch.org/functorch/
BSD 3-Clause "New" or "Revised" License
1.38k
stars
102
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
installation issue
#1062
Tsingularity
opened
1 year ago
4
runtime performance gain on model ensembling
#1061
xuyxu
closed
1 year ago
2
aten::all not implemented
#1060
iliTheFallen
closed
1 year ago
7
Cuda Memory Overflow in Jacobian Computation
#1058
kayween
opened
1 year ago
6
Installing functorch breaks torchaudio
#1057
dellis23
closed
1 year ago
11
function(Jacobian)-dot-vector and vector-Jacobian-vector function
#1056
veya2ztn
opened
1 year ago
0
[1.13] Tighten packaging versions
#1055
zou3519
closed
1 year ago
0
Better "oops your tensor escaped from vmap" error message
#1054
zou3519
closed
1 year ago
2
_get_nested_attr returning None due to missing a return in the general case
#1053
andreykramer
closed
1 year ago
1
Semantic discrepancy on requires_grad after compiling Tensor.detach
#1052
sangongs
closed
1 year ago
14
[release] Build functorch dummy binaries
#1051
zou3519
closed
1 year ago
0
vmap over conv_transpose on CUDA appears incorrect on A100 GPUs
#1050
zou3519
closed
1 year ago
1
.item() error when computing Jacobian with vmap and `torch.autograd.set_detect_anomaly(True)`
#1049
jotix16
closed
1 year ago
7
don't install torchvision, it's not ready yet
#1048
zou3519
closed
1 year ago
0
[release] Set up version; set up testing
#1047
zou3519
closed
1 year ago
0
Training with functorch causes constant to be embedded in graph and returned
#1046
dellis23
closed
1 year ago
1
[release] Set up version; set up testing
#1045
zou3519
closed
1 year ago
1
Bump functorch version
#1044
zou3519
closed
1 year ago
0
Is there a way to parallelize or accelerate a loop of column-by-column jvp?
#1043
kwmaeng91
opened
1 year ago
3
Re-add some notebooks back for BC
#1041
zou3519
closed
1 year ago
0
Small (<5%) regression in functorch_maml_omniglot_cuda model from 0.2.1 to latest
#1040
zou3519
opened
1 year ago
3
Possible (-2 to 4%) regression in functorch_dp_cifar10_cuda model from 0.1.1 to latest
#1039
zou3519
opened
1 year ago
3
functorch 1.13 release tracker
#1038
zou3519
opened
1 year ago
0
Get intermediate derivatives with nested jacobian and has_aux
#1037
i-a-morozov
closed
1 year ago
2
support scan
#1036
GallagherCommaJack
opened
1 year ago
3
Preparations for the upcoming release
#1035
zou3519
closed
1 year ago
0
Update README
#1034
zou3519
closed
1 year ago
0
fix logsigmoid_backward batching rule
#1033
zou3519
closed
1 year ago
2
Module not found after installation of nightly build
#1032
zonexo
opened
1 year ago
2
Support for custom gradient functions
#1031
esteveste
closed
1 year ago
10
Add support for `tree_map` or document recommended alternative
#1030
dellis23
opened
1 year ago
9
Better contiguity testing in functorch opinfo testing
#1029
zou3519
opened
1 year ago
2
Speed up functorch tests in CI
#1028
zou3519
closed
1 year ago
0
functorch doesn't work with saved variable hooks
#1027
zou3519
opened
1 year ago
1
Printing unwrapped tensors doesn't work under grad-type transform
#1026
samdow
closed
1 year ago
2
Change vmap to work around "vmap-incompatible in-place errors"
#1025
zou3519
opened
1 year ago
0
Get .item() error without calling .item()
#1024
LiXinrong1012
opened
1 year ago
1
Silent incorrectness for vmap x conv on M1 Macs
#1023
zou3519
closed
1 year ago
9
Regression in vmap x jvp support for torch.linalg.det
#1022
zou3519
closed
1 year ago
1
finish setting up regression benchmarks in torchbench
#1021
zou3519
closed
1 year ago
1
Investigate per-sample-grad performance relative to other mechanism
#1020
zou3519
opened
1 year ago
1
Silently incorrect behavior caught by test_vmapjvpvjp
#1019
zou3519
closed
1 year ago
3
Hi-pri failures for vmapvjpvjp test
#1018
zou3519
closed
1 year ago
1
Create dummy package for functorch
#1017
zou3519
closed
1 year ago
0
[hotfix] fix build error from FuncTorch dispatch key
#1016
winnie1994
opened
1 year ago
2
Distribute functorch for PyTorch 1.13
#1015
zou3519
opened
1 year ago
0
Default Partition: Backward module does not contain computation of the parameters' gradients.
#1014
ConnollyLeon
opened
1 year ago
3
Simultaneous computation of per-sample gradient and per-batch gradient
#1012
SJShin-AI
opened
1 year ago
1
memory_efficient_fusion leads to RuntimeError for higher-order gradients calculation. RuntimeError: You are attempting to call Tensor.requires_grad_()
#1011
AlphaBetaGamma96
opened
1 year ago
3
Multiple gradient calculation for single sample
#1010
JoaoLages
closed
1 year ago
23
Previous
Next