[GUIDE] Getting C++ custom ops to work with torch.compile
|
|
3
|
691
|
January 16, 2024
|
[RFC] LazyScheduler for operator reordering
|
|
0
|
159
|
January 13, 2024
|
Decomposition slows down the lazy tensor tracing
|
|
5
|
263
|
January 11, 2024
|
Question about the necessity of some guards in Dynamo
|
|
4
|
130
|
January 3, 2024
|
Problems with torch.compile generated code in tutorial
|
|
3
|
174
|
January 2, 2024
|
Inductor Triton Custom Op
|
|
4
|
411
|
January 2, 2024
|
TorchDynamo Update 10: Integrating with PyTorch/XLA for Inference and Training
|
|
9
|
4499
|
December 29, 2023
|
TorchDynamo Update 11: Making FSDP and Dynamo Work Together
|
|
5
|
3463
|
December 27, 2023
|
What is the relationship requirement among original bytecode, transformed bytecode, and bytecode returned by hooks in Dynamo?
|
|
6
|
288
|
December 16, 2023
|
How to get subgraph scheduled by inductor?
|
|
1
|
177
|
December 15, 2023
|
How to re-decompose operator in Inductor pass
|
|
0
|
130
|
December 6, 2023
|
How to trace torch.autograd.backward or torch.autograd.grad?
|
|
3
|
338
|
December 5, 2023
|
Arithmatic expression of prims operator
|
|
0
|
129
|
December 3, 2023
|
Inductor CUDA Backend
|
|
0
|
191
|
November 30, 2023
|
Call for backward compatability to enable users to understand and adapt to pytorch compiler
|
|
2
|
396
|
November 18, 2023
|
Torch.compile() tech talks at PTC'23
|
|
0
|
432
|
October 31, 2023
|
Torch.compile with AOT Autograd can be debugged now!
|
|
1
|
405
|
October 31, 2023
|
Torch.compile can be debugged now!
|
|
2
|
794
|
October 24, 2023
|
Interactively explore what torch.compile does to your code!
|
|
0
|
583
|
October 15, 2023
|
PyTorch/XLA 2.1 Release Dev Update
|
|
1
|
1005
|
October 13, 2023
|
Is Dynamo enabled per-thread or per-process?
|
|
2
|
458
|
October 7, 2023
|
A minimal working example of standalone usage for dynamo eval_frame
|
|
2
|
378
|
September 29, 2023
|
Defining the Core ATen Opset
|
|
11
|
2715
|
September 25, 2023
|
TorchInductor Update 6: CPU backend performance update and new features in PyTorch 2.1
|
|
0
|
1029
|
September 22, 2023
|
Why Dynamo fails to capture the computation graph in this function?
|
|
1
|
311
|
September 14, 2023
|
A proposal to make Dynamo more understandable to users
|
|
3
|
772
|
September 5, 2023
|
I build a decompiler to convert bytecode generated by dynamo into readable source code!
|
|
4
|
444
|
August 29, 2023
|
CUDAGraphs in Pytorch 2.0
|
|
3
|
1950
|
August 11, 2023
|
Supporting Dynamo in Python 3.11 - CPython Frame Evaluation
|
|
0
|
728
|
August 8, 2023
|
How are guards installed on frames that are transient objects?
|
|
9
|
407
|
August 6, 2023
|