|
TorchDynamo Update 5: Improved Capture & Bigger Graphs
|
|
4
|
4008
|
April 14, 2022
|
|
TorchDynamo Update 6: Training support with AOTAutograd
|
|
0
|
5655
|
March 29, 2022
|
|
Universal binaries for libtorch Mac?
|
|
10
|
3787
|
March 14, 2022
|
|
Prim::autocast_promote operation?
|
|
4
|
1373
|
December 31, 2021
|
|
TorchDynamo Update: 1.48x geomean speedup on TorchBench CPU Inference
|
|
0
|
5679
|
November 12, 2021
|
|
Next Steps for PyTorch Compilers
|
|
9
|
10534
|
October 21, 2021
|
|
Loop_tool's lazy frontend - Experimenting with symbolic laziness
|
|
0
|
934
|
October 11, 2021
|
|
torch-MLIR presentation from Google / nod.ai
|
|
1
|
1153
|
October 8, 2021
|
|
Depthwise conv2d: An NNC Case Study
|
|
0
|
1465
|
April 7, 2021
|
|
Symbolic Shape Inference
|
|
1
|
1574
|
March 31, 2021
|
|
Summary of Issues Found in Adapting Models to use TorchScript
|
|
1
|
959
|
March 16, 2021
|
|
ExecutionPlan caching in ProfilingGraphExecutorImpl?
|
|
0
|
732
|
March 3, 2021
|
|
Adapting Models to use TorchScript and Getting them to Produce Fusions
|
|
15
|
6355
|
February 26, 2021
|
|
Single-op fusion benchmarking
|
|
0
|
837
|
February 4, 2021
|
|
NNC Per-Operator Benchmarks (on CPU)
|
|
5
|
1055
|
January 27, 2021
|
|
TorchScript usability
|
|
2
|
1916
|
January 26, 2021
|
|
Eliminating Framework Overhead by Compiling a Model Directly
|
|
3
|
1285
|
January 25, 2021
|
|
Understanding TorchScript Type System
|
|
2
|
1835
|
January 25, 2021
|
|
Perf counters for fun and profit
|
|
0
|
1008
|
January 23, 2021
|
|
Quick Intro to E-Graphs
|
|
0
|
969
|
January 22, 2021
|