Can toch.cond be used recursively?
|
|
2
|
84
|
April 22, 2024
|
Error when backpropagating through a compiled PyTorch module multiple times
|
|
1
|
74
|
April 21, 2024
|
Is possible export the model on cpu and run it gpu?
|
|
4
|
111
|
April 17, 2024
|
Compile and deepcopy in the context of early stopping
|
|
0
|
86
|
April 16, 2024
|
Obscure error message when trying to export a compiled function
|
|
0
|
68
|
April 15, 2024
|
Torch.compile when /home is a read only filesystem
|
|
11
|
301
|
April 14, 2024
|
Torch.compile cache_size_limit best practice
|
|
1
|
108
|
April 12, 2024
|
Export and save problem(torch-tensorrt)
|
|
0
|
82
|
April 11, 2024
|
Compile function that uses constant tensor
|
|
1
|
112
|
April 10, 2024
|
How to achieve Torch-MLIR compatibility with Dynamo backend?
|
|
1
|
111
|
April 9, 2024
|
Is it possible to define a custom back-end wrapped with aot_autograd which does not use fake tensors?
|
|
0
|
64
|
April 9, 2024
|
The difference between `torch.ops.aten.addmm.default` and `torch.ops.aten.linear.default`
|
|
0
|
82
|
April 8, 2024
|
[dynamo] [onnx] Models exported with torch.onnx.dynamo_export show worse performance during inference
|
|
0
|
109
|
April 8, 2024
|
Return handle from compiled function from C++ to be used by backend
|
|
1
|
287
|
April 4, 2024
|
Unique node name across graph re-compilations and graph breaks
|
|
0
|
104
|
April 3, 2024
|
Why am I getting segfault when loading an AOT compiled model twice?
|
|
1
|
100
|
April 3, 2024
|
Speedup with max_autotune even though all the triton mm kernels are slower
|
|
0
|
91
|
April 2, 2024
|
Convolution speedup (slowdown) with torch.compile
|
|
4
|
102
|
April 2, 2024
|
Getting RuntimeError While Loading Model Weights
|
|
6
|
214
|
April 2, 2024
|
Memory Footprint of Eager Mode vs Torch Compiled (Default, with CUDA Graph, and CUDA Graph Dynamic)
|
|
2
|
217
|
April 2, 2024
|
Conda CMAKE CXX Compiler error while compiling Pytorch
|
|
0
|
138
|
March 31, 2024
|
Runtime error: Detected that your are using FX to symbolically trace a dynamo optimized function. This is not supported at the moment
|
|
1
|
187
|
March 26, 2024
|
Graph tracing false when using torch dynamo
|
|
0
|
111
|
March 25, 2024
|
Workarounds for eval_frame.py potentially leaking dram?
|
|
0
|
101
|
March 24, 2024
|
Exceptions in Torch Export
|
|
0
|
94
|
March 24, 2024
|
Torch.compile raises error when compiled function calls other functions
|
|
1
|
1371
|
March 22, 2024
|
A big picture to show the relationship among Compile, Dynamo, Inductor, AOTAutograd, AOTInductor, torch. export and related modules since PT2
|
|
0
|
106
|
March 21, 2024
|
How to find which ops are not supported by torch.compile
|
|
0
|
141
|
March 20, 2024
|
Skip incompatible - lcuda when searching for/usr/lib/i386 Linux gnu/libcuda. so
|
|
1
|
137
|
March 19, 2024
|
Cuda graphs kv_cache implementation slows inference significantly
|
|
4
|
279
|
March 18, 2024
|