compile

torch._inductor The default compiler in torch.compile, builds loops schedules and codegens triton or C++ kernels. torch._dynamo PyTorch’s compiler frontend which can output subgraphs when programs aren’t traceable. FX (Functional Transformations) FX is a toolkit for developers to use to transform nn.Module instances. FX consists of three main components: a symbolic tracer, an intermediate representation , and Python code generation .

Topic	Replies	Views	Activity
About the compile category compile	0	112	January 9, 2023
What can cause "[WARNING] DeviceCopy in input program"? compile	8	32	March 23, 2023
Model compilation error - BrokenProcessPool compile	1	24	March 23, 2023
How to serialize models with torch.compile properly compile	2	40	March 23, 2023
Using TorchTRT-FX backend on C++ FX (Functional Transformations)	5	112	March 23, 2023
How can compiled model generates a torch.fx.GraphModule compile	4	67	March 23, 2023
ValueError: --optim adamw_torch_fused with --fp16 requires PyTorch>2.0 torch._dynamo	12	67	March 23, 2023
"PTX .version 7.4 does not support .target sm_89" Again with latest Pytorch2.0, 4090 with CUDA12 and Ubuntu Linux 22.04 torch._dynamo	9	145	March 22, 2023
`torch.compile` + `torch.no_grad` not working for Mask R-CNN torch._dynamo	2	49	March 22, 2023
[TorchDynamo] Too many function args in a fx graph torch._dynamo	2	87	March 22, 2023
Pytorch 2.0 compile warning meaning torch._dynamo	1	115	March 22, 2023
How to get whole graph when subgraph_compile triggle torch._dynamo	1	58	March 22, 2023
UserDefinedObjectVariable should not be break renson torch._dynamo	1	37	March 22, 2023
Triton Error [CUDA]: an illegal memory access was encountered torch._dynamo	9	155	March 22, 2023
Trying to use triton on torch inductor torch._inductor	1	89	March 21, 2023
How can I tell whether the output of one subgraph is the input to the next? compile	0	21	March 20, 2023
Why pytorch 2.0 makes code slower on 2080ti? compile	4	85	March 17, 2023
PyTorch 2.0 torch.compile works on CPU but not GPU torch._dynamo	2	87	March 17, 2023
Question about graph break related to RETURN_VALUE and INLINE function call torch._dynamo	1	48	March 17, 2023
Is PyTorch 2.0 GA or Beta? compile	7	89	March 17, 2023
How to replace modules like TorchFX for compiled model compile	1	40	March 16, 2023
Understanding decompositions done in primtorch and autograd compile	1	42	March 16, 2023
Aot_eager mode segfault during bert training torch._dynamo	0	31	March 14, 2023
Any wheel for CUDA 12 coming? compile	2	101	March 14, 2023
TorchDynamo - how to fallback to the fine grained subgraph? compile	1	48	March 14, 2023
Getting the fx graph of submodules, instead of 'call_module' nodes? compile	0	28	March 14, 2023
Questions about Pytorch 2.0 tutorial and TORCH_COMPILE_DEBUG=1 compile	3	174	March 9, 2023
Torch.compile() negative performance compile	1	204	March 9, 2023
How to register new backend compile	1	52	March 9, 2023
PT 2.0 - Are compiled models savable compile	4	288	March 9, 2023