mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-06 12:20:52 +01:00
f36f285953
This PR implements the semantics change to `torch._dynamo.error_on_graph_break`: - ~`torch.compile` now has a new `error_on_graph_break` kwarg that serves as a lower-priority toggle for erroring/continuing on graph breaks~ - `error_on_graph_break` is a new internal `torch.compile `setting that is lower-priority than `fullgraph`. It allows the user to toggle erroring/continuing on graph breaks. - `error_on_graph_break` does nothing when `fullgraph=True` - `error_on_graph_break` does NOT guarantee a single graph Followup [DONE]: need to change the programming model docs to reflect the 3 graph break modes for compilation: - `fullgraph=True`: enforce one graph, no graph breaks, cannot be toggled - `fullgraph=False, error_on_graph_break=True`: errors on graph breaks, latter can be toggled during compile time - `fullgraph=False, error_on_graph_break=False`: resumes tracing on graph breaks, latter can be toggled during compile time Pull Request resolved: https://github.com/pytorch/pytorch/pull/161747 Approved by: https://github.com/mlazos ghstack dependencies: #161739 |
||
|---|---|---|
| .. | ||
| _static | ||
| header_code.py | ||
| programming_model.common_graph_breaks.md | ||
| programming_model.compiler_disable.md | ||
| programming_model.custom_ops.md | ||
| programming_model.dynamo_core_concepts.md | ||
| programming_model.dynamo_nonstrict_trace.md | ||
| programming_model.error_on_graph_break.md | ||
| programming_model.fullgraph_false.md | ||
| programming_model.fullgraph_true.md | ||
| programming_model.graph_breaks_index.md | ||
| programming_model.md | ||
| programming_model.nested_graph_breaks.md | ||
| programming_model.non_strict_tracing_model.md | ||
| programming_model.observability.md | ||
| programming_model.recompilation.md | ||
| programming_model.reporting_issues.md | ||
| programming_model.skipped_functions.md | ||
| programming_model.where_to_apply_compile.md | ||