torch.fx.experimental.proxy_tensor.maybe_enable_thunkify¶
- torch.fx.experimental.proxy_tensor.maybe_enable_thunkify()[source]¶
Within this context manager, if you are doing make_fx tracing, we will thunkify all SymNode compute and avoid tracing it into the graph unless it is actually needed. You should prefer to avoid using this as much as possible, as lazy evaluation of SymNode tracing can lead to long chains of thunks which will stack overflow if you evaluate them. However, this is currently sometimes necessary as there are buggy parts of PT2 which will fail with “s0 is not tracked with proxy” error due to insufficient tracing of SymNode computation.
- Return type
Generator[None, None, None]