[quote="jjohnson-arm, post:12, topic:6055"]
Is that right?
[/quote]
Yes, exactly right.
[quote="jjohnson-arm, post:13, topic:6055"]
I am wondering if it would be possible just to append some conversion entries
to the `output_map_index` instead and it would achieve the same thing?
[/quote]
I'm not completely sure what you mean here, but since `_get_op_inputs` looks
for the original Torch IR input names, we need to overwrite the input names or
add additional entries to `outputs` and `output_index_map`.
Overwrite can be done by `setDebugName` method. For the latter solution, we can
add
```
for torch_input_name, relay_var in zip(get_graph_input_names(script_module),
input_vars.values()):
output_index_map[torch_input_name] = len(outputs)
outputs.append(relay_var)
```
after
https://github.com/apache/incubator-tvm/blob/3aabbd9c30d247a31eb19ebf997d6074b14b5dd9/python/tvm/relay/frontend/pytorch.py#L1398-L1399
---
[Visit
Topic](https://discuss.tvm.ai/t/pytorch-frontend-graph-input-names-can-change-using-loaded-torchscript/6055/14)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https://discuss.tvm.ai/email/unsubscribe/9f7813373d9d82b953951ed468ddd43bce115ed9105e6f945c53bf89023506c8).
