Issues · Lightning-AI/lightning-thunder · GitHub

Label tracking meta-issue (edit me to get automatically CC'ed on issues!)
#72 · carmocca opened on Mar 25, 2024
1

Labels Milestones New issue

Refactor Triton cross-entropy executor: cleanup | Reduce activation memory in cross-entropy backward

#2818

· oonyshch opened

on Mar 23, 2026

Remove `xfail`s once PyTorch autograd stabilizes

#2807

· mattteochen opened

on Dec 16, 2025

`args_tensor_mask` has been removed from `torch.ops.higher_order.autograd_function_apply`

#2803

· mattteochen opened

on Dec 15, 2025

In-place ops on cloned tensor propagates to the original tensor

#2793

· shino16 opened

on Dec 10, 2025

Investigate `torch.Tensor.setitem` as unsupported in `thunderfx`

#2790

· mattteochen opened

on Dec 3, 2025

[symbolic values] Trace bakes in the initial value of input used for trace construction

symbolic values

#2789

· kshitij12345 opened

on Dec 3, 2025

`AssertionError` for mutation on intermediates being reordered after `return`

#2776

· shino16 opened

on Nov 27, 2025

[inference] High thunder.jit/thunderfx overhead for small module

#2775

· kshitij12345 opened

on Nov 27, 2025

Too many `update_aliases` after in-place op

#2768

· shino16 opened

on Nov 25, 2025

Errors on in-place ops on tensor aliases unresolved by proxy substitution

#2766

· shino16 opened

on Nov 24, 2025

Replace usage of nvFuser's broadcast_in_dim in executor with broadcast+expand ops

#2765

· IvanYashchuk opened

on Nov 23, 2025

thunderfx: Applying thunderfx should return a nn.Module like object.

#2757

· kshitij12345 opened

on Nov 19, 2025