# Accelerate 1.15.0 — warning_once keeps adapter instances alive (measured)

> Source: <https://discuss.huggingface.co/t/accelerate-1-15-0-warning-once-keeps-adapter-instances-alive-measured/191953#post_2>
> Published: 2026-11-18 09:26:15+00:00

**Recommendation:** When converting LoRA adapters across output-order conventions, permute the output rows of `lora_B` and leave `lora_A` unchanged, then verify both that `A` matches the original and that `B_converted @ A_converted` matches `B @ A` with a recorded maximum absolute difference.
More in the playbook "Fine-tuning and data": <https://wpnews.pro/playbooks/training-data>

My mistake… The following is a reply to the thread about Diffusers: [Diffusers #15305 — tested the fix in PR #15307: exact](https://discuss.huggingface.co/t/diffusers-15305-tested-the-fix-in-pr-15307-exact/191952)

For now, I took a quick look at a few related implementations and earlier fixes:

Thanks for putting the result in the forum, and for being precise about *which* code you tested. Your report is especially useful because it keeps three things separate: the released `diffusers==0.43.0` behavior, the behavior of the proposed fix at the PR head you tested, and the fact that `lora_A` itself was left unchanged. Those are substantially more informative observations than a generic “the conversion works now” claim.

My reading of [issue #15305](https://github.com/huggingface/diffusers/issues/15305), [PR #15307](https://github.com/huggingface/diffusers/pull/15307), and the corresponding model/converter implementations is that **the output ordering is what must change, not the LoRA rank coordinates**. In the affected final-layer mapping, the source convention is `(shift, scale)` and the Diffusers target convention is `(scale, shift)`. With a conventional two-factor LoRA update `delta_W = B @ A`, a permutation of output channels belongs on the output rows of `B`. `A` projects into the rank space; its rows are *not* the source model’s shift and scale output channels.

That gives the small but important invariant:

```
source delta:       B @ A
target delta:       swap_output_halves(B @ A)
converted factors:  swap_output_halves(B) @ A
```

The same mapping can be written as `P(BA) = (PB)A`, where `P` swaps the two output halves. Under that specific contract, `A` is deliberately unchanged. This is the reason the particular result in your report is meaningful, not just the fact that a numerical comparison returned zero.

As reported in [your test](https://discuss.huggingface.co/t/diffusers-15305-tested-the-fix-in-pr-15307-exact/191952), stock `0.43.0` gave a maximum absolute delta difference of about `39.0527`, whereas the tested head of #15307 gave `0.0`, with the converted `lora_A` equal to the original. I have not independently rerun that Diffusers converter comparison here, so I would treat those numbers as **your measured result for that reproducer and head**, supported by the algebra and the PR’s own converter-specific tests, rather than as an assertion about every checkpoint or later release.

For the exact `final_layer.adaLN_modulation.1` LoRA-A/LoRA-B issue, the default reading seems straightforward:

`lora_A` unchanged and reorder the output rows of `lora_B`. If this conversion path also accepts a bias attached to `0.43.0` with a particular PR head. At my November 18, 2026 check, `P(BA) = (PB)A` argument is strong for this two-factor, output-row permutation. A full-weight checkpoint, a non-LoRA factorization, or a converter with a different tensor layout may require its own mapping. That is a scope distinction, not a contradiction of the reported reproduction.
For a future reader who encounters a different result, I’d start with three cheap checks in this order. None is intended as extra homework for the original poster.

`A` matches the original `B_converted @ A_converted` matches the `B @ A`. A single `allclose` boolean does not say which half of that contract held. Recording a maximum absolute difference as well makes failures easier to diagnose.
To me, the strongest part of your report is the combination of a fixed reproducer, an explicit version comparison, the unchanged A factor, and a composed update that matches the intended output order. The existing PR tests and the earlier, separate fixes make the reasoning easy to follow. That is already a clear and reusable technical result, with no need to turn it into a general claim about all adapter formats or to ask you to run a larger model.
