Repository navigation
Description
Describe the bug
Two pipeline families in the Qwen-Image lineage still build their RoPE table with torch.polar and then index/multiply it as complex64 on the device, with no real-valued fallback:
transformer_qwenimage21.py—QwenImage21Rope.forwardhardcodesuse_real=Falsetransformer_z_image.py—torch.polar+view_as_complex
transformer_qwenimage.py (Qwen-Image 1.0) already has ROPE_PER_DEVICE from the AWS Neuron work in #13718, but only "neuron" is registered, so other accelerators keep taking the complex path.
On backends without complex64 kernels this is a hard failure rather than a slowdown, and it affects three targets, not one:
- Ascend NPU (torch_npu) — the op-plugin has no complex64
index/mul.QwenImage21Rope.forwardraisesaclnnIndex EZ1001as soon as the complex frequency table is indexed on device. Qwen-Image 2.1 and Z-Image currently cannot run on Ascend NPUs without an out-of-tree monkey-patch. - CoreML export —
torch.polarhas no MIL equivalent (RoPE implemetation, Torch.polar op not available in MIL requesting support apple/coremltools#2258). torch.compile— inductor cannot codegen complex ops (Complex numbers in Qwen/Z-Image Image pipeline incompatible with torch.compile #12668), which matches this same code path.
CUDA eager is unaffected, which is presumably why this has stayed low-pressure upstream.
I understand #12668 already tracks the complex-op problem from the torch.compile angle; opening this as a separate issue because the backend failure mode and the fix scope are different (see "Describe the solution you'd like"). Happy to close this if you would rather consolidate into #12668.
Reproduction
The failure is device-specific, but the code path is easy to isolate. On an Ascend NPU with a from-source diffusers checkout:
import torch
from diffusers.models.transformers.transformer_qwenimage21 import QwenImage21Rope
# Any small instantiation of the rope module is enough; the failure happens at the
# first on-device indexing of the complex frequency table.
rope = QwenImage21Rope(...).to("npu") # or .to(torch.device("npu"))
rope(x, freqs_cis) # -> aclnnIndex EZ1001A backend-agnostic sanity check for the same assumption (runs anywhere, no model weights needed):
import torch
freqs = torch.polar(torch.ones(8, 4), torch.rand(8, 4)).to("npu") # complex64 on device
freqs[torch.arange(8).to("npu")] # index -> unsupported on NPUDescribe the solution you'd like
Extend the existing #13718 pattern instead of adding a new one:
- port
ROPE_PER_DEVICE+ the real-valuedapply_rotary_emb_qwen_neuronpath intotransformer_qwenimage21.pyandtransformer_z_image.py; - keep the frequency table and its indexing on CPU, crossing the device boundary as reals (
_get_device_freqsalready does this for Neuron); - register
"npu"inROPE_PER_DEVICEfor Qwen-Image 1.0.
Context
I hit this while bringing up image models on a single Ascend 910B2; Qwen-Image 2.1 was the only pipeline in the set that could not run at all without patching.
Disclosure, and a note on scope: I opened PR #14905 for exactly this in September and it was auto-closed because it did not link an issue — that is on me, not on the change. Before re-opening anything I would like an explicit ack here, as the AI-assisted contribution section asks. I am also happy to split the work into one PR per pipeline instead of one size/L PR covering both, if that is easier to review.
Logs
RuntimeError: The inner error is reported as above. The process exits for this inner error. (Inner error: EZ1001: ... aclnnIndex ...)
File "diffusers/models/transformers/transformer_qwenimage21.py", line ..., in forward
...
(The exact op-level message from aclnnIndex is the one that names the unsupported complex64 dtype.)
System Info
- diffusers: main (from-source checkout at the time of report; PR #14905 was opened 2026-09-29)
- platform: Linux aarch64
- Python: 3.12.13
- torch: 2.10.0+cpu
- torch_npu: 2.10.0.post4
- CANN: 9.1.0
- accelerator: 1x Ascend 910B2 (HBM 64 GB)
Who can help?
Anyone with familiarity with the Qwen-Image RoPE implementations, or with Apple/Neuron backends where the same real-valued path already exists.
Activity