UnboundLocalError: local variable 'h' referenced before assignment

[LeonZhao28]

A exception occurs when using sd-webui to generate a image.
This is the stack:

dump json failed,data:{'file': <_io.TextIOWrapper name='<stderr>' mode='w' encoding='utf-8'>}
    Traceback (most recent call last):
      File "/app/modules/api/api.py", line 204, in exception_handling
        return await call_next(request)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 84, in call_next
        raise app_exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 70, in coro
        await self.app(scope, receive_or_disconnect, send_no_error)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 108, in __call__
        response = await self.dispatch_func(request, call_next)
      File "/app/modules/api/api.py", line 151, in log_and_time
        res: Response = await call_next(req)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 84, in call_next
        raise app_exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 70, in coro
        await self.app(scope, receive_or_disconnect, send_no_error)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 108, in __call__
        response = await self.dispatch_func(request, call_next)
      File "/app/modules/he_ware/run_mode_middle_ware.py", line 28, in dispatch
        response = await call_next(request)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 84, in call_next
        raise app_exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 70, in coro
        await self.app(scope, receive_or_disconnect, send_no_error)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 108, in __call__
        response = await self.dispatch_func(request, call_next)
      File "/app/modules/he_ware/progress_res_middle_ware.py", line 24, in dispatch
        response = await call_next(request)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 84, in call_next
        raise app_exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 70, in coro
        await self.app(scope, receive_or_disconnect, send_no_error)
dump json failed,data:{'file': <_io.TextIOWrapper name='<stderr>' mode='w' encoding='utf-8'>}
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 108, in __call__
        response = await self.dispatch_func(request, call_next)
      File "/app/modules/he_ware/task_state_middle_ware.py", line 52, in dispatch
        response = await call_next(request)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 84, in call_next
        raise app_exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 70, in coro
        await self.app(scope, receive_or_disconnect, send_no_error)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 108, in __call__
        response = await self.dispatch_func(request, call_next)
      File "/app/venv/lib/python3.10/site-packages/starlette_prometheus/middleware.py", line 57, in dispatch
        raise e from None
      File "/app/venv/lib/python3.10/site-packages/starlette_prometheus/middleware.py", line 53, in dispatch
        response = await call_next(request)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 84, in call_next
        raise app_exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/base.py", line 70, in coro
        await self.app(scope, receive_or_disconnect, send_no_error)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/cors.py", line 84, in __call__
        await self.app(scope, receive, send)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/gzip.py", line 24, in __call__
        await responder(scope, receive, send)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/gzip.py", line 44, in __call__
        await self.app(scope, receive, self.send_with_gzip)
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 79, in __call__
        raise exc
      File "/app/venv/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 68, in __call__
        await self.app(scope, receive, sender)
      File "/app/venv/lib/python3.10/site-packages/fastapi/middleware/asyncexitstack.py", line 21, in __call__
        raise e
      File "/app/venv/lib/python3.10/site-packages/fastapi/middleware/asyncexitstack.py", line 18, in __call__
        await self.app(scope, receive, send)
      File "/app/venv/lib/python3.10/site-packages/starlette/routing.py", line 718, in __call__
        await route.handle(scope, receive, send)
      File "/app/venv/lib/python3.10/site-packages/starlette/routing.py", line 276, in handle
        await self.app(scope, receive, send)
      File "/app/venv/lib/python3.10/site-packages/starlette/routing.py", line 66, in app
        response = await func(request)
      File "/app/venv/lib/python3.10/site-packages/fastapi/routing.py", line 237, in app
        raw_response = await run_endpoint_function(
      File "/app/venv/lib/python3.10/site-packages/fastapi/routing.py", line 165, in run_endpoint_function
        return await run_in_threadpool(dependant.call, **values)
      File "/app/venv/lib/python3.10/site-packages/starlette/concurrency.py", line 41, in run_in_threadpool
        return await anyio.to_thread.run_sync(func, *args)
      File "/app/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 33, in run_sync
        return await get_asynclib().run_sync_in_worker_thread(
      File "/app/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 877, in run_sync_in_worker_thread
        return await future
      File "/app/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 807, in run
        result = context.run(func, *args)
      File "/app/modules/api/api.py", line 457, in img2imgapi
        processed = process_images(p)
      File "/app/modules/processing.py", line 733, in process_images
        res = process_images_inner(p)
      File "/app/extensions/sd-webui-controlnet/scripts/batch_hijack.py", line 42, in processing_process_images_hijack
        return getattr(processing, '__controlnet_original_process_images_inner')(p, *args, **kwargs)
      File "/app/modules/processing.py", line 868, in process_images_inner
        samples_ddim = p.sample(conditioning=p.c, unconditional_conditioning=p.uc, seeds=p.seeds, subseeds=p.subseeds, subseed_strength=p.subseed_strength, prompts=p.prompts)
      File "/app/modules/processing.py", line 1529, in sample
        samples = self.sampler.sample_img2img(self, self.init_latent, x, conditioning, unconditional_conditioning, image_conditioning=self.image_conditioning)
      File "/app/modules/sd_samplers_kdiffusion.py", line 188, in sample_img2img
        samples = self.launch_sampling(t_enc + 1, lambda: self.func(self.model_wrap_cfg, xi, extra_args=self.sampler_extra_args, disable=False, callback=self.callback_state, **extra_params_kwargs))
      File "/app/modules/sd_samplers_common.py", line 261, in launch_sampling
        return func()
      File "/app/modules/sd_samplers_kdiffusion.py", line 188, in <lambda>
        samples = self.launch_sampling(t_enc + 1, lambda: self.func(self.model_wrap_cfg, xi, extra_args=self.sampler_extra_args, disable=False, callback=self.callback_state, **extra_params_kwargs))
      File "/app/venv/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
        return func(*args, **kwargs)
      File "/app/repositories/k-diffusion/k_diffusion/sampling.py", line 651, in sample_dpmpp_2m_sde
        h_last = h
    UnboundLocalError: local variable 'h' referenced before assignment
---

This is the code link:

k-diffusion/k_diffusion/sampling.py

Line 651 in 0455157

h_last = h

Change the code like this can avoid the exception, but I'm not sure whether the logic is correct.

The request parameters are sent as the files
req.json

[zhangj1an]

I run into the same issue.

UnboundLocalError: h in sample_dpmpp_3m_sde when the only step is terminal (steps=1 + append_zero)

Description

sample_dpmpp_3m_sde can raise:

UnboundLocalError: cannot access local variable 'h' where it is not associated with a value

when an iteration uses the terminal branch sigmas[i + 1] == 0 without any earlier iteration having run the else branch that assigns h (see minimal repro below).

What happens

1. get_sigmas_polyexponential ends with append_zero, so the last sigma in the schedule is 0.
2. On any iteration with sigmas[i + 1] == 0, the code sets x = denoised and does not assign h for that iteration.
3. After the if / else, the loop unconditionally runs:
   • denoised_1, denoised_2 = denoised, denoised_1
   • h_1, h_2 = h, h_1
4. UnboundLocalError: if this is the only loop iteration (e.g. steps=1 → len(sigmas)==2), h is never assigned in the function.
   For steps>1, h may still exist from the previous iteration (stale), so the same lines may not crash but still look like the wrong place to advance (h_1, h_2) after a pure terminal denoise step.

Expected behavior

On the terminal denoising step (sigmas[i + 1] == 0), sampling should end with x = denoised and must not read h or advance (h_1, h_2) as if an SDE step had run.

Suggested fix

Only update denoised_1, denoised_2, h_1, h_2 when sigmas[i + 1] != 0 (i.e. when the else branch ran and h was assigned)—for example, indent those two lines inside the non-terminal branch.

Location in code

File: k_diffusion/sampling.py, function sample_dpmpp_3m_sde (around the loop over len(sigmas) - 1, immediately after the if sigmas[i + 1] == 0 / else block).

Minimal repro (k-diffusion only)

Needs torch, k-diffusion, and a CUDA device (or change device to "cpu"; behavior is the same for this bug).

python -c "
import torch
import k_diffusion as K

device = 'cuda' if torch.cuda.is_available() else 'cpu'
steps = 1  # critical: one iteration, sigma goes straight to 0
sigmas = K.sampling.get_sigmas_polyexponential(steps, 0.03, 1000.0, 1.0, device=device)

def model_fn(x, sigma, **kwargs):
    return x

denoiser = K.external.VDenoiser(model_fn)
x = torch.randn(1, 2, 32, device=device) * sigmas[0]
K.sampling.sample_dpmpp_3m_sde(denoiser, x, sigmas, extra_args={}, disable=True)
"

Expected: UnboundLocalError: cannot access local variable 'h' where it is not associated with a value.

Why steps=1: get_sigmas_polyexponential(1, ...) returns two sigmas [sigma_max, 0]; the sampler loop runs once; that iteration hits sigmas[1] == 0 immediately, so h is never set, then lines h_1, h_2 = h, h_1 run.

Environment

• OS: Linux 6.5.0-35-generic (added support for non square Images #35~22.04.1-Ubuntu SMP), x86_64
• Python: 3.11.10
• k-diffusion: 0.1.1 (pip)
• PyTorch: 2.10.0+cu128
• CUDA (PyTorch build): 12.8
• torch.cuda.is_available(): True