LoRA shape mismatches should fail fast instead of corrupting GPU state and causing OOM/system instability

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

Either cross-compatibility, or error message indicating LoRA is incompatible with model architecture, operation aborts cleanly.

### Actual Behavior

ComfyUI logs ~1500 "lora key not loaded" warnings followed by ~228 shape mismatch errors, then continues execution anyway:
lora key not loaded: lora_unet_double_blocks_0_img_mod_lin.lora_up.weight
[...~1500 similar warnings...]
ERROR lora diffusion_model.single_blocks.X.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[...~228 similar errors, each layer 3x...]
Execution continues despite ~1700 logged problems. Corrupted tensor state gets pushed to GPU, causing OOM on 96GB card and system instability requiring reboot.

### Steps to Reproduce

Load Flux 2 Dev model
Apply a Flux 1 LoRA (architecturally incompatible - different tensor dimensions in single_blocks)
Attempt to run generation
Observe error spam followed by OOM and potential system instability

### Debug Logs

```powershell
[2026-01-14 01:57:06.591] got prompt
[2026-01-14 01:57:06.685] Using pytorch attention in VAE[2026-01-14 01:57:06.687] Using pytorch attention in VAE
[2026-01-14 01:57:06.859] VAE load device: cuda:0, offload device: cuda:0, dtype: torch.bfloat16
[2026-01-14 01:57:06.901] ============================================================
[2026-01-14 01:57:06.902] Phase 1: Loading CLIP
[2026-01-14 01:57:06.902] ============================================================
[2026-01-14 01:57:06.902] Loading CLIP: mistral_3_small_flux2_bf16.safetensors
[2026-01-14 01:57:06.967] [MultiGPU Core Patching] text_encoder_device_patched returning device: cuda:0 (current_text_encoder_device=cuda:0)
[2026-01-14 01:57:36.953] Requested to load Flux2TEModel_
[2026-01-14 01:57:36.961] loaded completely; 33080.59 MB loaded, full load: True
[2026-01-14 01:57:36.962] CLIP/text encoder model load device: cuda:0, offload device: cuda:0, current: cuda:0, dtype: torch.float16
[2026-01-14 01:57:37.019] [CHECK] Prompt encoded
[2026-01-14 01:57:37.019]
Phase 2: Unloading CLIP
[2026-01-14 01:57:41.072] [CHECK] CLIP unloaded
[2026-01-14 01:57:41.072]
Phase 3: Loading Diffusion Model
[2026-01-14 01:57:41.173] model weight dtype torch.bfloat16, manual cast: None
[2026-01-14 01:57:41.174] model_type FLUX
[2026-01-14 01:58:51.789] [CHECK] Model loaded
[2026-01-14 01:58:51.792]
[2026-01-14 01:58:51.918] lora key not loaded: lora_unet_double_blocks_0_img_mod_lin.alpha
[2026-01-14 01:58:51.918] lora key not loaded: lora_unet_double_blocks_0_img_mod_lin.lora_down.weight

[1470 lines snipped]

[2026-01-14 01:58:52.157] lora key not loaded: lora_unet_single_blocks_8_modulation_lin.lora_up.weight
[2026-01-14 01:58:52.157] lora key not loaded: lora_unet_single_blocks_9_modulation_lin.alpha
[2026-01-14 01:58:52.157] lora key not loaded: lora_unet_single_blocks_9_modulation_lin.lora_down.weight
[2026-01-14 01:58:52.157] lora key not loaded: lora_unet_single_blocks_9_modulation_lin.lora_up.weight
[2026-01-14 01:58:52.179] Requested to load Flux2
[2026-01-14 01:58:52.250] ERROR lora diffusion_model.single_blocks.9.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.261] ERROR lora diffusion_model.single_blocks.9.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.263] ERROR lora diffusion_model.single_blocks.9.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.274] ERROR lora diffusion_model.single_blocks.8.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.278] ERROR lora diffusion_model.single_blocks.8.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.278] ERROR lora diffusion_model.single_blocks.8.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.288] ERROR lora diffusion_model.single_blocks.7.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.292] ERROR lora diffusion_model.single_blocks.7.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.293] ERROR lora diffusion_model.single_blocks.7.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288
[2026-01-14 01:58:52.304] ERROR lora diffusion_model.single_blocks.6.linear1.weight shape '[55296, 6144]' is invalid for input of size 66060288

[208 lines snipped]

[2026-01-14 01:59:09.976] ERROR lora diffusion_model.single_blocks.26.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:10.403] Error on reading from the event loop self pipe
loop:
[2026-01-14 01:59:11.403] [Crystools ERROR] Could not get GPU temperature. Turning off this feature. The operating system has blocked the request.
[2026-01-14 01:59:19.273] ERROR lora diffusion_model.single_blocks.25.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:19.275] ERROR lora diffusion_model.single_blocks.25.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:19.276] ERROR lora diffusion_model.single_blocks.25.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:19.723] send error: Cannot write to closing transport
[2026-01-14 01:59:19.723] send error: Cannot write to closing transport
[2026-01-14 01:59:19.723] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:19.724] send error: Cannot write to closing transport
[2026-01-14 01:59:25.338] ERROR lora diffusion_model.single_blocks.24.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:25.343] ERROR lora diffusion_model.single_blocks.24.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:25.343] ERROR lora diffusion_model.single_blocks.24.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:27.319] ERROR lora diffusion_model.single_blocks.23.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:27.323] ERROR lora diffusion_model.single_blocks.23.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:27.324] ERROR lora diffusion_model.single_blocks.23.linear2.weight shape '[6144, 24576]' is invalid for input of size 47185920
[2026-01-14 01:59:31.436] Error on reading from the event loop self pipe
loop:
[2026-01-14 01:59:32.624] Error on reading from the event loop self pipe
loop:
[2026-01-14 01:59:32.625] Exception in thread Thread-23 (prompt_worker):
[2026-01-14 01:59:32.626] Traceback (most recent call last):
[2026-01-14 01:59:32.626] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 518, in execute
[2026-01-14 01:59:32.627] output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
[2026-01-14 01:59:32.627] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.632] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 329, in get_output_data
[2026-01-14 01:59:32.632] return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
[2026-01-14 01:59:32.632] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.636] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 303, in _async_map_node_over_list
[2026-01-14 01:59:32.636] await process_inputs(input_dict, i)
[2026-01-14 01:59:32.636] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 291, in process_inputs
[2026-01-14 01:59:32.636] result = f(**inputs)
[2026-01-14 01:59:32.636] ^^^^^^^^^^^
[2026-01-14 01:59:32.636] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy_api\internal\__init__.py", line 149, in wrapped_func
[2026-01-14 01:59:32.637] return method(locked_class, **inputs)
[2026-01-14 01:59:32.637] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.638] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy_api\latest\_io.py", line 1582, in EXECUTE_NORMALIZED
[2026-01-14 01:59:32.638] to_return = cls.execute(*args, **kwargs)
[2026-01-14 01:59:32.638] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.639] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy_extras\nodes_custom_sampler.py", line 950, in execute
[2026-01-14 01:59:32.640] samples = guider.sample(noise.generate_noise(latent), latent_image, sampler, sigmas, denoise_mask=noise_mask, callback=callback, disable_pbar=disable_pbar, seed=noise.seed)
[2026-01-14 01:59:32.640] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.643] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\samplers.py", line 1050, in sample
[2026-01-14 01:59:32.643] output = executor.execute(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed, latent_shapes=latent_shapes)
[2026-01-14 01:59:32.643] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.646] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\patcher_extension.py", line 112, in execute
[2026-01-14 01:59:32.647] return self.original(*args, **kwargs)
[2026-01-14 01:59:32.647] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.647] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\samplers.py", line 984, in outer_sample
[2026-01-14 01:59:32.648] self.inner_model, self.conds, self.loaded_models = comfy.sampler_helpers.prepare_sampling(self.model_patcher, noise.shape, self.conds, self.model_options)
[2026-01-14 01:59:32.648] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.651] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\sampler_helpers.py", line 130, in prepare_sampling
[2026-01-14 01:59:32.651] return executor.execute(model, noise_shape, conds, model_options=model_options, force_full_load=force_full_load)
[2026-01-14 01:59:32.651] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.653] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\patcher_extension.py", line 112, in execute
[2026-01-14 01:59:32.653] return self.original(*args, **kwargs)
[2026-01-14 01:59:32.653] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.654] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\sampler_helpers.py", line 138, in _prepare_sampling
[2026-01-14 01:59:32.654] comfy.model_management.load_models_gpu([model] + models, memory_required=memory_required + inference_memory, minimum_memory_required=minimum_memory_required + inference_memory, force_full_load=force_full_load)
[2026-01-14 01:59:32.655] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\model_management.py", line 715, in load_models_gpu
[2026-01-14 01:59:32.655] loaded_model.model_load(lowvram_model_memory, force_patch_weights=force_patch_weights)
[2026-01-14 01:59:32.655] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\model_management.py", line 520, in model_load
[2026-01-14 01:59:32.655] self.model_use_more_vram(use_more_vram, force_patch_weights=force_patch_weights)
[2026-01-14 01:59:32.656] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\model_management.py", line 550, in model_use_more_vram
[2026-01-14 01:59:32.656] return self.model.partially_load(self.device, extra_memory, force_patch_weights=force_patch_weights)
[2026-01-14 01:59:32.656] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.657] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\model_patcher.py", line 983, in partially_load
[2026-01-14 01:59:32.657] raise e
[2026-01-14 01:59:32.657] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\model_patcher.py", line 980, in partially_load
[2026-01-14 01:59:32.657] self.load(device_to, lowvram_model_memory=current_used + extra_memory, force_patch_weights=force_patch_weights, full_load=full_load)
[2026-01-14 01:59:32.658] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\comfy\model_patcher.py", line 780, in load
[2026-01-14 01:59:32.658] torch.cuda.synchronize()
[2026-01-14 01:59:32.658] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\cuda\__init__.py", line 1083, in synchronize
[2026-01-14 01:59:32.659] return torch._C._cuda_synchronize()
[2026-01-14 01:59:32.659] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.659] torch.AcceleratorError: CUDA error: out of memory
Search for `cudaErrorMemoryAllocation' in https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html for more information.
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
[2026-01-14 01:59:32.659]
[2026-01-14 01:59:32.659]
[2026-01-14 01:59:32.659] During handling of the above exception, another exception occurred:
[2026-01-14 01:59:32.659]
[2026-01-14 01:59:32.659] Traceback (most recent call last):
[2026-01-14 01:59:32.659] File "C:\home\creation\studio_1\sm_data\Assets\Python\cpython-3.12.11-windows-x86_64-none\Lib\threading.py", line 1075, in _bootstrap_inner
[2026-01-14 01:59:32.660] self.run()
[2026-01-14 01:59:32.660] File "C:\home\creation\studio_1\sm_data\Assets\Python\cpython-3.12.11-windows-x86_64-none\Lib\threading.py", line 1012, in run
[2026-01-14 01:59:32.660] self._target(*self._args, **self._kwargs)
[2026-01-14 01:59:32.661] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\main.py", line 231, in prompt_worker
[2026-01-14 01:59:32.661] e.execute(item[2], prompt_id, extra_data, item[4])
[2026-01-14 01:59:32.661] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 673, in execute
[2026-01-14 01:59:32.662] asyncio.run(self.execute_async(prompt, prompt_id, extra_data, execute_outputs))
[2026-01-14 01:59:32.662] File "C:\home\creation\studio_1\sm_data\Assets\Python\cpython-3.12.11-windows-x86_64-none\Lib\asyncio\runners.py", line 195, in run
[2026-01-14 01:59:32.662] return runner.run(main)
[2026-01-14 01:59:32.662] ^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.662] File "C:\home\creation\studio_1\sm_data\Assets\Python\cpython-3.12.11-windows-x86_64-none\Lib\asyncio\runners.py", line 118, in run
[2026-01-14 01:59:32.663] return self._loop.run_until_complete(task)
[2026-01-14 01:59:32.663] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.663] File "C:\home\creation\studio_1\sm_data\Assets\Python\cpython-3.12.11-windows-x86_64-none\Lib\asyncio\base_events.py", line 691, in run_until_complete
[2026-01-14 01:59:32.664] return future.result()
[2026-01-14 01:59:32.664] ^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.664] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 722, in execute_async
[2026-01-14 01:59:32.664] result, error, ex = await execute(self.server, dynamic_prompt, self.caches, node_id, extra_data, executed, prompt_id, execution_list, pending_subgraph_results, pending_async_nodes, ui_node_outputs)
[2026-01-14 01:59:32.665] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.666] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 596, in execute
[2026-01-14 01:59:32.667] input_data_formatted[name] = [format_value(x) for x in inputs]
[2026-01-14 01:59:32.667] ^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.667] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\execution.py", line 407, in format_value
[2026-01-14 01:59:32.667] return str(x)
[2026-01-14 01:59:32.667] ^^^^^^
[2026-01-14 01:59:32.668] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor.py", line 568, in __repr__
[2026-01-14 01:59:32.668] return torch._tensor_str._str(self, tensor_contents=tensor_contents)
[2026-01-14 01:59:32.668] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.669] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 722, in _str
[2026-01-14 01:59:32.669] return _str_intern(self, tensor_contents=tensor_contents)
[2026-01-14 01:59:32.669] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.669] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 643, in _str_intern
[2026-01-14 01:59:32.670] tensor_str = _tensor_str(self, indent)
[2026-01-14 01:59:32.670] ^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.670] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 375, in _tensor_str
[2026-01-14 01:59:32.670] formatter = _Formatter(get_summarized_data(self) if summarize else self)
[2026-01-14 01:59:32.670] ^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.671] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 413, in get_summarized_data
[2026-01-14 01:59:32.671] return torch.stack([get_summarized_data(x) for x in self])
[2026-01-14 01:59:32.671] ^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.671] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 411, in get_summarized_data
[2026-01-14 01:59:32.671] return torch.stack([get_summarized_data(x) for x in (start + end)])
[2026-01-14 01:59:32.671] ^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.672] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 411, in get_summarized_data
[2026-01-14 01:59:32.672] return torch.stack([get_summarized_data(x) for x in (start + end)])
[2026-01-14 01:59:32.672] ^^^^^^^^^^^^^^^^^^^^^^
[2026-01-14 01:59:32.672] File "C:\home\creation\studio_1\sm_data\Packages\ComfyUI\venv\Lib\site-packages\torch\_tensor_str.py", line 401, in get_summarized_data
[2026-01-14 01:59:32.673] return torch.cat(
[2026-01-14 01:59:32.673] ^^^^^^^^^^
[2026-01-14 01:59:32.673] torch.AcceleratorError: CUDA error: out of memory
Search for `cudaErrorMemoryAllocation' in https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html for more information.
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
[2026-01-14 01:59:32.673]
```

### Other

Environment: ComfyUI 0.9.1, PyTorch 2.9.1+cu130, CUDA 13.0, NVIDIA RTX PRO 6000 (96GB), Windows 11
~1700 total warnings/errors logged before crash. Suggest failing after first incompatibility detection rather than exhaustively logging every mismatch.