MiniMax H3 Mem Eff Sage Attention Patch ERROR: Execution Failed.
- Dominant language
- Python
- Stars
- 133k
- Forks
- 15.7k
- Avg merge
- 1d 7h
- Merged PRs (30d)
- 158
Description
### Custom Node Testing
- [ ] 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)
### Your question
The issue arises because the MiniMax H3 Memory Efficient Sage Attention Patch node contains a flaw in its internal validation logic. Even with the essential files updated and the environment running Torch 2.13.0+cu130 alongside Sage Attention 2.2.0 (sageattention-2.2.0+cu130torch2.10.0andhigher.post6-cp310), the node fails when attempting to apply the optimization during video generation, as it is unable to correctly recognize the CUDA version or architecture due to a bug in the node's own check. Without the node, the video is generated without issues.
### Logs
```powershell
[INFO] [ComfyUI-Manager] The ComfyRegistry cache update is still in progress, so an outdated cache is being used.
FETCH DATA from: C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\.venv\Lib\site-packages\comfyui_manager\custom-node-list.json [DONE]
[INFO] got prompt
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float16
[INFO] Found quantization metadata version 1
[INFO] Using MixedPrecisionOps for text encoder
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
[INFO] Requested to load MiniMaxH3TEModel_
[INFO] Model MiniMaxH3TEModel_ prepared for dynamic VRAM loading. 25882MB Staged. 0 patches attached. Force pre-loaded 310 weights: 1272 KB.
[INFO] Requested to load MiniMaxH3VideoVAE
[INFO] Model MiniMaxH3VideoVAE prepared for dynamic VRAM loading. 4965MB Staged. 0 patches attached. Force pre-loaded 128 weights: 348 KB.
[INFO] Found quantization metadata version 1
[INFO] Detected mixed precision quantization
[INFO] Using mixed precision operations
[INFO] Native ops: float8_e4m3fn, int8_tensorwise, convrot_w4a4, asym_w4a8_int8, float8_e5m2 , emulated ops: mxfp8, nvfp4
[INFO] model weight dtype torch.bfloat16, manual cast: torch.bfloat16
[INFO] model_type FLOW_AV
[ERROR] !!! Exception during processing !!! sageattention is not new enough version or could not determine CUDA architecture, cannot apply MiniMax H3 Memory Efficient Sage Attention Patch.
[ERROR] Traceback (most recent call last):
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\execution.py", line 545, in execute
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)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\execution.py", line 344, in get_output_data
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)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\execution.py", line 318, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\execution.py", line 306, in process_inputs
result = f(**inputs)
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\comfy_api\internal\__init__.py", line 149, in wrapped_func
return method(locked_class, **inputs)
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\comfy_api\latest\_io.py", line 1990, in EXECUTE_NORMALIZED
to_return = cls.execute(*args, **kwargs)
File "C:\ComfyUI\ComfyUI Install\ComfyUI\ComfyUI\custom_nodes\comfyui-kjnodes\nodes\ltxv_nodes.py", line 2183, in execute
raise RuntimeError("sageattention is not new enough version or could not determine CUDA architecture, cannot apply MiniMax H3 Memory Efficient Sage Attention Patch.")
RuntimeError: sageattention is not new enough version or could not determine CUDA architecture, cannot apply MiniMax H3 Memory Efficient Sage Attention Patch.
```
### Other
_No response_
Contributor guide
Research direction
Start at custom_nodes/comfyui-kjnodes/nodes/ltxv_nodes.py:2183, where the MiniMax H3 Memory Efficient Sage Attention Patch raises the validation error. Reproduce with the reported Torch 2.13.0+cu130 and Sage Attention 2.2.0 environment, then inspect how the CUDA architecture and Sage Attention version are checked. Done means the valid environment is accepted and MiniMax H3 video generation proceeds with the patch enabled.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- ai
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 68/100