Ay bara CUDA home not found

### Expected Behavior

Get a video

### Actual Behavior

Like WTF, i feel like giving up and go grow potatoes in a field somewhere, really, after a whole day of rendering it comes to the GIMMVFI model and this is all I have to show for, 'CUDA_HOME' not set, unable to find cuda-toolkit installation.

I sat here waiting for longer than 24 hrs for a lousy:
'CUDA_HOME' not set, unable to find cuda-toolkit installation.
like really?

### Steps to Reproduce

press queue prompt, and go to bed, sleep in and wait for f*ckin ever untill it gets to the GIMMVFI model node and then you get the error that only I have apparently, there is no known issue like this, no error in advance that I need to download this model or have some version of TORCH installed or anything, all you have to do is wait and get the golden ticket of error codes with no solution anywhere

### Debug Logs

```powershell
[5/04/2025 21:12:00] INFO PromptTask 6bae471c04ec4285b73487281e5a5bf3
Input: You are a large language model specialized in generating a short 5 second video
prompt. You will take the input prompt and turn it into a single coherent prompt
describing a single scene. Keep body and camera motions simple but dynamic. Keep the
prompt short and simple, and and do not add any introduction sentances or anything other
than the short video prompt.
HTTP Request: POST http://127.0.0.1:11434/api/chat "HTTP/1.1 200 OK"
[5/04/2025 21:12:06] INFO PromptTask 6bae471c04ec4285b73487281e5a5bf3
Output: A kitten playfully chases a laser pointer around a living room, its paws and tail
flickering in the beam's bright light, while a houseplant subtly sways in the background.
INFO PromptTask 6bae471c04ec4285b73487281e5a5bf3
Input: Make an elaborate and detailed prompt for image generation, from the input given,
take the information from the input string and you make a prompt that holds all the
elements given, but you add none of your own fantasies or notes, ONLY USE THE INFORMATION
GIVEN and take every detail in to account, Only the prompt no other commentary, thought
processes or notes

Remove any logos or watermarks

The image is of a young woman standing on a beach. She is wearing a red bikini with white
polka dots. The bikini top has a halter neckline and thin straps. She has long dark hair
with bangs and is smiling at the camera. The background shows a beautiful blue ocean and a
small island in the distance. The sky is clear and the weather appears to be sunny and
warm.
HTTP Request: POST http://127.0.0.1:11434/api/chat "HTTP/1.1 200 OK"
[5/04/2025 21:12:09] INFO PromptTask 6bae471c04ec4285b73487281e5a5bf3
Output: A young woman, dressed in a red bikini with white polka dots and a halter
neckline, stands on a pristine beach with long dark hair and bangs. She is smiling
directly at the camera while gazing out onto a clear blue ocean that stretches towards a
distant island. The weather appears to be sunny and warm with a cloudless sky above. The
background features no logos or watermarks.
in_channels: 36
Model type: i2v, num_heads: 40, num_layers: 40
Model variant detected: i2v_480
TeaCache: Using cache device: cpu
model_type FLOW
Using accelerate to load and assign model weights to device...
Loading transformer parameters to cpu: 100%|██████████████████████████████████████| 1303/1303 [00:02<00:00, 556.90it/s]
image_cond torch.Size([20, 26, 90, 90])
Seq len: 52650
previewer: None
Swapping 20 transformer blocks
Initializing block swap: 100%|█████████████████████████████████████████████████████████| 40/40 [00:44<00:00, 1.12s/it]
----------------------
Block swap memory summary:
Transformer blocks on cpu: 7705.21MB
Transformer blocks on cuda:0: 7705.21MB
Total memory used by transformer blocks: 15410.43MB
Non-blocking memory transfer: True
----------------------
Sampling 101 frames at 720x720 with 15 steps
60%|█████████████████████████████████████████████▌ | 9/15 [11:56:38<7:21:46, 4417.70s/it]Loras page loaded successfully with 984 items
Unauthorized access to resource: https://civitai.com/api/download/models/1626588 (Status 401)
Early access download failed: Invalid or missing CivitAI API key, or early access restriction.
100%|█████████████████████████████████████████████████████████████████████████████| 15/15 [21:39:58<00:00, 5199.93s/it]
TeaCache skipped: 1 conditional steps: [14]
TeaCache skipped: 1 unconditional steps: [14]
Allocated memory: memory=0.131 GB
Max allocated memory: max_memory=18.051 GB
Max reserved memory: max_reserved=20.688 GB
VAE decoding: 100%|██████████████████████████████████████████████████████████████████████| 1/1 [00:57<00:00, 57.10s/it]
Unload Model:
- Unloading all models...
- Clearing Cache...
Downloading GMMI-VFI model to: D:\ComfyUI_windows_portable\ComfyUI\models\interpolation\gimm-vfi\gimmvfi_f_arb_lpips_fp32.safetensors
Fetching 1 files: 0%| | 0/1 [00:00", line 753, in setdefault
File "", line 719, in __setitem__
File "", line 779, in check_str
TypeError: str expected, not NoneType

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 327, in execute
output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 202, in get_output_data
return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 174, in _map_node_over_list
process_inputs(input_dict, i)
File "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 163, in process_inputs
results.append(getattr(obj, func)(**inputs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI\nodes.py", line 213, in interpolate
all_outputs = gimmvfi_model(xs, coord_inputs, t=timesteps, ds_factor=ds_factor)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\nn\modules\module.py", line 1739, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\nn\modules\module.py", line 1750, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI\gimmvfi\generalizable_INR\gimmvfi_f.py", line 334, in forward
normal_inr_flows = self.predict_flow(normal_flows, coord, t, flows)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI\gimmvfi\generalizable_INR\gimmvfi_f.py", line 158, in predict_flow
tmp_pixel_latent_0 = softsplat(
^^^^^^^^^^
File "D:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\_dynamo\eval_frame.py", line 745, in _fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI\gimmvfi\generalizable_INR\modules\softsplat.py", line 321, in softsplat
tenOut = softsplat_func.apply(tenIn, tenFlow)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\autograd\function.py", line 575, in apply
return super().apply(*args, **kwargs) # type: ignore[misc]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\amp\autocast_mode.py", line 514, in decorate_fwd
return fwd(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI\gimmvfi\generalizable_INR\modules\softsplat.py", line 370, in forward
cuda_launch(
File "cupy\\_util.pyx", line 64, in cupy._util.memoize.decorator.ret
File "D:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\_dynamo\eval_frame.py", line 745, in _fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI\gimmvfi\generalizable_INR\modules\softsplat.py", line 270, in cuda_launch
raise RuntimeError("'CUDA_HOME' not set, unable to find cuda-toolkit installation.")
RuntimeError: 'CUDA_HOME' not set, unable to find cuda-toolkit installation.

Prompt executed in 78742.26 seconds
```

### Other

_No response_