Not got custom node value.

[yaosheng216]

Expected Behavior

Hi, author.
First, thanks for your work.
I met a question during of development custom node.
Below is my code, second time to execute workflow(update other node, this node value don't modify), why not got random_seed value of this node(this node of other values can got).

Wish your can answer my question.
Thank you!

Actual Behavior

second time to execute workflow(update other node, this node value don't modify), why not got random_seed value of this node(this node of other values can got)

Steps to Reproduce

second time to execute workflow(update other node, this node value don't modify), why not got random_seed value of this node(this node of other values can got)

Debug Logs

got prompt
parameters is -> {'sample': 'ddim', 'seed': 666, 'sample_steps': 5, 'guide_scale': 5.5, 'guide_rescale': 0.5, 'discretization': 'trailing', 'target_size_as_tuple': [1024, 1024]}
data is -> [{'prompt': 'Change the style of {image} to colored pencil style', 'negative_prompt': '', 'image': [<PIL.Image.Image image mode=RGB size=1024x1024 at 0x7F74A90BEBF0>], 'sample': 'ddim', 'sample_steps': 5, 'guide_scale': 5.5, 'guide_rescale': 0.5, 'discretization': 'trailing', 'target_size_as_tuple': [1024, 1024]}, {'diffusion_model': 'ACE_0.6B_512_ACE', 'first_stage_model': 'ACE_0.6B_512_AutoencoderKL', 'cond_stage_model': 'ACE_0.6B_512_T5EmbedderHF', 'seed': 666}]
/home/yaosheng/.local/lib/python3.10/site-packages/scepter/modules/model/network/autoencoder/ae_kl.py:132: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  sd = torch.load(path, map_location='cpu')
Downloading Model to directory: models/scepter/hub/iic/ACE-0.6B-512px
2024-11-22 17:59:37,108 - modelscope - INFO - Creating symbolic link [models/scepter/hub/iic/ACE-0.6B-512px].
2024-11-22 17:59:37,108 - modelscope - WARNING - Failed to create symbolic link models/scepter/hub/iic/ACE-0.6B-512px.
Loading checkpoint shards: 100%|████████████████████████████████████████████████████████████████████| 5/5 [00:02<00:00,  2.43it/s]
/home/yaosheng/.local/lib/python3.10/site-packages/scepter/modules/model/backbone/ace/ace.py:154: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  model = torch.load(local_path, map_location='cpu')
Restored from ms://iic/ACE-0.6B-512px@models/dit/ace_0.6b_512px.pth with 0 missing and 0 unexpected keys
100%|███████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:00<00:00,  6.28it/s]
Prompt executed in 24.28 seconds


got prompt
parameters is -> {'sample': 'ddim', 'sample_steps': 5, 'guide_scale': 5.5, 'guide_rescale': 0.5, 'discretization': 'trailing', 'target_size_as_tuple': [1024, 1024]}
data is -> [{'prompt': 'Change the style of {image} to colored pencil style', 'negative_prompt': '', 'image': [<PIL.Image.Image image mode=RGB size=1024x1024 at 0x7F74A8F340D0>], 'sample': 'ddim', 'sample_steps': 5, 'guide_scale': 5.5, 'guide_rescale': 0.5, 'discretization': 'trailing', 'target_size_as_tuple': [1024, 1024]}, {'diffusion_model': 'ACE_0.6B_1024_ACE', 'first_stage_model': 'ACE_0.6B_1024_AutoencoderKL', 'cond_stage_model': 'ACE_0.6B_1024_T5EmbedderHF', 'seed': -1}]
Downloading Model to directory: models/scepter/hub/iic/ACE-0.6B-1024px
2024-11-22 18:00:08,710 - modelscope - INFO - Creating symbolic link [models/scepter/hub/iic/ACE-0.6B-1024px]

Other

No response

[ltdrdata]

If you are developing an API-based program, modifications to the random_seed should be handled within your program. In ComfyUI, randomization of the seed is also performed on the front end rather than the backend.