Incorrect path_or_model_id: '/data/apps/models/checkpoints/Flux1//text_encoder'. Please provide either the path to a local folder or the repo_id of a model on the Hub.

[adamreading]

I am trying to run StoryDiffusion in a Comfyui in the cloud on MimicPC.

I have followed all of the instructions - I downloaded the kijai version of Flux fp 8 and put the path to it in the repo id - I get the error - Incorrect path_or_model_id: '/data/apps/models/checkpoints/Flux1//text_encoder'. Please provide either the path to a local folder or the repo_id of a model on the Hub.

If I put the repo id - Kijai/flux-fp8 instead I get a different error - like it's looking for a unet diffuser info rather than a checkpoint. What am I doing wrong please?

(If I do not put anything in the Repo Id space - it runs for 30 mins plus (on A10G) and eventually gives another error message - File "/data/app/execution.py", line 323, 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 "/data/app/execution.py", line 198, 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 "/data/app/execution.py", line 169, in _map_node_over_list process_inputs(input_dict, i) File "/data/app/execution.py", line 158, in process_inputs results.append(getattr(obj, func)(**inputs)) File "/data/app/custom_nodes/ComfyUI_StoryDiffusion/Storydiffusion_node.py", line 1790, in story_model_loader pipe = FluxGenerator(flux_pulid_name, ckpt_path, "cuda", offload=offload, File "/data/app/custom_nodes/ComfyUI_StoryDiffusion/PuLID/app_flux.py", line 46, in init self.pulid_model = PuLIDPipeline(self.model, device, self.clip_vision_path,weight_dtype=torch.bfloat16) File "/data/app/custom_nodes/ComfyUI_StoryDiffusion/PuLID/pulid/pipeline_flux.py", line 161, in init self.app = FaceAnalysis( File "/data/miniconda3/pkgs/comfyui/lib/python3.10/site-packages/insightface/app/face_analysis.py", line 43, in init assert 'detection' in self.models AssertionError

Prompt executed in 923.86 seconds

message (1).txt

[smthemex]

"assert 'detection' in self.models" is a insightface model load error i will fix it soon， if using repo repo should be like this: /aaa/Flux1 .Do not directly point the path to text_encoder, but to the previous level path of /aaa/Flux1

[adamreading]

"assert 'detection' in self.models" is a insightface model load error i will fix it soon， if using repo repo should be like this: /aaa/Flux1 .Do not directly point the path to text_encoder, but to the previous level path of /aaa/Flux1

Thank you - I have the Kijai/flux-fp8 safetensor downloaded to my checkpoints/Flux1/ folder - that is the path I am putting in the Repo ID But with the checkpoints there are no text_encoder files or directories to load - just a single safetensor file?

If I use Kijai/flux-fp8 as a repo id it just says something else is missing

[adamreading]

[adamreading]

and if I use Kijai/flux-fp8

[smthemex]

"Kijai/flux-fp8" Just need its monolithic model and place it in the comfy/models/checkpoints directory ,if using pulid don't fill repo_id

[adamreading]

"Kijai/flux-fp8" Just need its monolithic model and place it in the comfy/models/checkpoints directory ,if using pulid don't fill repo_id

Ok, so this is what I did right in the beginning - and it took 16 minutes on a A10G 24gbVram 32Gb Ram cloud GPU then crashed with the assert error - so I guess at the moment, there is no way for me to make it work?

[smthemex]

it need a big Ram(>64G) to load T5, int fp8, Vram 24G is enough. The current version has a bug in CUDA uninstallation, and I am currently fixing it. It will take about half an hour

[adamreading]

Thank you - at least in cloud i can access big Ram - though it's expensive

[adamreading]

I know you are now away - but it crashes every time now in the sampler - saying invalid key [Taylor] or whatever name you choose. If you don't put a name it give a List error.