InvokeAI
InvokeAI copied to clipboard
invoke-ai
[bug]: guidance_in.in_layer.weight error
Is there an existing issue for this problem?
- [x] I have searched the existing issues
Operating system
Windows
GPU vendor
Nvidia (CUDA)
GPU model
rtx 3090
GPU VRAM
24gb
Version number
6.0.2
Browser
edge
Python dependencies
No response
What happened
new invoke install, just installed the flux starter models. using flux.1 schnell quantized. text to image. keep getting errors about guidance_in.in_layer.weight
[2025-07-13 14:20:09,648]::[InvokeAI]::INFO --> Executing queue item 7, session 2a7fdd98-6005-4890-8a56-ee100e87c6b3
[2025-07-13 14:20:12,911]::[ModelManagerService]::INFO --> [MODEL CACHE] Loaded model '821a8568-9076-45e8-9a8f-8358fc54b316:text_encoder_2' (T5EncoderModel) onto cuda device in 3.00s. Total model size: 4667.39MB, VRAM: 4667.39MB (100.0%)
You set add_prefix_space. The tokenizer needs to be converted from the slow tokenizers
[2025-07-13 14:20:13,103]::[ModelManagerService]::INFO --> [MODEL CACHE] Loaded model '821a8568-9076-45e8-9a8f-8358fc54b316:tokenizer_2' (T5TokenizerFast) onto cuda device in 0.00s. Total model size: 0.03MB, VRAM: 0.00MB (0.0%)
C:\ai\InvokeAI.venv\Lib\site-packages\bitsandbytes\autograd_functions.py:185: UserWarning: MatMul8bitLt: inputs will be cast from torch.bfloat16 to float16 during quantization
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization")
[2025-07-13 14:20:14,232]::[ModelManagerService]::INFO --> [MODEL CACHE] Loaded model '6413f756-567c-4866-b2dc-26aac4c6a927:text_encoder' (CLIPTextModel) onto cuda device in 0.07s. Total model size: 469.44MB, VRAM: 469.44MB (100.0%)
[2025-07-13 14:20:14,289]::[ModelManagerService]::INFO --> [MODEL CACHE] Loaded model '6413f756-567c-4866-b2dc-26aac4c6a927:tokenizer' (CLIPTokenizer) onto cuda device in 0.00s. Total model size: 0.00MB, VRAM: 0.00MB (0.0%)
[2025-07-13 14:20:14,541]::[InvokeAI]::ERROR --> Error while invoking session 2a7fdd98-6005-4890-8a56-ee100e87c6b3, invocation bca007b5-3c21-4573-8eb5-573af6bd6279 (flux_denoise): Error(s) in loading state_dict for Flux:
Unexpected key(s) in state_dict: "guidance_in.in_layer.bias", "guidance_in.in_layer.weight", "guidance_in.in_layer.weight.absmax", "guidance_in.in_layer.weight.quant_map", "guidance_in.in_layer.weight.quant_state.bitsandbytes__nf4", "guidance_in.out_layer.bias", "guidance_in.out_layer.weight", "guidance_in.out_layer.weight.absmax", "guidance_in.out_layer.weight.quant_map", "guidance_in.out_layer.weight.quant_state.bitsandbytes__nf4".
[2025-07-13 14:20:14,541]::[InvokeAI]::ERROR --> Traceback (most recent call last):
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\services\session_processor\session_processor_default.py", line 130, in run_node
output = invocation.invoke_internal(context=context, services=self._services)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\invocations\baseinvocation.py", line 241, in invoke_internal
output = self.invoke(context)
^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\torch\utils_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\invocations\flux_denoise.py", line 164, in invoke
latents = self._run_diffusion(context)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\invocations\flux_denoise.py", line 345, in _run_diffusion
context.models.load(self.transformer.transformer).model_on_device()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\services\shared\invocation_context.py", line 394, in load
return self._services.model_manager.load.load_model(model, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\services\model_load\model_load_default.py", line 71, in load_model
).load_model(model_config, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\load_default.py", line 56, in load_model
cache_record = self._load_and_cache(model_config, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\load_default.py", line 77, in _load_and_cache
loaded_model = self._load_model(config, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\model_loaders\flux.py", line 306, in _load_model
return self._load_from_singlefile(config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\model_loaders\flux.py", line 330, in _load_from_singlefile
model.load_state_dict(sd, assign=True)
File "C:\ai\InvokeAI.venv\Lib\site-packages\torch\nn\modules\module.py", line 2593, in load_state_dict
raise RuntimeError(
RuntimeError: Error(s) in loading state_dict for Flux:
Unexpected key(s) in state_dict: "guidance_in.in_layer.bias", "guidance_in.in_layer.weight", "guidance_in.in_layer.weight.absmax", "guidance_in.in_layer.weight.quant_map", "guidance_in.in_layer.weight.quant_state.bitsandbytes__nf4", "guidance_in.out_layer.bias", "guidance_in.out_layer.weight", "guidance_in.out_layer.weight.absmax", "guidance_in.out_layer.weight.quant_map", "guidance_in.out_layer.weight.quant_state.bitsandbytes__nf4".
[2025-07-13 14:20:14,555]::[InvokeAI]::INFO --> Graph stats: 2a7fdd98-6005-4890-8a56-ee100e87c6b3 Node Calls Seconds VRAM Used flux_model_loader 1 0.014s 0.000G string 1 0.000s 0.000G flux_text_encoder 1 4.735s 5.035G collect 1 0.002s 5.031G integer 1 0.001s 5.031G flux_denoise 1 0.128s 5.036G TOTAL GRAPH EXECUTION TIME: 4.880s TOTAL GRAPH WALL TIME: 4.883s RAM used by InvokeAI process: 6.23G (+5.892G) RAM used to load models: 5.02G VRAM in use: 5.031G RAM cache statistics: Model cache hits: 4 Model cache misses: 5 Models cached: 4 Models cleared from cache: 0 Cache high water mark: 5.02/0.00G
C:\ai\InvokeAI.venv\Lib\site-packages\huggingface_hub\utils_deprecation.py:131: FutureWarning: 'get_token_permission' (from 'huggingface_hub.hf_api') is deprecated and will be removed from version '1.0'. Permissions are more complex than when get_token_permission was first introduced. OAuth and fine-grain tokens allows for more detailed permissions. If you need to know the permissions associated with a token, please use whoami and check the 'auth' key.
warnings.warn(warning_message, FutureWarning)
[2025-07-13 14:25:43,866]::[InvokeAI]::INFO --> Executing queue item 8, session 85f7e209-d291-4910-a488-9b833c5d673f
[2025-07-13 14:25:43,957]::[InvokeAI]::ERROR --> Error while invoking session 85f7e209-d291-4910-a488-9b833c5d673f, invocation 0f39b916-21f0-4b08-82bf-40c45c49aece (flux_denoise): Error(s) in loading state_dict for Flux:
Unexpected key(s) in state_dict: "guidance_in.in_layer.bias", "guidance_in.in_layer.weight", "guidance_in.in_layer.weight.absmax", "guidance_in.in_layer.weight.quant_map", "guidance_in.in_layer.weight.quant_state.bitsandbytes__nf4", "guidance_in.out_layer.bias", "guidance_in.out_layer.weight", "guidance_in.out_layer.weight.absmax", "guidance_in.out_layer.weight.quant_map", "guidance_in.out_layer.weight.quant_state.bitsandbytes__nf4".
[2025-07-13 14:25:43,957]::[InvokeAI]::ERROR --> Traceback (most recent call last):
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\services\session_processor\session_processor_default.py", line 130, in run_node
output = invocation.invoke_internal(context=context, services=self._services)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\invocations\baseinvocation.py", line 241, in invoke_internal
output = self.invoke(context)
^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\torch\utils_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\invocations\flux_denoise.py", line 164, in invoke
latents = self._run_diffusion(context)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\invocations\flux_denoise.py", line 345, in _run_diffusion
context.models.load(self.transformer.transformer).model_on_device()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\services\shared\invocation_context.py", line 394, in load
return self._services.model_manager.load.load_model(model, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\app\services\model_load\model_load_default.py", line 71, in load_model
).load_model(model_config, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\load_default.py", line 56, in load_model
cache_record = self._load_and_cache(model_config, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\load_default.py", line 77, in _load_and_cache
loaded_model = self._load_model(config, submodel_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\model_loaders\flux.py", line 306, in _load_model
return self._load_from_singlefile(config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\ai\InvokeAI.venv\Lib\site-packages\invokeai\backend\model_manager\load\model_loaders\flux.py", line 330, in _load_from_singlefile
model.load_state_dict(sd, assign=True)
File "C:\ai\InvokeAI.venv\Lib\site-packages\torch\nn\modules\module.py", line 2593, in load_state_dict
raise RuntimeError(
RuntimeError: Error(s) in loading state_dict for Flux:
Unexpected key(s) in state_dict: "guidance_in.in_layer.bias", "guidance_in.in_layer.weight", "guidance_in.in_layer.weight.absmax", "guidance_in.in_layer.weight.quant_map", "guidance_in.in_layer.weight.quant_state.bitsandbytes__nf4", "guidance_in.out_layer.bias", "guidance_in.out_layer.weight", "guidance_in.out_layer.weight.absmax", "guidance_in.out_layer.weight.quant_map", "guidance_in.out_layer.weight.quant_state.bitsandbytes__nf4".
[2025-07-13 14:25:43,967]::[InvokeAI]::INFO --> Graph stats: 85f7e209-d291-4910-a488-9b833c5d673f Node Calls Seconds VRAM Used flux_model_loader 1 0.000s 5.031G string 1 0.001s 5.031G flux_text_encoder 1 0.000s 5.031G collect 1 0.001s 5.031G integer 1 0.000s 5.031G flux_denoise 1 0.086s 5.036G TOTAL GRAPH EXECUTION TIME: 0.088s TOTAL GRAPH WALL TIME: 0.088s RAM used by InvokeAI process: 6.23G (+0.005G) RAM used to load models: 0.00G VRAM in use: 5.031G RAM cache statistics: Model cache hits: 0 Model cache misses: 1 Models cached: 0 Models cleared from cache: 0 Cache high water mark: 0.00/0.00G
What you expected to happen
invoke images
How to reproduce the problem
No response
Additional context
No response
Discord username
No response
This is a serious blocker. Basically, the default "starter pack" model just fails to work completely.
Ran into this issue as well. I pivoted to use the FLUX.1 dev model (non-quantized: ~24GB) and that seemed to do the trick for me.
