How to convert YOLO NAS .pth output into Tflite.

[adkbbx]

💡 Your Question

I was trying to convert the yolo nas output of my custom object detection model from .pth into .tflite by refering to this repo by converting it into ONNX -> TF -> Tflite.

!pip install super-gradients==3.1.0
!pip install imutils
!pip install pytube --upgrade
!pip install --upgrade pillow==9.2.0

import torch
from super_gradients.training import models

dataset_params = {
    'data_dir':'/content/drive/MyDrive/Numberplate/Yolo/Yolov8/data',
    'train_images_dir':'train/images',
    'train_labels_dir':'train/labels',
    'val_images_dir':'validate/images',
    'val_labels_dir':'validate/labels',
    'test_images_dir':'test/images',
    'test_labels_dir':'test/labels',
    'classes': ['blue','red','green','black','yellow','white']
}


model = models.get('yolo_nas_l', 
                   num_classes=len(dataset_params['classes']), 
                   pretrained_weights="coco"
                   ) 


pt_model_path = "/path/to/ckpt_best.pth" 
model.load_state_dict(torch.load(pt_model_path, map_location='cpu'))

but there seems to be a runtime error when I try to load the model :

RuntimeError: Error(s) in loading state_dict for YoloNAS_L:
	Missing key(s) in state_dict: "backbone.stem.conv.branch_3x3.conv.weight", "backbone.stem.conv.branch_3x3.bn.weight", "backbone.stem.conv.branch_3x3.bn.bias", "backbone.stem.conv.branch_3x3.bn.running_mean", "backbone.stem.conv.branch_3x3.bn.running_var", "backbone.stem.conv.branch_1x1.weight", "backbone.stem.conv.branch_1x1.bias", "backbone.stem.conv.post_bn.weight", "backbone.stem.conv.post_bn.bias", "backbone.stem.conv.post_bn.running_mean", "backbone.stem.conv.post_bn.running_var", "backbone.stem.conv.rbr_reparam.weight", "backbone.stem.conv.rbr_reparam.bias", "backbone.stage1.downsample.branch_3x3.conv.weight", "backbone.stage1.downsample.branch_3x3.bn.weight", "backbone.stage1.downsample.branch_3x3.bn.bias", "backbone.stage1.downsample.branch_3x3.bn.running_mean", "backbone.stage1.downsample.branch_3x3.bn.running_var", "backbone.stage1.downsample.branch_1x1.weight", "backbone.stage1.downsample.branch_1x1.bias", "backbone.stage1.downsample.post_bn.weight", "backbone.stage1.downsample.post_bn.bias", "backbone.stage1.downsample.post_bn.running_mean", "backbone.stage1.downsample.post_bn.running_var", "backbone.stage1.downsample.rbr_reparam.weight", "backbone.stage1.downsample.rbr_reparam.bias", "backbone.stage1.blocks.conv1.conv.weight", "backbone.stage1.blocks.conv1.bn.weight", "backbone.stage1.blocks.conv1.bn.bias", "backbone.stage1.blocks.conv1.bn.running_mean", "backbone.stage1.blocks.conv1.bn.running_var", "backbone.stage1.blocks.conv2.conv.weight", "backbone.stage...
	Unexpected key(s) in state_dict: "net", "acc", "epoch", "optimizer_state_dict", "scaler_state_dict", "ema_net", "processing_params". 

I am using google colab to run this code. Do let me know if you have any solution this issue.

Thank you.

Versions

No response

[NatanBagrov]

The weights in the pth include other "states". The model's weights are located under ema_net or net

[adkbbx]

The weights in the pth include other "states". The model's weights are located under ema_net or net

Thank you for you response.

I tried to use the net key to get the model weights as below.

my_model = (torch.load(pt_model_path, map_location='cpu'))
model.load_state_dict(my_model['net'])

but it agian gives a similar runtime error as before

RuntimeError                              Traceback (most recent call last)
[<ipython-input-34-68a3d4abfa3b>](https://localhost:8080/#) in <cell line: 1>()
----> 1 model.load_state_dict(my_model['ema_net'])

[/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py](https://localhost:8080/#) in load_state_dict(self, state_dict, strict)
   1669 
   1670         if len(error_msgs) > 0:
-> 1671             raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
   1672                                self.__class__.__name__, "\n\t".join(error_msgs)))
   1673         return _IncompatibleKeys(missing_keys, unexpected_keys)

RuntimeError: Error(s) in loading state_dict for YoloNAS_L:
	Missing key(s) in state_dict: "backbone.stem.conv.branch_3x3.conv.weight", "backbone.stem.conv.branch_3x3.bn.weight", "backbone.stem.conv.branch_3x3.bn.bias", "backbone.stem.conv.branch_3x3.bn.running_mean", "backbone.stem.conv.branch_3x3.bn.running_var", "backbone.stem.conv.branch_1x1.weight", "backbone.stem.conv.branch_1x1.bias", "backbone.stem.conv.post_bn.weight", "backbone.stem.conv.post_bn.bias", "backbone.stem.conv.post_bn.running_mean", "backbone.stem.conv.post_bn.running_var", "backbone.stem.conv.rbr_reparam.weight", "backbone.stem.conv.rbr_reparam.bias", "backbone.stage1.downsample.branch_3x3.conv.weight", "backbone.stage1.downsample.branch_3x3.bn.weight", "backbone.stage1.downsample.branch_3x3.bn.bias", "backbone.stage1.downsample.branch_3x3.bn.running_mean", "backbone.stage1.downsample.branch_3x3.bn.running_var", "backbone.stage1.downsample.branch_1x1.weight", "backbone.stage1.downsample.branch_1x1.bias", "backbone.stage1.downsample.post_bn.weight", "backbone.stage1.downsample.post_bn.bias", "backbone.stage1.downsample.post_bn.running_mean", "backbone.stage1.downsample.post_bn.running_var", "backbone.stage1.downsample.rbr_reparam.weight", "backbone.stage1.downsample.rbr_reparam.bias", "backbone.stage1.blocks.conv1.conv.weight", "backbone.stage1.blocks.conv1.bn.weight", "backbone.stage1.blocks.conv1.bn.bias", "backbone.stage1.blocks.conv1.bn.running_mean", "backbone.stage1.blocks.conv1.bn.running_var", "backbone.stage1.blocks.conv2.conv.weight", "backbone.stage...
	Unexpected key(s) in state_dict: "module.backbone.stem.conv.branch_3x3.conv.weight", "module.backbone.stem.conv.branch_3x3.bn.weight", "module.backbone.stem.conv.branch_3x3.bn.bias", "module.backbone.stem.conv.branch_3x3.bn.running_mean", "module.backbone.stem.conv.branch_3x3.bn.running_var", "module.backbone.stem.conv.branch_3x3.bn.num_batches_tracked", "module.backbone.stem.conv.branch_1x1.weight", "module.backbone.stem.conv.branch_1x1.bias", "module.backbone.stem.conv.post_bn.weight", "module.backbone.stem.conv.post_bn.bias", "module.backbone.stem.conv.post_bn.running_mean", "module.backbone.stem.conv.post_bn.running_var", "module.backbone.stem.conv.post_bn.num_batches_tracked", "module.backbone.stem.conv.rbr_reparam.weight", "module.backbone.stem.conv.rbr_reparam.bias", "module.backbone.stage1.downsample.branch_3x3.conv.weight", "module.backbone.stage1.downsample.branch_3x3.bn.weight", "module.backbone.stage1.downsample.branch_3x3.bn.bias", "module.backbone.stage1.downsample.branch_3x3.bn.running_mean", "module.backbone.stage1.downsample.branch_3x3.bn.running_var", "module.backbone.stage1.downsample.branch_3x3.bn.num_batches_tracked", "module.backbone.stage1.downsample.branch_1x1.weight", "module.backbone.stage1.downsample.branch_1x1.bias", "module.backbone.stage1.downsample.post_bn.weight", "module.backbone.stage1.downsample.post_bn.bias", "module.backbone.stage1.downsample.post_bn.running_mean", "module.backbone.stage1.downsample.post_bn.running_var", "module.backbo...

Do let me know if there is any other solution I can try to solve this.

Thank you in advance.

[NatanBagrov]

As you can see, the keys pretty much match, and the only difference is module. prefix. To successfully load state dict you may use the following existing method in SG:

adaptive_load_state_dict(model, my_model, strict="no_key_matching")

[adkbbx]

Thank you for your response.

I did try to run the code above as

from super_gradients.training.utils.checkpoint_utils import adaptive_load_state_dict
adaptive_load_state_dict(model, my_model, strict="no_key_matching")

but I got another runtime error as below:

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/super_gradients/training/utils/checkpoint_utils.py](https://localhost:8080/#) in adaptive_load_state_dict(net, state_dict, strict, solver)
     57         strict_bool = strict if isinstance(strict, bool) else strict != StrictLoad.OFF
---> 58         net.load_state_dict(state_dict, strict=strict_bool)
     59     except (RuntimeError, ValueError, KeyError) as ex:

3 frames
RuntimeError: Error(s) in loading state_dict for YoloNAS_L:
	Missing key(s) in state_dict: "backbone.stem.conv.branch_3x3.conv.weight", "backbone.stem.conv.branch_3x3.bn.weight", "backbone.stem.conv.branch_3x3.bn.bias", "backbone.stem.conv.branch_3x3.bn.running_mean", "backbone.stem.conv.branch_3x3.bn.running_var", "backbone.stem.conv.branch_1x1.weight", "backbone.stem.conv.branch_1x1.bias", "backbone.stem.conv.post_bn.weight", "backbone.stem.conv.post_bn.bias", "backbone.stem.conv.post_bn.running_mean", "backbone.stem.conv.post_bn.running_var", "backbone.stem.conv.rbr_reparam.weight", "backbone.stem.conv.rbr_reparam.bias", "backbone.stage1.downsample.branch_3x3.conv.weight", "backbone.stage1.downsample.branch_3x3.bn.weight", "backbone.stage1.downsample.branch_3x3.bn.bias", "backbone.stage1.downsample.branch_3x3.bn.running_mean", "backbone.stage1.downsample.branch_3x3.bn.running_var", "backbone.stage1.downsample.branch_1x1.weight", "backbone.stage1.downsample.branch_1x1.bias", "backbone.stage1.downsample.post_bn.weight", "backbone.stage1.downsample.post_bn.bias", "backbone.stage1.downsample.post_bn.running_mean", "backbone.stage1.downsample.post_bn.running_var", "backbone.stage1.downsample.rbr_reparam.weight", "backbone.stage1.downsample.rbr_reparam.bias", "backbone.stage1.blocks.conv1.conv.weight", "backbone.stage1.blocks.conv1.bn.weight", "backbone.stage1.blocks.conv1.bn.bias", "backbone.stage1.blocks.conv1.bn.running_mean", "backbone.stage1.blocks.conv1.bn.running_var", "backbone.stage1.blocks.conv2.conv.weight", "backbone.stage...
	Unexpected key(s) in state_dict: "module.backbone.stem.conv.branch_3x3.conv.weight", "module.backbone.stem.conv.branch_3x3.bn.weight", "module.backbone.stem.conv.branch_3x3.bn.bias", "module.backbone.stem.conv.branch_3x3.bn.running_mean", "module.backbone.stem.conv.branch_3x3.bn.running_var", "module.backbone.stem.conv.branch_3x3.bn.num_batches_tracked", "module.backbone.stem.conv.branch_1x1.weight", "module.backbone.stem.conv.branch_1x1.bias", "module.backbone.stem.conv.post_bn.weight", "module.backbone.stem.conv.post_bn.bias", "module.backbone.stem.conv.post_bn.running_mean", "module.backbone.stem.conv.post_bn.running_var", "module.backbone.stem.conv.post_bn.num_batches_tracked", "module.backbone.stem.conv.rbr_reparam.weight", "module.backbone.stem.conv.rbr_reparam.bias", "module.backbone.stage1.downsample.branch_3x3.conv.weight", "module.backbone.stage1.downsample.branch_3x3.bn.weight", "module.backbone.stage1.downsample.branch_3x3.bn.bias", "module.backbone.stage1.downsample.branch_3x3.bn.running_mean", "module.backbone.stage1.downsample.branch_3x3.bn.running_var", "module.backbone.stage1.downsample.branch_3x3.bn.num_batches_tracked", "module.backbone.stage1.downsample.branch_1x1.weight", "module.backbone.stage1.downsample.branch_1x1.bias", "module.backbone.stage1.downsample.post_bn.weight", "module.backbone.stage1.downsample.post_bn.bias", "module.backbone.stage1.downsample.post_bn.running_mean", "module.backbone.stage1.downsample.post_bn.running_var", "module.backbo...

During handling of the above exception, another exception occurred:

RuntimeError                              Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/super_gradients/training/utils/checkpoint_utils.py](https://localhost:8080/#) in raise_informative_runtime_error(state_dict, checkpoint, exception_msg)
    178         exception_msg = f"\n{'=' * 200} \nThe checkpoint and model shapes do no fit, e.g.: {ex}\n{'=' * 200}"
    179     finally:
--> 180         raise RuntimeError(exception_msg)
    181 
    182 

RuntimeError: 
======================================================================================================================================================================================================== 
The checkpoint and model shapes do no fit, e.g.: ckpt layer module.heads.head1.cls_pred.weight with shape torch.Size([3, 128, 1, 1]) does not match heads.head1.cls_pred.weight with shape torch.Size([6, 128, 1, 1]) in the model
========================================================================================================================================================================================================

[NatanBagrov]

This looks like an issue with different number of classes.
Please follow the instruction of this issue #949, and let me know if that helps

[adkbbx]

Thank you for your response.

I tried running the code as in #949

It shows me a runtime error similar to above when I used:
model = models.get('yolo_nas_l', num_classes=6, pretrained_weights="coco")
error:

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/super_gradients/training/utils/checkpoint_utils.py](https://localhost:8080/#) in adaptive_load_state_dict(net, state_dict, strict, solver)
     57         strict_bool = strict if isinstance(strict, bool) else strict != StrictLoad.OFF
---> 58         net.load_state_dict(state_dict, strict=strict_bool)
     59     except (RuntimeError, ValueError, KeyError) as ex:

3 frames
RuntimeError: Error(s) in loading state_dict for YoloNAS_L:
	Missing key(s) in state_dict: "backbone.stem.conv.branch_3x3.conv.weight", "backbone.stem.conv.branch_3x3.bn.weight", "backbone.stem.conv.branch_3x3.bn.bias", "backbone.stem.conv.branch_3x3.bn.running_mean", "backbone.stem.conv.branch_3x3.bn.running_var", "backbone.stem.conv.branch_1x1.weight", "backbone.stem.conv.branch_1x1.bias", "backbone.stem.conv.post_bn.weight", "backbone.stem.conv.post_bn.bias", "backbone.stem.conv.post_bn.running_mean", "backbone.stem.conv.post_bn.running_var", "backbone.stem.conv.rbr_reparam.weight", "backbone.stem.conv.rbr_reparam.bias", "backbone.stage1.downsample.branch_3x3.conv.weight", "backbone.stage1.downsample.branch_3x3.bn.weight", "backbone.stage1.downsample.branch_3x3.bn.bias", "backbone.stage1.downsample.branch_3x3.bn.running_mean", "backbone.stage1.downsample.branch_3x3.bn.running_var", "backbone.stage1.downsample.branch_1x1.weight", "backbone.stage1.downsample.branch_1x1.bias", "backbone.stage1.downsample.post_bn.weight", "backbone.stage1.downsample.post_bn.bias", "backbone.stage1.downsample.post_bn.running_mean", "backbone.stage1.downsample.post_bn.running_var", "backbone.stage1.downsample.rbr_reparam.weight", "backbone.stage1.downsample.rbr_reparam.bias", "backbone.stage1.blocks.conv1.conv.weight", "backbone.stage1.blocks.conv1.bn.weight", "backbone.stage1.blocks.conv1.bn.bias", "backbone.stage1.blocks.conv1.bn.running_mean", "backbone.stage1.blocks.conv1.bn.running_var", "backbone.stage1.blocks.conv2.conv.weight", "backbone.stage...
	Unexpected key(s) in state_dict: "module.backbone.stem.conv.branch_3x3.conv.weight", "module.backbone.stem.conv.branch_3x3.bn.weight", "module.backbone.stem.conv.branch_3x3.bn.bias", "module.backbone.stem.conv.branch_3x3.bn.running_mean", "module.backbone.stem.conv.branch_3x3.bn.running_var", "module.backbone.stem.conv.branch_3x3.bn.num_batches_tracked", "module.backbone.stem.conv.branch_1x1.weight", "module.backbone.stem.conv.branch_1x1.bias", "module.backbone.stem.conv.post_bn.weight", "module.backbone.stem.conv.post_bn.bias", "module.backbone.stem.conv.post_bn.running_mean", "module.backbone.stem.conv.post_bn.running_var", "module.backbone.stem.conv.post_bn.num_batches_tracked", "module.backbone.stem.conv.rbr_reparam.weight", "module.backbone.stem.conv.rbr_reparam.bias", "module.backbone.stage1.downsample.branch_3x3.conv.weight", "module.backbone.stage1.downsample.branch_3x3.bn.weight", "module.backbone.stage1.downsample.branch_3x3.bn.bias", "module.backbone.stage1.downsample.branch_3x3.bn.running_mean", "module.backbone.stage1.downsample.branch_3x3.bn.running_var", "module.backbone.stage1.downsample.branch_3x3.bn.num_batches_tracked", "module.backbone.stage1.downsample.branch_1x1.weight", "module.backbone.stage1.downsample.branch_1x1.bias", "module.backbone.stage1.downsample.post_bn.weight", "module.backbone.stage1.downsample.post_bn.bias", "module.backbone.stage1.downsample.post_bn.running_mean", "module.backbone.stage1.downsample.post_bn.running_var", "module.backbo...

During handling of the above exception, another exception occurred:

RuntimeError                              Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/super_gradients/training/utils/checkpoint_utils.py](https://localhost:8080/#) in raise_informative_runtime_error(state_dict, checkpoint, exception_msg)
    178         exception_msg = f"\n{'=' * 200} \nThe checkpoint and model shapes do no fit, e.g.: {ex}\n{'=' * 200}"
    179     finally:
--> 180         raise RuntimeError(exception_msg)
    181 
    182 

RuntimeError: 
======================================================================================================================================================================================================== 
The checkpoint and model shapes do no fit, e.g.: ckpt layer module.heads.head1.cls_pred.weight with shape torch.Size([3, 128, 1, 1]) does not match heads.head1.cls_pred.weight with shape torch.Size([6, 128, 1, 1]) in the model
========================================================================================================================================================================================================

but it seems to run and showed me another runtime error when I changed the number of classes to 3 as below:

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/super_gradients/training/utils/checkpoint_utils.py](https://localhost:8080/#) in adaptive_load_state_dict(net, state_dict, strict, solver)
     57         strict_bool = strict if isinstance(strict, bool) else strict != StrictLoad.OFF
---> 58         net.load_state_dict(state_dict, strict=strict_bool)
     59     except (RuntimeError, ValueError, KeyError) as ex:

3 frames
RuntimeError: Error(s) in loading state_dict for YoloNAS_L:
	Missing key(s) in state_dict: "backbone.stem.conv.branch_3x3.conv.weight", "backbone.stem.conv.branch_3x3.bn.weight", "backbone.stem.conv.branch_3x3.bn.bias", "backbone.stem.conv.branch_3x3.bn.running_mean", "backbone.stem.conv.branch_3x3.bn.running_var", "backbone.stem.conv.branch_1x1.weight", "backbone.stem.conv.branch_1x1.bias", "backbone.stem.conv.post_bn.weight", "backbone.stem.conv.post_bn.bias", "backbone.stem.conv.post_bn.running_mean", "backbone.stem.conv.post_bn.running_var", "backbone.stem.conv.rbr_reparam.weight", "backbone.stem.conv.rbr_reparam.bias", "backbone.stage1.downsample.branch_3x3.conv.weight", "backbone.stage1.downsample.branch_3x3.bn.weight", "backbone.stage1.downsample.branch_3x3.bn.bias", "backbone.stage1.downsample.branch_3x3.bn.running_mean", "backbone.stage1.downsample.branch_3x3.bn.running_var", "backbone.stage1.downsample.branch_1x1.weight", "backbone.stage1.downsample.branch_1x1.bias", "backbone.stage1.downsample.post_bn.weight", "backbone.stage1.downsample.post_bn.bias", "backbone.stage1.downsample.post_bn.running_mean", "backbone.stage1.downsample.post_bn.running_var", "backbone.stage1.downsample.rbr_reparam.weight", "backbone.stage1.downsample.rbr_reparam.bias", "backbone.stage1.blocks.conv1.conv.weight", "backbone.stage1.blocks.conv1.bn.weight", "backbone.stage1.blocks.conv1.bn.bias", "backbone.stage1.blocks.conv1.bn.running_mean", "backbone.stage1.blocks.conv1.bn.running_var", "backbone.stage1.blocks.conv2.conv.weight", "backbone.stage...
	Unexpected key(s) in state_dict: "module.backbone.stem.conv.branch_3x3.conv.weight", "module.backbone.stem.conv.branch_3x3.bn.weight", "module.backbone.stem.conv.branch_3x3.bn.bias", "module.backbone.stem.conv.branch_3x3.bn.running_mean", "module.backbone.stem.conv.branch_3x3.bn.running_var", "module.backbone.stem.conv.branch_3x3.bn.num_batches_tracked", "module.backbone.stem.conv.branch_1x1.weight", "module.backbone.stem.conv.branch_1x1.bias", "module.backbone.stem.conv.post_bn.weight", "module.backbone.stem.conv.post_bn.bias", "module.backbone.stem.conv.post_bn.running_mean", "module.backbone.stem.conv.post_bn.running_var", "module.backbone.stem.conv.post_bn.num_batches_tracked", "module.backbone.stem.conv.rbr_reparam.weight", "module.backbone.stem.conv.rbr_reparam.bias", "module.backbone.stage1.downsample.branch_3x3.conv.weight", "module.backbone.stage1.downsample.branch_3x3.bn.weight", "module.backbone.stage1.downsample.branch_3x3.bn.bias", "module.backbone.stage1.downsample.branch_3x3.bn.running_mean", "module.backbone.stage1.downsample.branch_3x3.bn.running_var", "module.backbone.stage1.downsample.branch_3x3.bn.num_batches_tracked", "module.backbone.stage1.downsample.branch_1x1.weight", "module.backbone.stage1.downsample.branch_1x1.bias", "module.backbone.stage1.downsample.post_bn.weight", "module.backbone.stage1.downsample.post_bn.bias", "module.backbone.stage1.downsample.post_bn.running_mean", "module.backbone.stage1.downsample.post_bn.running_var", "module.backbo...

During handling of the above exception, another exception occurred:

RuntimeError                              Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/super_gradients/training/utils/checkpoint_utils.py](https://localhost:8080/#) in raise_informative_runtime_error(state_dict, checkpoint, exception_msg)
    178         exception_msg = f"\n{'=' * 200} \nThe checkpoint and model shapes do no fit, e.g.: {ex}\n{'=' * 200}"
    179     finally:
--> 180         raise RuntimeError(exception_msg)
    181 
    182 

RuntimeError: 
========================================================================================================================================================================================================
Error(s) in loading state_dict for YoloNAS_L:
	Missing key(s) in state_dict: "backbone.stem.conv.branch_3x3.conv.weight", "backbone.stem.conv.branch_3x3.bn.weight", "backbone.stem.conv.branch_3x3.bn.bias", "backbone.stem.conv.branch_3x3.bn.running_mean", "backbone.stem.conv.branch_3x3.bn.running_var", "backbone.stem.conv.branch_1x1.weight", "backbone.stem.conv.branch_1x1.bias", "backbone.stem.conv.post_bn.weight", "backbone.stem.conv.post_bn.bias", "backbone.stem.conv.post_bn.running_mean", "backbone.stem.conv.post_bn.running_var", "backbone.stem.conv.rbr_reparam.weight", "backbone.stem.conv.rbr_reparam.bias", "backbone.stage1.downsample.branch_3x3.conv.weight", "backbone.stage1.downsample.branch_3x3.bn.weight", "backbone.stage1.downsample.branch_3x3.bn.bias", "backbone.stage1.downsample.branch_3x3.bn.running_mean", "backbone.stage1.downsample.branch_3x3.bn.running_var", "backbone.stage1.downsample.branch_1x1.weight", "backbone.stage1.downsample.branch_1x1.bias", "backbone.stage1.downsample.post_bn.weight", "backbone.stage1.downsample.post_bn.bias", "backbone.stage1.downsample.post_bn.running_mean", "backbone.stage1.downsample.post_bn.running_var", "backbone.stage1.downsample.rbr_reparam.weight", "backbone.stage1.downsample.rbr_reparam.bias", "backbone.stage1.blocks.conv1.conv.weight", "backbone.stage1.blocks.conv1.bn.weight", "backbone.stage1.blocks.conv1.bn.bias", "backbone.stage1.blocks.conv1.bn.running_mean", "backbone.stage1.blocks.conv1.bn.running_var", "backbone.stage1.blocks.conv2.conv.weight", "backbone.stage...
	Unexpected key(s) in state_dict: "module.backbone.stem.conv.branch_3x3.conv.weight", "module.backbone.stem.conv.branch_3x3.bn.weight", "module.backbone.stem.conv.branch_3x3.bn.bias", "module.backbone.stem.conv.branch_3x3.bn.running_mean", "module.backbone.stem.conv.branch_3x3.bn.running_var", "module.backbone.stem.conv.branch_3x3.bn.num_batches_tracked", "module.backbone.stem.conv.branch_1x1.weight", "module.backbone.stem.conv.branch_1x1.bias", "module.backbone.stem.conv.post_bn.weight", "module.backbone.stem.conv.post_bn.bias", "module.backbone.stem.conv.post_bn.running_mean", "module.backbone.stem.conv.post_bn.running_var", "module.backbone.stem.conv.post_bn.num_batches_tracked", "module.backbone.stem.conv.rbr_reparam.weight", "module.backbone.stem.conv.rbr_reparam.bias", "module.backbone.stage1.downsample.branch_3x3.conv.weight", "module.backbone.stage1.downsample.branch_3x3.bn.weight", "module.backbone.stage1.downsample.branch_3x3.bn.bias", "module.backbone.stage1.downsample.branch_3x3.bn.running_mean", "module.backbone.stage1.downsample.branch_3x3.bn.running_var", "module.backbone.stage1.downsample.branch_3x3.bn.num_batches_tracked", "module.backbone.stage1.downsample.branch_1x1.weight", "module.backbone.stage1.downsample.branch_1x1.bias", "module.backbone.stage1.downsample.post_bn.weight", "module.backbone.stage1.downsample.post_bn.bias", "module.backbone.stage1.downsample.post_bn.running_mean", "module.backbone.stage1.downsample.post_bn.running_var", "module.backbo...
convert ckpt via the utils.adapt_state_dict_to_fit_model_layer_names method
a converted checkpoint file was saved in the path /tmp/tmpp97s1mos.pt
========================================================================================================================================================================================================

Does this mean the model assumes that I only have 3 classes when I had initialised 6 classes during trining the model?

[BloodAxe]

After we merged #1184 this should be resolved