Something wrong with fixing ema

[MolchanovYA]

Before submitting a bug report, please be aware that your issue must be reproducible with all of the following, otherwise it is non-actionable, and we can not help you:

• Current repo: run git fetch && git status -uno to check and git pull to update repo
• Common dataset: coco.yaml or coco128.yaml
• Common environment: Colab, Google Cloud, or Docker image. See https://github.com/ultralytics/yolov5#environments

If this is a custom dataset/training question you must include your train*.jpg, test*.jpg and results.png figures, or we can not help you. You can generate these with utils.plot_results().

🐛 Bug

A clear and concise description of what the bug is.

I get an error while trying to resume a training

To Reproduce (REQUIRED)

Input:
``
!python /content/yolov5/train.py --resume /content/drive/MyDrive/XRay/202102283/weights/last.pt

Output:
wandb: Run `wandb offline` to turn off syncing.

Traceback (most recent call last):
  File "train.py", line 532, in <module>
    train(hyp, opt, device, tb_writer, wandb)
  File "train.py", line 154, in train
    ema.ema.load_state_dict(ckpt['ema'].float().state_dict())
AttributeError: 'tuple' object has no attribute 'float'

wandb: Waiting for W&B process to finish, PID 8842

## Expected behavior
Resuming of training


## Environment
Google Colab

 - OS: [e.g. Ubuntu]
 - GPU [e.g. 2080 Ti]


## Additional context
Removing float() fixes the bug.

[glenn-jocher]

@MolchanovYA sorry, there have been recent PRs to improve EMA resume behavior. Everything should work correctly now when starting and resuming a new run, however if you started and resumed your run with different versions of the code, this will likely lead to problems.

I would suggest to restart training from a new clone.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.