[Bug] The dype error in Dreambooth.

[wangqiang9]

Prerequisite

• I have searched Issues and Discussions but cannot get the expected help.
• I have read the FAQ documentation but cannot get the expected help.
• The bug has not been fixed in the latest version (main) or latest version (0.x).

Task

I'm using the official example scripts/configs for the officially supported tasks/models/datasets.

Branch

main branch https://github.com/open-mmlab/mmagic

Environment

When mmaigc was first released, I reproduced the dreambooth and did not encounter this problem... But today, when I cloned the latest code experiment, I encountered a dtype-related problem. The complete error report is as follows:

/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/diffusers/configuration_utils.py:135: FutureWarning: Accessing config attribute `num_train_timesteps` directly via 'DDPMScheduler' object attribute is deprecated. Please access 'num_train_timesteps' over 'DDPMScheduler's config object instead, e.g. 'scheduler.config.num_train_timesteps'.
  deprecate("direct config name access", "1.0.0", deprecation_message, standard_warn=False)
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /root/modelscope/mmagic/tools/train.py:114 in <module>                                           │
│                                                                                                  │
│   111                                                                                            │
│   112                                                                                            │
│   113 if __name__ == '__main__':                                                                 │
│ ❱ 114 │   main()                                                                                 │
│   115                                                                                            │
│                                                                                                  │
│ /root/modelscope/mmagic/tools/train.py:107 in main                                               │
│                                                                                                  │
│   104 │   print_colored_log(f'Log directory: {runner._log_dir}')                                 │
│   105 │                                                                                          │
│   106 │   # start training                                                                       │
│ ❱ 107 │   runner.train()                                                                         │
│   108 │                                                                                          │
│   109 │   print_colored_log(f'Log saved under {runner._log_dir}')                                │
│   110 │   print_colored_log(f'Checkpoint saved under {cfg.work_dir}')                            │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/runner/runner.py:1721 in train  │
│                                                                                                  │
│   1718 │   │   # This must be called **AFTER** model has been wrapped.                           │
│   1719 │   │   self._maybe_compile('train_step')                                                 │
│   1720 │   │                                                                                     │
│ ❱ 1721 │   │   model = self.train_loop.run()  # type: ignore                                     │
│   1722 │   │   self.call_hook('after_run')                                                       │
│   1723 │   │   return model                                                                      │
│   1724                                                                                           │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/runner/loops.py:278 in run      │
│                                                                                                  │
│   275 │   │   │   self.runner.model.train()                                                      │
│   276 │   │   │                                                                                  │
│   277 │   │   │   data_batch = next(self.dataloader_iterator)                                    │
│ ❱ 278 │   │   │   self.run_iter(data_batch)                                                      │
│   279 │   │   │                                                                                  │
│   280 │   │   │   self._decide_current_val_interval()                                            │
│   281 │   │   │   if (self.runner.val_loop is not None                                           │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/runner/loops.py:302 in run_iter │
│                                                                                                  │
│   299 │   │   # synchronization during gradient accumulation process.                            │
│   300 │   │   # outputs should be a dict of loss.                                                │
│   301 │   │   outputs = self.runner.model.train_step(                                            │
│ ❱ 302 │   │   │   data_batch, optim_wrapper=self.runner.optim_wrapper)                           │
│   303 │   │                                                                                      │
│   304 │   │   self.runner.call_hook(                                                             │
│   305 │   │   │   'after_train_iter',                                                            │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/model/wrappers/seperate_distrib │
│ uted.py:102 in train_step                                                                        │
│                                                                                                  │
│    99 │   │   Returns:                                                                           │
│   100 │   │   │   Dict[str, torch.Tensor]: A dict of tensor for logging.                         │
│   101 │   │   """                                                                                │
│ ❱ 102 │   │   return self.module.train_step(data, optim_wrapper)                                 │
│   103 │                                                                                          │
│   104 │   def val_step(self, data: Union[dict, tuple, list]) -> list:                            │
│   105 │   │   """Gets the prediction of module during validation process.                        │
│                                                                                                  │
│ /root/modelscope/mmagic/mmagic/models/editors/dreambooth/dreambooth.py:293 in train_step         │
│                                                                                                  │
│   290 │   │   │   model_output = self.unet(                                                      │
│   291 │   │   │   │   noisy_latents.float(),                                                     │
│   292 │   │   │   │   timesteps,                                                                 │
│ ❱ 293 │   │   │   │   encoder_hidden_states=encoder_hidden_states.float())                       │
│   294 │   │   │   model_pred = model_output['sample']                                            │
│   295 │   │   │                                                                                  │
│   296 │   │   │   loss_dict = dict()                                                             │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/parallel/distributed.py:1040 in │
│ forward                                                                                          │
│                                                                                                  │
│   1037 │   │   │   │   # Notify joined ranks whether they should sync in backwards pass or not.  │
│   1038 │   │   │   │   self._check_global_requires_backward_grad_sync(is_joined_rank=False)      │
│   1039 │   │   │                                                                                 │
│ ❱ 1040 │   │   │   output = self._run_ddp_forward(*inputs, **kwargs)                             │
│   1041 │   │   │                                                                                 │
│   1042 │   │   │   # sync params according to location (before/after forward) user               │
│   1043 │   │   │   # specified as part of hook, if hook was specified.                           │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/parallel/distributed.py:1000 in │
│ _run_ddp_forward                                                                                 │
│                                                                                                  │
│    997 │   │   │   │   self.use_side_stream_for_tensor_copies                                    │
│    998 │   │   │   )                                                                             │
│    999 │   │   │   with self._inside_ddp_forward():                                              │
│ ❱ 1000 │   │   │   │   return module_to_run(*inputs[0], **kwargs[0])                             │
│   1001 │   │   else:                                                                             │
│   1002 │   │   │   with self._inside_ddp_forward():                                              │
│   1003 │   │   │   │   return module_to_run(*inputs, **kwargs)                                   │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/modelscope/mmagic/mmagic/models/archs/wrapper.py:179 in forward                            │
│                                                                                                  │
│   176 │   │   Returns:                                                                           │
│   177 │   │   │   Any: The output of wrapped module's forward function.                          │
│   178 │   │   """                                                                                │
│ ❱ 179 │   │   return self.model(*args, **kwargs)                                                 │
│   180                                                                                            │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/diffusers/models/unet_2d_condition.py:71 │
│ 8 in forward                                                                                     │
│                                                                                                  │
│   715 │   │   │   encoder_hidden_states = self.encoder_hid_proj(encoder_hidden_states)           │
│   716 │   │                                                                                      │
│   717 │   │   # 2. pre-process                                                                   │
│ ❱ 718 │   │   sample = self.conv_in(sample)                                                      │
│   719 │   │                                                                                      │
│   720 │   │   # 3. down                                                                          │
│   721 │   │   down_block_res_samples = (sample,)                                                 │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/conv.py:463 in forward  │
│                                                                                                  │
│    460 │   │   │   │   │   │   self.padding, self.dilation, self.groups)                         │
│    461 │                                                                                         │
│    462 │   def forward(self, input: Tensor) -> Tensor:                                           │
│ ❱  463 │   │   return self._conv_forward(input, self.weight, self.bias)                          │
│    464                                                                                           │
│    465 class Conv3d(_ConvNd):                                                                    │
│    466 │   __doc__ = r"""Applies a 3D convolution over an input signal composed of several inpu  │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/conv.py:460 in          │
│ _conv_forward                                                                                    │
│                                                                                                  │
│    457 │   │   │   │   │   │   │   weight, bias, self.stride,                                    │
│    458 │   │   │   │   │   │   │   _pair(0), self.dilation, self.groups)                         │
│    459 │   │   return F.conv2d(input, weight, bias, self.stride,                                 │
│ ❱  460 │   │   │   │   │   │   self.padding, self.dilation, self.groups)                         │
│    461 │                                                                                         │
│    462 │   def forward(self, input: Tensor) -> Tensor:                                           │
│    463 │   │   return self._conv_forward(input, self.weight, self.bias)                          │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
RuntimeError: Input type (float) and bias type (c10::Half) should be the same
ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 49302) of binary: /root/anaconda3/envs/mmagic/bin/python
Traceback (most recent call last):
  File "/root/anaconda3/envs/mmagic/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/root/anaconda3/envs/mmagic/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/distributed/launch.py", line 195, in <module>
    main()
  File "/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/distributed/launch.py", line 191, in main
    launch(args)
  File "/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/distributed/launch.py", line 176, in launch
    run(args)
  File "/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/distributed/run.py", line 756, in run
    )(*cmd_args)
  File "/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/distributed/launcher/api.py", line 132, in __call__
    return launch_agent(self._config, self._entrypoint, list(args))
  File "/root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/distributed/launcher/api.py", line 248, in launch_agent
    failures=result.failures,
torch.distributed.elastic.multiprocessing.errors.ChildFailedError: 

Reproduces the problem - code sample

When mmaigc was first released, I reproduced the dreambooth and did not encounter this problem... But today, when I cloned the latest code experiment, I encountered a dtype-related problem. The complete error report is as follows:

╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /root/modelscope/mmagic/tools/train.py:114 in <module>                                           │
│                                                                                                  │
│   111                                                                                            │
│   112                                                                                            │
│   113 if __name__ == '__main__':                                                                 │
│ ❱ 114 │   main()                                                                                 │
│   115                                                                                            │
│                                                                                                  │
│ /root/modelscope/mmagic/tools/train.py:107 in main                                               │
│                                                                                                  │
│   104 │   print_colored_log(f'Log directory: {runner._log_dir}')                                 │
│   105 │                                                                                          │
│   106 │   # start training                                                                       │
│ ❱ 107 │   runner.train()                                                                         │
│   108 │                                                                                          │
│   109 │   print_colored_log(f'Log saved under {runner._log_dir}')                                │
│   110 │   print_colored_log(f'Checkpoint saved under {cfg.work_dir}')                            │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/runner/runner.py:1721 in train  │
│                                                                                                  │
│   1718 │   │   # This must be called **AFTER** model has been wrapped.                           │
│   1719 │   │   self._maybe_compile('train_step')                                                 │
│   1720 │   │                                                                                     │
│ ❱ 1721 │   │   model = self.train_loop.run()  # type: ignore                                     │
│   1722 │   │   self.call_hook('after_run')                                                       │
│   1723 │   │   return model                                                                      │
│   1724                                                                                           │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/runner/loops.py:278 in run      │
│                                                                                                  │
│   275 │   │   │   self.runner.model.train()                                                      │
│   276 │   │   │                                                                                  │
│   277 │   │   │   data_batch = next(self.dataloader_iterator)                                    │
│ ❱ 278 │   │   │   self.run_iter(data_batch)                                                      │
│   279 │   │   │                                                                                  │
│   280 │   │   │   self._decide_current_val_interval()                                            │
│   281 │   │   │   if (self.runner.val_loop is not None                                           │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/runner/loops.py:302 in run_iter │
│                                                                                                  │
│   299 │   │   # synchronization during gradient accumulation process.                            │
│   300 │   │   # outputs should be a dict of loss.                                                │
│   301 │   │   outputs = self.runner.model.train_step(                                            │
│ ❱ 302 │   │   │   data_batch, optim_wrapper=self.runner.optim_wrapper)                           │
│   303 │   │                                                                                      │
│   304 │   │   self.runner.call_hook(                                                             │
│   305 │   │   │   'after_train_iter',                                                            │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/mmengine/model/wrappers/seperate_distrib │
│ uted.py:102 in train_step                                                                        │
│                                                                                                  │
│    99 │   │   Returns:                                                                           │
│   100 │   │   │   Dict[str, torch.Tensor]: A dict of tensor for logging.                         │
│   101 │   │   """                                                                                │
│ ❱ 102 │   │   return self.module.train_step(data, optim_wrapper)                                 │
│   103 │                                                                                          │
│   104 │   def val_step(self, data: Union[dict, tuple, list]) -> list:                            │
│   105 │   │   """Gets the prediction of module during validation process.                        │
│                                                                                                  │
│ /root/modelscope/mmagic/mmagic/models/editors/dreambooth/dreambooth.py:293 in train_step         │
│                                                                                                  │
│   290 │   │   │   model_output = self.unet(                                                      │
│   291 │   │   │   │   noisy_latents.float(),                                                     │
│   292 │   │   │   │   timesteps,                                                                 │
│ ❱ 293 │   │   │   │   encoder_hidden_states=encoder_hidden_states.float())                       │
│   294 │   │   │   model_pred = model_output['sample']                                            │
│   295 │   │   │                                                                                  │
│   296 │   │   │   loss_dict = dict()                                                             │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/parallel/distributed.py:1040 in │
│ forward                                                                                          │
│                                                                                                  │
│   1037 │   │   │   │   # Notify joined ranks whether they should sync in backwards pass or not.  │
│   1038 │   │   │   │   self._check_global_requires_backward_grad_sync(is_joined_rank=False)      │
│   1039 │   │   │                                                                                 │
│ ❱ 1040 │   │   │   output = self._run_ddp_forward(*inputs, **kwargs)                             │
│   1041 │   │   │                                                                                 │
│   1042 │   │   │   # sync params according to location (before/after forward) user               │
│   1043 │   │   │   # specified as part of hook, if hook was specified.                           │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/parallel/distributed.py:1000 in │
│ _run_ddp_forward                                                                                 │
│                                                                                                  │
│    997 │   │   │   │   self.use_side_stream_for_tensor_copies                                    │
│    998 │   │   │   )                                                                             │
│    999 │   │   │   with self._inside_ddp_forward():                                              │
│ ❱ 1000 │   │   │   │   return module_to_run(*inputs[0], **kwargs[0])                             │
│   1001 │   │   else:                                                                             │
│   1002 │   │   │   with self._inside_ddp_forward():                                              │
│   1003 │   │   │   │   return module_to_run(*inputs, **kwargs)                                   │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/modelscope/mmagic/mmagic/models/archs/wrapper.py:179 in forward                            │
│                                                                                                  │
│   176 │   │   Returns:                                                                           │
│   177 │   │   │   Any: The output of wrapped module's forward function.                          │
│   178 │   │   """                                                                                │
│ ❱ 179 │   │   return self.model(*args, **kwargs)                                                 │
│   180                                                                                            │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/diffusers/models/unet_2d_condition.py:71 │
│ 8 in forward                                                                                     │
│                                                                                                  │
│   715 │   │   │   encoder_hidden_states = self.encoder_hid_proj(encoder_hidden_states)           │
│   716 │   │                                                                                      │
│   717 │   │   # 2. pre-process                                                                   │
│ ❱ 718 │   │   sample = self.conv_in(sample)                                                      │
│   719 │   │                                                                                      │
│   720 │   │   # 3. down                                                                          │
│   721 │   │   down_block_res_samples = (sample,)                                                 │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/module.py:1194 in       │
│ _call_impl                                                                                       │
│                                                                                                  │
│   1191 │   │   # this function, and just call forward.                                           │
│   1192 │   │   if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks o  │
│   1193 │   │   │   │   or _global_forward_hooks or _global_forward_pre_hooks):                   │
│ ❱ 1194 │   │   │   return forward_call(*input, **kwargs)                                         │
│   1195 │   │   # Do not call functions when jit is used                                          │
│   1196 │   │   full_backward_hooks, non_full_backward_hooks = [], []                             │
│   1197 │   │   if self._backward_hooks or _global_backward_hooks:                                │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/conv.py:463 in forward  │
│                                                                                                  │
│    460 │   │   │   │   │   │   self.padding, self.dilation, self.groups)                         │
│    461 │                                                                                         │
│    462 │   def forward(self, input: Tensor) -> Tensor:                                           │
│ ❱  463 │   │   return self._conv_forward(input, self.weight, self.bias)                          │
│    464                                                                                           │
│    465 class Conv3d(_ConvNd):                                                                    │
│    466 │   __doc__ = r"""Applies a 3D convolution over an input signal composed of several inpu  │
│                                                                                                  │
│ /root/anaconda3/envs/mmagic/lib/python3.7/site-packages/torch/nn/modules/conv.py:460 in          │
│ _conv_forward                                                                                    │
│                                                                                                  │
│    457 │   │   │   │   │   │   │   weight, bias, self.stride,                                    │
│    458 │   │   │   │   │   │   │   _pair(0), self.dilation, self.groups)                         │
│    459 │   │   return F.conv2d(input, weight, bias, self.stride,                                 │
│ ❱  460 │   │   │   │   │   │   self.padding, self.dilation, self.groups)                         │
│    461 │                                                                                         │
│    462 │   def forward(self, input: Tensor) -> Tensor:                                           │
│    463 │   │   return self._conv_forward(input, self.weight, self.bias)                          │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
RuntimeError: Input type (float) and bias type (c10::Half) should be the same

Reproduces the problem - command or script

ditto

Reproduces the problem - error message

ditto

Additional information

No response

[wangqiang9]

I used the following command to solve this bug:

git checkout 0d37c51391

Indicates that the commit after this commit introduced this new problem.

https://github.com/open-mmlab/mmagic/commit/15c6dadefd82254d922b58408625b8065d4e043f