[BUG]save model fail for RuntimeError: "mse_cpu" not implemented for 'BFloat16' #1279
Description
Describe the bug
I used cellpose gui to hand annotation and retrain the model. I meet the error in saving model with RuntimeError: "mse_cpu" not implemented for 'BFloat16'
I used the MacBook Pro, 2.3 GHz 8-Core Intel Core i9
Run log
python -m cellpose
,,,
2025-07-22 10:26:48,093 [INFO] >>> saving model to /Users/cell_seg_res/S1R_5/analysis_outputs/models/cpsam_20250722_090158
Traceback (most recent call last):
File "/Users/yunxiawang/Downloads/cellpose-main/cellpose/gui/gui.py", line 1809, in new_model
self.train_model(restore=restore, normalize_params=normalize_params)
File "/Users/yunxiawang/Downloads/cellpose-main/cellpose/gui/gui.py", line 1826, in train_model
self.new_model_path, train_losses = train.train_seg(
File "/Users/yunxiawang/Downloads/cellpose-main/cellpose/train.py", line 463, in train_seg
loss = _loss_fn_seg(lbl, y, device)
File "/Users/yunxiawang/Downloads/cellpose-main/cellpose/train.py", line 49, in _loss_fn_seg
loss = criterion(y[:, -3:-1], veci)
File "/Users/yunxiawang/miniconda3/envs/cellpose/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/Users/yunxiawang/miniconda3/envs/cellpose/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/Users/yunxiawang/miniconda3/envs/cellpose/lib/python3.8/site-packages/torch/nn/modules/loss.py", line 535, in forward
return F.mse_loss(input, target, reduction=self.reduction)
File "/Users/yunxiawang/miniconda3/envs/cellpose/lib/python3.8/site-packages/torch/nn/functional.py", line 3339, in mse_loss
return torch._C._nn.mse_loss(expanded_input, expanded_target, _Reduction.get_enum(reduction))
RuntimeError: "mse_cpu" not implemented for 'BFloat16'
,,,
Screenshots
process figures
Error figures
Please help me. Thank you