fixing bug in writing computing_masks... and updating docs

README.md

@@ -35,6 +35,8 @@ and updated GUI:
 
 ![gui screenshot](https://www.cellpose.org/static/images/cellpose_gui.png)
 
+Mxnet is no longer supported in cellpose. To use mxnet, please use v1.0.2.
+
 ### UPDATE v1.0 (Jan 2022)
 
 Cellpose has been relatively stable for a while now. Small bugs will continue to be fixed, but we are now releasing a reference 1.0 version. Larger updates to Cellpose will go towards a new 2.0 candidate version to be released soon. 
@@ -45,6 +47,11 @@ from cellpose.io import logger_setup
 logger_setup();
 ```
 
+To install this version please use
+```
+pip install cellpose==1.0.2
+```
+
 ### UPDATE v0.7 (Nov 2021)
 
 [Omnipose](https://github.com/kevinjohncutler/omnipose) is now officially available and supported as part of Cellpose. `pip install omnipose` and use the 'omni' flag and models to take advantage of it for long cells! (<i>E.g.</i>, filamentous bacteria.) Omnipose was written by Kevin Cutler ([@kevinjohncutler](https://github.com/kevinjohncutler)). To learn about Omnipose, read the [paper](https://biorxiv.org/content/early/2021/11/04/2021.11.03.467199). If you use Omnipose in your work please cite the Cellpose paper and the Omnipose paper.
@@ -130,17 +137,6 @@ Follow the instructions [here](https://pytorch.org/get-started/locally/) to dete
 conda install pytorch cudatoolkit=10.2 -c pytorch
 ~~~~
 
-For the GPU version of mxnet, you will need to install the cuda toolkit first if you haven't already (on Windows it may be necessary to install via anaconda as below):
-
-~~~
-conda install -c anaconda cudatoolkit
-~~~
-
-When upgrading GPU Cellpose in the future, you will want to ignore dependencies (to ensure that the pip version of torch does not install):
-~~~
-pip install --no-deps cellpose --upgrade
-~~~
-
 ### Installation of github version
 
 Follow steps from above to install the dependencies. Then run 
@@ -246,7 +242,7 @@ See the [docs](https://cellpose.readthedocs.io/en/latest/outputs.html) for info.
 
 ## Dependencies
 cellpose relies on the following excellent packages (which are automatically installed with conda/pip if missing):
-- [mxnet_mkl](https://mxnet.apache.org/)
+- [pytorch](https://pytorch.org/)
 - [pyqtgraph](https://pyqtgraph.org/)
 - [PyQt5](https://pyqt.sourceforge.net/Docs/PyQt5/)
 - [numpy](https://www.numpy.org/) (>=1.16.0)

cellpose/gui/gui.py

@@ -1642,11 +1642,7 @@ def get_thresholds(self):
  cellprob_threshold = float(self.cellprob_threshold.text())
  if flow_threshold==0.0 or self.NZ>1:
  flow_threshold = None
- logger.info('computing masks with cell prob=%0.3f, no flow error threshold'%
- (cellprob_threshold))
- else:
- logger.info('computing masks with cell prob=%0.3f, flow error threshold=%0.3f'%
- (cellprob_threshold, flow_threshold))
+
  return flow_threshold, cellprob_threshold
  except Exception as e:
  print('flow threshold or cellprob threshold not a valid number, setting to defaults')
@@ -1657,6 +1653,12 @@ def get_thresholds(self):
  def compute_cprob(self):
  if self.recompute_masks:
  flow_threshold, cellprob_threshold = self.get_thresholds()
+ if flow_threshold is None:
+ logger.info('computing masks with cell prob=%0.3f, no flow error threshold'%
+ (cellprob_threshold))
+ else:
+ logger.info('computing masks with cell prob=%0.3f, flow error threshold=%0.3f'%
+ (cellprob_threshold, flow_threshold))
  maski = dynamics.compute_masks(self.flows[4][:-1], 
  self.flows[4][-1],
  p=self.flows[3].copy(),

docs/command.rst

@@ -31,11 +31,14 @@ These are the same :ref:`settings`, but set up for the command line, e.g.
  * use_gpu: (bool)
  run network on GPU
 
+ * save_outlines: FLAG
+ save outlines as text file for ImageJ
+
  * save_png: FLAG
- save masks as png and outlines as text file for ImageJ
+ save masks as png  
 
  * save_tif: FLAG
- save masks as tif and outlines as text file for ImageJ
+ save masks as tif
 
  * no_npy: FLAG 
  turn off saving of _seg.npy file 
@@ -73,6 +76,7 @@ Options
 You can run the help string and see all the options:
 
 ::
+ 
  usage: __main__.py [-h] [--use_gpu] [--check_mkl] [--dir DIR]
  [--look_one_level_down] [--img_filter IMG_FILTER]
  [--channel_axis CHANNEL_AXIS] [--z_axis Z_AXIS]

docs/gui.rst

@@ -15,6 +15,18 @@ You can **drag and drop** images (.tif, .png, .jpg, .gif) into the GUI and run C
 
 For multi-channel, multi-Z tiff's, the expected format is Z x channels x Ly x Lx.
 
+.. note::
+ The output file with the masks is in the same folder as the loaded
+ image with ``_seg.npy`` appended. The GUI automatically saves after you 
+ draw an ROI but NOT after running a model for
+ segmentation and NOT after 3D mask drawing (too slow). Save in the file
+ menu or with Ctrl+S. 
+
+.. note::
+ Since the output file is in the same folder as the loaded
+ image with ``_seg.npy`` appended, make sure you have WRITE access 
+ in the folder, otherwise the file will not save.
+
 Using the GUI 
 ~~~~~~~~~~~~~~~~~~~~~~~
 
@@ -41,14 +53,10 @@ in 2D should be single strokes (if *single_stroke* is checked).
 
 If you want to draw masks in 3D, then you can turn *single_stroke*
 option off and draw a stroke on each plane with the cell and then press
-ENTER. 3D labelling will fill in unlabelled z-planes so that you do not
+ENTER (cellpose 1.0 only currently). 
+3D labelling will fill in unlabelled z-planes so that you do not
 have to as densely label.
 
-.. note::
- The GUI automatically saves after you draw a mask but NOT after
- segmentation and NOT after 3D mask drawing (too slow). Save in the file
- menu or with Ctrl+S. The output file is in the same folder as the loaded
- image with ``_seg.npy`` appended.
 
 Segmentation options
 ~~~~~~~~~~~~~~~~~~~~~~~~

docs/installation.rst

@@ -33,16 +33,6 @@ this code at the beginning of your notebook before you import cellpose:
 Common issues
 ~~~~~~~~~~~~~~~~~~~~~~~
 
-If you receive the error: ``Illegal instruction (core dumped)``, then
-likely mxnet does not recognize your MKL version. Please uninstall and
-reinstall mxnet without mkl:
-
-::
-
- pip uninstall mxnet-mkl
- pip uninstall mxnet
- pip install mxnet==1.4.0
-
 If you receive the error: ``No module named PyQt5.sip``, then try
 uninstalling and reinstalling pyqt5
 
@@ -84,7 +74,7 @@ Dependencies
 cellpose relies on the following excellent packages (which are
 automatically installed with conda/pip if missing):
 
-- `mxnet_mkl`_
+- `pytorch`_
 - `pyqtgraph`_
 - `PyQt5`_
 - `numpy`_ (>=1.16.0)
@@ -98,7 +88,7 @@ automatically installed with conda/pip if missing):
 .. _environment.yml: https://github.com/MouseLand/cellpose/blob/master/environment.yml?raw=true
 .. _here: https://pypi.org/project/cellpose/
 
-.. _mxnet_mkl: https://mxnet.apache.org/
+.. _pytorch: https://pytorch.org/
 .. _pyqtgraph: https://pyqtgraph.org/
 .. _PyQt5: https://pyqt.sourceforge.net/Docs/PyQt5/
 .. _numpy: https://www.numpy.org/

docs/models.rst

@@ -1,8 +1,15 @@
 Models
 ------------------------------
 
+``from cellpose import models``
+
 Cellpose 2.0 now has a model zoo and options for user model training. 
-See paper for more details.
+Each model will be downloaded automatically to your ``models.MODELS_DIR`` 
+(see Installation instructions for more details on MODELS_DIR). 
+See paper for more details on the model zoo. You can also directly download a 
+model by going to the URL, e.g.:
+
+``https://www.cellpose.org/models/MODEL_NAME``
 
 Model Zoo
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -26,6 +33,11 @@ run with ``diameter=0`` or ``--diameter 0`` and the model can estimate the ROI s
 These models can be loaded and used in the notebook with ``models.Cellpose(model_type='cyto')`` 
 or in the command line with ``python -m cellpose --pretrained_model cyto``.
 
+These models' names (to download all the models for a class run with ``--net_avg``): 
+* `'cyto'`: ``cytotorch_0``, ``cytotorch_1``, ``cytotorch_2``, ``cytotorch_3``, ``size_cytotorch_0.npy``
+* `'nuclei'`: ``nucleitorch_0``, ``nucleitorch_1``, ``nucleitorch_2``, ``nucleitorch_3``, ``size_nucleitorch_0.npy``
+* `'cyto2'`: ``cyto2torch_0``, ``cyto2torch_1``, ``cyto2torch_2``, ``cyto2torch_3``, ``size_cyto2torch_0.npy``
+
 Cytoplasm model (`'cyto'`)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
@@ -71,6 +83,8 @@ These models can be loaded and used in the notebook with e.g.
 ``models.CellposeModel(model_type='tissuenet')`` or ``models.CellposeModel(model_type='LC2')``, 
 or in the command line with ``python -m cellpose --pretrained_model tissuenet``.
 
+These models' names are the same as their strings in the GUI.
+
 TissueNet models
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^