@@ -83,7 +83,7 @@ class CellposeModel():
8383 __init__(self, gpu=False, pretrained_model=False, model_type=None, diam_mean=30., device=None):
8484 Initialize the CellposeModel.
8585
86- eval(self, x, batch_size=8, resample=True, channels=None, channel_axis=None, z_axis=None, normalize=True, invert=False, rescale=None, diameter=None, flow_threshold=0.4, cellprob_threshold=0.0, do_3D=False, anisotropy=None, stitch_threshold=0.0, min_size=15, niter=None, augment=False, tile_overlap=0.1, bsize=224 , interp=True, compute_masks=True, progress=None):
86+ eval(self, x, batch_size=8, resample=True, channels=None, channel_axis=None, z_axis=None, normalize=True, invert=False, rescale=None, diameter=None, flow_threshold=0.4, cellprob_threshold=0.0, do_3D=False, anisotropy=None, stitch_threshold=0.0, min_size=15, niter=None, augment=False, tile_overlap=0.1, bsize=256 , interp=True, compute_masks=True, progress=None):
8787 Segment list of images x, or 4D array - Z x C x Y x X.
8888
8989 """
@@ -196,7 +196,7 @@ def eval(self, x, batch_size=8, resample=True, channels=None, channel_axis=None,
196196 niter (int, optional): number of iterations for dynamics computation. if None, it is set proportional to the diameter. Defaults to None.
197197 augment (bool, optional): tiles image with overlapping tiles and flips overlapped regions to augment. Defaults to False.
198198 tile_overlap (float, optional): fraction of overlap of tiles when computing flows. Defaults to 0.1.
199- bsize (int, optional): block size for tiles, recommended to keep at 224 , like in training. Defaults to 224 .
199+ bsize (int, optional): block size for tiles, recommended to keep at 256 , like in training. Defaults to 256 .
200200 interp (bool, optional): interpolate during 2D dynamics (not available in 3D) . Defaults to True.
201201 compute_masks (bool, optional): Whether or not to compute dynamics and return masks. Returns empty array if False. Defaults to True.
202202 progress (QProgressBar, optional): pyqt progress bar. Defaults to None.
@@ -453,7 +453,7 @@ def _resize_gradients(self, grads: np.ndarray, to_y_size: int, to_x_size: int, t
453453 def _run_net (self , x ,
454454 augment = False ,
455455 batch_size = 8 , tile_overlap = 0.1 ,
456- bsize = 224 , anisotropy = 1.0 , do_3D = False ):
456+ bsize = 256 , anisotropy = 1.0 , do_3D = False ):
457457 """ run network on image x """
458458 tic = time .time ()
459459 shape = x .shape
0 commit comments