ZhiJieLiu

On the mistakes of DeepLab in arcgis.learn

Discussion created by ZhiJieLiu on Feb 25, 2020
In [9]  :  dl.lr_find()

LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.
---------------------------------------------------------------------------TypeError                                 Traceback (most recent call last)<ipython-input-9-ff54e9d20eac> in <module>----> 1 dl.lr_find()~\AppData\Local\ESRI\conda\envs\TensorFlow\lib\site-packages\arcgis\learn\models\_arcgis_model.py in lr_find(self, allow_plot)    283         self._check_requisites()    284 --> 285         self.learn.lr_find()    286         from IPython.display import clear_output    287         clear_output()~\AppData\Local\ESRI\conda\envs\TensorFlow\lib\site-packages\fastai\train.py in lr_find(learn, start_lr, end_lr, num_it, stop_div, wd)     30     cb = LRFinder(learn, start_lr, end_lr, num_it, stop_div)     31     epochs = int(np.ceil(num_it/len(learn.data.train_dl)))---> 32     learn.fit(epochs, start_lr, callbacks=[cb], wd=wd)     33      34 def to_fp16(learn:Learner, loss_scale:float=None, max_noskip:int=1000, dynamic:bool=True, clip:float=None, ~\AppData\Local\ESRI\conda\envs\TensorFlow\lib\site-packages\fastai\basic_train.py in fit(self, epochs, lr, wd, callbacks)    198         callbacks = [cb(self) for cb in self.callback_fns + listify(defaults.extra_callback_fns)] + listify(callbacks)    199         if defaults.extra_callbacks is not None: callbacks += defaults.extra_callbacks--> 200         fit(epochs, self, metrics=self.metrics, callbacks=self.callbacks+callbacks)    201     202     def create_opt(self, lr:Floats, wd:Floats=0.)->None:~\AppData\Local\ESRI\conda\envs\TensorFlow\lib\site-packages\fastai\basic_train.py in fit(epochs, learn, callbacks, metrics)     99             for xb,yb in progress_bar(learn.data.train_dl, parent=pbar):    100                 xb, yb = cb_handler.on_batch_begin(xb, yb)--> 101                 loss = loss_batch(learn.model, xb, yb, learn.loss_func, learn.opt, cb_handler)    102                 if cb_handler.on_batch_end(loss): break    103  ~\AppData\Local\ESRI\conda\envs\TensorFlow\lib\site-packages\fastai\basic_train.py in loss_batch(model, xb, yb, loss_func, opt, cb_handler)     28      29     if not loss_func: return to_detach(out), yb[0].detach()---> 30     loss = loss_func(out, *yb)     31      32     if opt is not None:TypeError: _deeplab_loss() takes 3 positional arguments but 4 were given

Outcomes