Call predictor with positional arguments

examples/ConvolutionalPoseMachines/load-cpm.py

@@ -116,7 +116,7 @@ def run_test(model_path, img_file):
 
     im = cv2.imread(img_file, cv2.IMREAD_COLOR).astype('float32')
     im = cv2.resize(im, (368, 368))
-    out = predict_func([[im]])[0][0]
+    out = predict_func(im[None, :, :, :])[0][0]
     hm = out[:, :, :14].sum(axis=2)
     viz = colorize(im, hm)
     cv2.imwrite("output.jpg", viz)

examples/DeepQNetwork/common.py

@@ -20,7 +20,7 @@ def play_one_episode(env, func, render=False):
         """
         Map from observation to action, with 0.001 greedy.
         """
-        act = func([[s]])[0][0].argmax()
+        act = func(s[None, :, :, :])[0][0].argmax()
         if random.random() < 0.001:
             spc = env.action_space
             act = spc.sample()

examples/DeepQNetwork/expreplay.py

@@ -199,7 +199,7 @@ class ExpReplay(DataFlow, Callback):
             history = np.stack(history, axis=2)
 
             # assume batched network
-            q_values = self.predictor([[history]])[0][0]  # this is the bottleneck
+            q_values = self.predictor(history[None, :, :, :])[0][0]  # this is the bottleneck
             act = np.argmax(q_values)
         self._current_ob, reward, isOver, info = self.player.step(act)
         if isOver:

examples/DoReFa-Net/alexnet-dorefa.py

@@ -284,7 +284,7 @@ def run_image(model, sess_init, inputs):
         assert img is not None
 
         img = transformers.augment(img)[np.newaxis, :, :, :]
-        outputs = predictor([img])[0]
+        outputs = predictor(img)[0]
         prob = outputs[0]
         ret = prob.argsort()[-10:][::-1]
 

examples/DoReFa-Net/resnet-dorefa.py

@@ -139,7 +139,7 @@ def run_image(model, sess_init, inputs):
         assert img is not None
 
         img = transformers.augment(img)[np.newaxis, :, :, :]
-        o = predict_func([img])
+        o = predict_func(img)
         prob = o[0][0]
         ret = prob.argsort()[-10:][::-1]
 

examples/DynamicFilterNetwork/steering-filter.py

@@ -78,7 +78,7 @@ class OnlineTensorboardExport(Callback):
             x /= x.max()
             return x
 
-        o = self.pred([self.theta])
+        o = self.pred(self.theta)
 
         gt_filters = np.concatenate([self.filters[i, :, :] for i in range(8)], axis=0)
         pred_filters = np.concatenate([o[0][i, :, :, 0] for i in range(8)], axis=0)

examples/FasterRCNN/eval.py

@@ -94,7 +94,7 @@ def detect_one_image(img, model_func):
     resizer = CustomResize(config.SHORT_EDGE_SIZE, config.MAX_SIZE)
     resized_img = resizer.augment(img)
     scale = (resized_img.shape[0] * 1.0 / img.shape[0] + resized_img.shape[1] * 1.0 / img.shape[1]) / 2
-    fg_probs, fg_boxes = model_func([resized_img])
+    fg_probs, fg_boxes = model_func(resized_img)
     fg_boxes = fg_boxes / scale
     fg_boxes = clip_boxes(fg_boxes, img.shape[:2])
     return nms_fastrcnn_results(fg_boxes, fg_probs)

examples/HED/hed.py

@@ -198,8 +198,10 @@ def run(model_path, image_path, output):
     predictor = OfflinePredictor(pred_config)
     im = cv2.imread(image_path)
     assert im is not None
-    im = cv2.resize(im, (im.shape[1] // 16 * 16, im.shape[0] // 16 * 16))
-    outputs = predictor([[im.astype('float32')]])
+    im = cv2.resize(
+        im, (im.shape[1] // 16 * 16, im.shape[0] // 16 * 16)
+    )[None, :, :, :].astype('float32')
+    outputs = predictor(im)
     if output is None:
         for k in range(6):
             pred = outputs[k][0]

examples/ResNet/load-resnet.py

@@ -98,7 +98,7 @@ def run_test(params, input):
     im = cv2.imread(input).astype('float32')
     im = prepro.augment(im)
     im = np.reshape(im, (1, 224, 224, 3))
-    outputs = predict_func([im])
+    outputs = predict_func(im)
     prob = outputs[0]
 
     ret = prob[0].argsort()[-10:][::-1]

examples/Saliency/saliency-maps.py

@@ -42,7 +42,7 @@ def run(model_path, image_path):
     im = cv2.resize(im, (IMAGE_SIZE, IMAGE_SIZE))
     im = im.astype(np.float32)[:, :, ::-1]
 
-    saliency_images = predictor([im])[0]
+    saliency_images = predictor(im)[0]
 
     abs_saliency = np.abs(saliency_images).max(axis=-1)
     pos_saliency = np.maximum(0, saliency_images)

examples/SimilarityLearning/mnist-embeddings.py

@@ -387,7 +387,7 @@ def visualize(model_path, model, algo_name):
 
     for offset, dp in enumerate(ds.get_data()):
         digit, label = dp
-        prediction = pred([digit])[0]
+        prediction = pred(digit)[0]
         embed[offset * BATCH_SIZE:offset * BATCH_SIZE + BATCH_SIZE, ...] = prediction
         images[offset * BATCH_SIZE:offset * BATCH_SIZE + BATCH_SIZE, ...] = digit
         offset += 1

examples/SpatialTransformer/mnist-addition.py

@@ -140,7 +140,7 @@ def view_warp(modelpath):
     ds.reset_state()
     for k in ds.get_data():
         img, label = k
-        outputs, affine1, affine2 = pred([img])
+        outputs, affine1, affine2 = pred(img)
         for idx, viz in enumerate(outputs):
             viz = cv2.cvtColor(viz, cv2.COLOR_GRAY2BGR)
             # Here we assume the second branch focuses on the first digit

examples/load-alexnet.py

@@ -65,7 +65,7 @@ def run_test(path, input):
     assert im is not None, input
     im = cv2.resize(im, (227, 227))[:, :, ::-1].reshape(
         (1, 227, 227, 3)).astype('float32') - 110
-    outputs = predictor([im])[0]
+    outputs = predictor(im)[0]
     prob = outputs[0]
     ret = prob.argsort()[-10:][::-1]
     print("Top10 predictions:", ret)

examples/load-vgg16.py

@@ -76,7 +76,7 @@ def run_test(path, input):
     im = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)
     im = cv2.resize(im, (224, 224)).reshape((1, 224, 224, 3)).astype('float32')
     im = im - 110
-    outputs = predict_func([im])[0]
+    outputs = predict_func(im)[0]
     prob = outputs[0]
     ret = prob.argsort()[-10:][::-1]
     print("Top10 predictions:", ret)

tensorpack/predict/base.py

@@ -10,6 +10,7 @@ import six
 from ..tfutils.common import get_tensors_by_names
 from ..tfutils.tower import TowerContext
 from ..input_source import PlaceholderInput
+from ..utils.develop import log_deprecated
 
 __all__ = ['PredictorBase', 'AsyncPredictorBase',
            'OnlinePredictor', 'OfflinePredictor',
@@ -30,22 +31,21 @@ class PredictorBase(object):
         """
         Call the predictor on some inputs.
 
-        If ``len(args) == 1``, assume ``args[0]`` is a datapoint (a list).
-        otherwise, assume ``args`` is a datapoinnt
-
         Examples:
-            When you have a predictor which takes a datapoint [e1, e2], you
-            can call it in two ways:
+            When you have a predictor defined with two inputs, call it with:
 
             .. code-block:: python
 
                 predictor(e1, e2)
-                predictor([e1, e2])
         """
-        if len(args) != 1:
-            dp = args
+        if len(args) == 1 and isinstance(args[0], (list, tuple)):
+            dp = args[0]    # backward-compatibility
+            log_deprecated(
+                "Calling a predictor with one datapoint",
+                "Call it with positional arguments instead!",
+                "2018-3-1")
         else:
-            dp = args[0]
+            dp = args
         output = self._do_call(dp)
         if self.return_input:
             return (dp, output)

tensorpack/predict/concurrency.py

@@ -3,6 +3,7 @@
 # File: concurrency.py
 # Author: Yuxin Wu <ppwwyyxxc@gmail.com>
 
+import numpy as np
 import multiprocessing
 import six
 from six.moves import queue, range
@@ -71,7 +72,7 @@ class MultiProcessQueuePredictWorker(MultiProcessPredictWorker):
                 self.outqueue.put((DIE, None))
                 return
             else:
-                self.outqueue.put((tid, self.predictor(dp)))
+                self.outqueue.put((tid, self.predictor(*dp)))
 
 
 class PredictorWorkerThread(StoppableThread, ShareSessionThread):
@@ -89,7 +90,7 @@ class PredictorWorkerThread(StoppableThread, ShareSessionThread):
             while not self.stopped():
                 batched, futures = self.fetch_batch()
                 try:
-                    outputs = self.func(batched)
+                    outputs = self.func(*batched)
                 except tf.errors.CancelledError:
                     for f in futures:
                         f.cancel()
@@ -122,6 +123,9 @@ class PredictorWorkerThread(StoppableThread, ShareSessionThread):
                 futures.append(f)
             except queue.Empty:
                 break   # do not wait
+
+        for k in range(nr_input_var):
+            batched[k] = np.asarray(batched[k])
         return batched, futures
 
 

tensorpack/predict/dataset.py

@@ -73,7 +73,7 @@ class SimpleDatasetPredictor(DatasetPredictorBase):
             sz = 0
         with get_tqdm(total=sz, disable=(sz == 0)) as pbar:
             for dp in self.dataset.get_data():
-                res = self.predictor(dp)
+                res = self.predictor(*dp)
                 yield res
                 pbar.update()
 

tensorpack/tfutils/tower.py

@@ -277,3 +277,8 @@ class TowerTensorHandle(object):
         The output returned by the tower function.
         """
         return self._output
+
+    # def make_callable(self, input_names, output_names):
+    #     input_tensors = self.get_tensors(input_names)
+    #     output_tensors = self.get_tensors(output_names)
+    #     pass