Ambiguous dimension while trying to load weights.

[federicoAntosiano]

Hi everyone, I'm trying to load the SSD300 for inferencing but I'm facing this ValueError:

---

ValueError Traceback (most recent call last) in 25 iou_threshold=0.45, 26 top_k=200, ---> 27 nms_max_output_size=400) 28 29 # 2: Load the trained weights into the model.

~/.../keras_ssd300.py in ssd_300(image_size, n_classes, mode, l2_regularization, min_scale, max_scale, scales, aspect_ratios_global, aspect_ratios_per_layer, two_boxes_for_ar1, steps, offsets, clip_boxes, variances, coords, normalize_coords, subtract_mean, divide_by_stddev, swap_channels, confidence_thresh, iou_threshold, top_k, nms_max_output_size, return_predictor_sizes) 340 conv4_3_norm_mbox_priorbox = AnchorBoxes(img_height, img_width, this_scale=scales[0], next_scale=scales[1], aspect_ratios=aspect_ratios[0], 341 two_boxes_for_ar1=two_boxes_for_ar1, this_steps=steps[0], this_offsets=offsets[0], clip_boxes=clip_boxes, --> 342 variances=variances, coords=coords, normalize_coords=normalize_coords, name='conv4_3_norm_mbox_priorbox')(conv4_3_norm_mbox_loc) 343 fc7_mbox_priorbox = AnchorBoxes(img_height, img_width, this_scale=scales[1], next_scale=scales[2], aspect_ratios=aspect_ratios[1], 344 two_boxes_for_ar1=two_boxes_for_ar1, this_steps=steps[1], this_offsets=offsets[1], clip_boxes=clip_boxes,

/opt/DL/tensorflow/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py in call(self, inputs, *args, **kwargs) 755 if not in_deferred_mode: 756 self._in_call = True --> 757 outputs = self.call(inputs, *args, **kwargs) 758 self._in_call = False 759 if outputs is None:

~/I.../keras_object_detection/ssd_keras/keras_layers/keras_layer_AnchorBoxes.py in call(self, x, mask) 203 offset_width = self.this_offsets 204 # Now that we have the offsets and step sizes, compute the grid of anchor box center points. --> 205 cy = np.linspace(offset_height * step_height, (offset_height + feature_map_height - 1) * step_height, feature_map_height) 206 cx = np.linspace(offset_width * step_width, (offset_width + feature_map_width - 1) * step_width, feature_map_width) 207 cx_grid, cy_grid = np.meshgrid(cx, cy)

/opt/DL/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/tensor_shape.py in radd(self, other) 180 A Dimension whose value is the sum of self and other. 181 """ --> 182 return self + other 183 184 def sub(self, other):

/opt/DL/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/tensor_shape.py in add(self, other) 165 A Dimension whose value is the sum of self and other. 166 """ --> 167 other = as_dimension(other) 168 if self._value is None or other.value is None: 169 return Dimension(None)

/opt/DL/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/tensor_shape.py in as_dimension(value) 480 return value 481 else: --> 482 return Dimension(value) 483 484

/opt/DL/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/tensor_shape.py in init(self, value) 38 if (not isinstance(value, compat.bytes_or_text_types) and 39 self._value != value): ---> 40 raise ValueError("Ambiguous dimension: %s" % value) 41 if self._value < 0: 42 raise ValueError("Dimension %d must be >= 0" % self._value)

ValueError: Ambiguous dimension: 0.5

Do you know why this error appears? How can I fix it?