Semantic Segmentation Issue

[dokunima]

I am following the semantic segmentation tutorial to train a custom one-class dataset but I get the error below when I want to visualize the data or train the model. I will be grateful if anyone can suggest how I can fix this error.

AttributeError                            Traceback (most recent call last)

[<ipython-input-104-d5499f594444>](https://localhost:8080/#) in <cell line: 2>()
      1 sample_records = random.choices(train_records, k=3)
----> 2 show_records(sample_records, ncols=3)

5 frames

[/usr/local/lib/python3.9/dist-packages/icevision/visualize/draw_data.py](https://localhost:8080/#) in draw_sample(sample, class_map, denormalize_fn, display_label, display_bbox, display_score, display_mask, display_keypoints, font_path, font_size, label_color, label_border_color, label_thin_border, label_pad_width_factor, label_pad_height_factor, mask_blend, mask_border_thickness, color_map, prettify, prettify_func, return_as_pil_img, exclude_labels, include_only, multiple_classification_spacing_factor, dynamic_font_size_div_factor, include_classification_task_names, include_instances_task_names, force_mask_file_reload)
     91     for task, composite in sample.task_composites.items():
     92         if task == "segmentation":
---> 93             cm = rand_cmap(sample.segmentation.class_map.num_classes, verbose=False)
     94             if composite.mask_array is None:
     95                 if isinstance(composite.masks[0], RLE):

AttributeError: 'list' object has no attribute 'num_classes'

Thank you.

@lgvaz @FraPochetti @dnth @rsomani95 @jerbly