lopezurrutia/dsb_2018

dsb_2018's Issues

Getting no (aspect_ratio) attribute error while trying to run inspect_nucleus_model.ipynb

Hi all,
I am very new to python kaggle and github so sorry for any inconvenience

I am trying to run inspect_nucleus_model.ipynb I got till Run Detection but there I have the following output:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-16-09616659d17f> in <module>()
      1 image_id = random.choice(dataset.image_ids)
----> 2 image, image_meta, gt_class_id, gt_bbox, gt_mask =modellib.load_image_gt(dataset, config, image_id, use_mini_mask=False)
      3 info = dataset.image_info[image_id]
      4 print("image ID: {}.{} ({}) {}".format(info["source"], info["id"], image_id, 
      5                                        dataset.image_reference(image_id)))

/data_l59/udinc/0_Downloads/DSB_2018-master_n/mrcnn/model.py in load_image_gt(dataset, config, image_id, augment, augmentation, use_mini_mask)
   1294         min_scale=config.IMAGE_MIN_SCALE,
   1295         max_dim=config.IMAGE_MAX_DIM,
-> 1296         aspect_ratio = config.ASPECT_RATIO,
   1297         min_enlarge = config.MIN_ENLARGE,
   1298         zoom = config.ZOOM,

AttributeError: 'NucleusInferenceConfig' object has no attribute 'ASPECT_RATIO'

Thanks
Ugur

How to Build New Data

Hi,
I want to train a new models with other data. but I didn't see you used the "stage1_train_labels.csv" file in the project.
My question:
Did you used the "stage1_train_labels.csv" file in your project?

The model weights link is dead

Could you please update the download link? The following link is dead:

https://drive.google.com/file/d/19kVton20JL9u0CpwGssD7EbBvsWcq1ty/view?usp=sharing

Thank you.

error

If I use 4 gpus,:
InvalidArgumentError (see above for traceback): Integer division by zero
[[Node: training/SGD/gradients/mrcnn_bbox_loss_1/concat_grad/mod = FloorMod[T=DT_INT32, _class=["loc:@mrcnn_bbox_loss_1/concat"], _device="/job:localhost/replica:0/task:0/device:CPU:0"](split_2/split_dim, training/SGD/gradients/mrcnn_bbox_loss_1/concat_grad/Rank)]]
[[Node: training/SGD/gradients/tower_3/mask_rcnn/res4t_branch2a/BiasAdd_grad/BiasAddGrad/_13835 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:3", send_device_incarnation=1, tensor_name="edge_76975_training/SGD/gradients/tower_3/mask_rcnn/res4t_branch2a/BiasAdd_grad/BiasAddGrad", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:GPU:0"]]

tensorflow 1.4
keras 2.1

model link

Can you share a link to your trained model? thanks!

Might it be that VAL_IMAGE_IDS are not in the stage1_train data?

Hi Again

I know this sounds crazy but I can't just find this set of images in the stage1_training data.

Here is the list of it as it is in the code (nucleus.py lines 69-95)

VAL_IMAGE_IDS = [
    "0c2550a23b8a0f29a7575de8c61690d3c31bc897dd5ba66caec201d201a278c2",
    "92f31f591929a30e4309ab75185c96ff4314ce0a7ead2ed2c2171897ad1da0c7",
    "1e488c42eb1a54a3e8412b1f12cde530f950f238d71078f2ede6a85a02168e1f",
    "c901794d1a421d52e5734500c0a2a8ca84651fb93b19cec2f411855e70cae339",
    "8e507d58f4c27cd2a82bee79fe27b069befd62a46fdaed20970a95a2ba819c7b",
    "60cb718759bff13f81c4055a7679e81326f78b6a193a2d856546097c949b20ff",
    "da5f98f2b8a64eee735a398de48ed42cd31bf17a6063db46a9e0783ac13cd844",
    "9ebcfaf2322932d464f15b5662cae4d669b2d785b8299556d73fffcae8365d32",
    "1b44d22643830cd4f23c9deadb0bd499fb392fb2cd9526d81547d93077d983df",
    "97126a9791f0c1176e4563ad679a301dac27c59011f579e808bbd6e9f4cd1034",
    "e81c758e1ca177b0942ecad62cf8d321ffc315376135bcbed3df932a6e5b40c0",
    "f29fd9c52e04403cd2c7d43b6fe2479292e53b2f61969d25256d2d2aca7c6a81",
    "0ea221716cf13710214dcd331a61cea48308c3940df1d28cfc7fd817c83714e1",
    "3ab9cab6212fabd723a2c5a1949c2ded19980398b56e6080978e796f45cbbc90",
    "ebc18868864ad075548cc1784f4f9a237bb98335f9645ee727dac8332a3e3716",
    "bb61fc17daf8bdd4e16fdcf50137a8d7762bec486ede9249d92e511fcb693676",
    "e1bcb583985325d0ef5f3ef52957d0371c96d4af767b13e48102bca9d5351a9b",
    "947c0d94c8213ac7aaa41c4efc95d854246550298259cf1bb489654d0e969050",
    "cbca32daaae36a872a11da4eaff65d1068ff3f154eedc9d3fc0c214a4e5d32bd",
    "f4c4db3df4ff0de90f44b027fc2e28c16bf7e5c75ea75b0a9762bbb7ac86e7a3",
    "4193474b2f1c72f735b13633b219d9cabdd43c21d9c2bb4dfc4809f104ba4c06",
    "f73e37957c74f554be132986f38b6f1d75339f636dfe2b681a0cf3f88d2733af",
    "a4c44fc5f5bf213e2be6091ccaed49d8bf039d78f6fbd9c4d7b7428cfcb2eda4",
    "cab4875269f44a701c5e58190a1d2f6fcb577ea79d842522dcab20ccb39b7ad2",
    "8ecdb93582b2d5270457b36651b62776256ade3aaa2d7432ae65c14f07432d49",
]

Here is the link to the data

https://www.kaggle.com/c/data-science-bowl-2018/data

Best,
Ugur

I get a worse score

Hi, thanks for your excellent work!
Now I am reproducing your work but have a problem here. After three steps of training, I get a score with 0.43549 on Kaggle (but the model weights which you provide get a score 0.61450).
Could you help me with how could that happen? I have the following pipeline:
1.python my_train_1.py
2.python my_train_2.py
3.python my_train_3.py
4.python my_inference_tta.py