Sparse Stardist training example

Once we have sparse labels the training process is the same as non-sparse.

import os
from tnia.plotting.plt_helper import imshow_multi2d
import tensorflow as tf
from pathlib import Path
import json

2024-10-23 06:02:12.641668: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2024-10-23 06:02:12.651885: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:479] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-10-23 06:02:12.666401: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:10575] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-10-23 06:02:12.666425: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1442] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2024-10-23 06:02:12.675863: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2024-10-23 06:02:13.194393: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT

Check what devices we have access to….

Not as important to have a beefy GPU for 2D as it is for 3D, but let’s check

visible_devices = tf.config.list_physical_devices()
print(visible_devices)

[PhysicalDevice(name='/physical_device:CPU:0', device_type='CPU'), PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]

2024-10-23 06:02:15.285568: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:02:15.327907: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:02:15.328063: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355

Load inputs and ground truth

We load directories called ‘input0’ and ‘ground truth0’ which should exist under train_path. The reason we append ‘0’ to the end of the name is simply because some of the code that generates image and label sets is meant to work on multiple channels (so the 0 is the channel number)

#tnia_images_path = Path(r'/home/bnorthan/images')
data_path = r'../../data'
parent_path = os.path.join(data_path, 'ladybugs_sparse')
#test_name='bsp1-2.jpg'
n_rays = 32

train_path = os.path.join(parent_path , 'patches')

with open(os.path.join(train_path , 'info.json'), 'r') as json_file:
    data = json.load(json_file)
    # Access the sub_sample parameter
    sub_sample = data['sub_sample']
    print('sub_sample',sub_sample)
    axes = data['axes']
    print('axes',axes)

image_patch_path = os.path.join(train_path , 'ground truth0')
label_patch_path = os.path.join(train_path , 'input0' )

model_path = os.path.join(parent_path , 'models')

if not os.path.exists(model_path):
    os.makedirs(model_path)

if not os.path.exists(image_patch_path):
    print('image_patch_path does not exist')

if not os.path.exists(label_patch_path):
    print('label_patch_path does not exist')

sub_sample 1
axes YXC

Use a helper to collect the training data

The helper will also optionally normalize the inputs.

Normalization is a tricky issue sometimes it makes sense to normalize before creating patches, such that the data is normalized based on statistics of a larger region, closer to the normalization range that will be used for prediction.

from tnia.deeplearning.dl_helper import collect_training_data
add_trivial_channel = False

X, Y = collect_training_data(train_path, sub_sample=1, downsample=False, normalize_input=False, add_trivial_channel = add_trivial_channel)

print('type X ', type(X))
print('type Y ', type(Y))

raster_geometry not imported.  This is only needed for the ellipsoid rendering in apply_stardist
type X  <class 'list'>
type Y  <class 'list'>

Inspect images

Output training data shapes and plot images to make sure image and label set look OK

n=35
print(X[n].shape, Y[n].shape)
print(X[n].min(), X[n].max())
print(Y[n].min(), Y[n].max())
fig=imshow_multi2d([X[n], Y[n]], ['input', 'label'], 1,2)

Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers). Got range [-0.007905139..1.0].

(256, 256, 3) (256, 256)
-0.007905139 1.0
0 65535

Check to make sure we have negative values in the labels

from tnia.deeplearning.dl_helper import divide_training_data

X_train, Y_train, X_val, Y_val = divide_training_data(X, Y, val_size=2)

Y_train = Y_train.astype('int16')
Y_val = Y_val.astype('int16')

print(Y_train.min(), Y_train.max())

-1 44

Create stardist model

In this cell we create the model. Make sure to rename the model and give it a descriptive name that conveys the training data and setting used.

from stardist.models import StarDist2D, Config2D
from tnia.deeplearning.dl_helper import augmenter

if axes == 'YXC':
    n_channel_in =3
else:
    n_channel_in = 1

model_name = "ladybug_sparse"
new_model = True 

if new_model:

    config = Config2D (n_rays=n_rays, axes=axes,n_channel_in=n_channel_in, train_patch_size = (256,256), unet_n_depth=3)
    model = StarDist2D(config=config, name=model_name, basedir=model_path)
else:
    model = StarDist2D(config=None, name=model_name, basedir=model_path)

base_model.py (198): output path for model already exists, files may be overwritten: /home/bnorthan/code/i2k/tnia/notebooks-and-napari-widgets-for-dl/data/ladybugs_sparse/models/ladybug_sparse
2024-10-23 06:06:23.260110: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:06:23.260265: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:06:23.260343: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:06:23.317818: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:06:23.317956: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:06:23.318048: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355
2024-10-23 06:06:23.318119: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1928] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 6264 MB memory:  -> device: 0, name: NVIDIA GeForce RTX 4070 Laptop GPU, pci bus id: 0000:01:00.0, compute capability: 8.9

Using default values: prob_thresh=0.5, nms_thresh=0.4.

Train the model

#model.train(X_train, Y_train, validation_data=(X_val,Y_val),epochs=100, steps_per_epoch=200, augmenter=augmenter) 
model.train(X_train, Y_train, validation_data=(X_val,Y_val),epochs=200, steps_per_epoch=200) 

Epoch 1/200

WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
I0000 00:00:1729678015.834298  126696 service.cc:145] XLA service 0x7de3a0002790 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
I0000 00:00:1729678015.834325  126696 service.cc:153]   StreamExecutor device (0): NVIDIA GeForce RTX 4070 Laptop GPU, Compute Capability 8.9
2024-10-23 06:06:55.919704: I tensorflow/compiler/mlir/tensorflow/utils/dump_mlir_util.cc:268] disabling MLIR crash reproducer, set env var `MLIR_CRASH_REPRODUCER_DIRECTORY` to enable.
2024-10-23 06:07:01.314635: I external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:465] Loaded cuDNN version 8907

  3/200 ━━━━━━━━━━━━━━━━━━━━ 9s 50ms/step - dist_dist_iou_metric: 7.4945e-06 - dist_relevant_mae: 16.6135 - dist_relevant_mse: 367.0920 - loss: 4.0206 - prob_kld: 0.3650  

I0000 00:00:1729678031.069700  126696 device_compiler.h:188] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.

200/200 ━━━━━━━━━━━━━━━━━━━━ 35s 85ms/step - dist_dist_iou_metric: 0.1627 - dist_relevant_mae: 11.3809 - dist_relevant_mse: 211.2150 - loss: 2.9400 - prob_kld: 0.3043 - val_dist_dist_iou_metric: 0.2841 - val_dist_relevant_mae: 9.9700 - val_dist_relevant_mse: 191.5270 - val_loss: 2.8396 - val_prob_kld: 0.3304 - learning_rate: 3.0000e-04
Epoch 2/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.3767 - dist_relevant_mae: 7.1337 - dist_relevant_mse: 91.8835 - loss: 2.0531 - prob_kld: 0.2562 - val_dist_dist_iou_metric: 0.2686 - val_dist_relevant_mae: 10.0539 - val_dist_relevant_mse: 198.0379 - val_loss: 2.6728 - val_prob_kld: 0.1469 - learning_rate: 3.0000e-04
Epoch 3/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.3968 - dist_relevant_mae: 6.8994 - dist_relevant_mse: 87.3241 - loss: 1.9318 - prob_kld: 0.1855 - val_dist_dist_iou_metric: 0.4432 - val_dist_relevant_mae: 7.3069 - val_dist_relevant_mse: 108.2121 - val_loss: 2.1132 - val_prob_kld: 0.1366 - learning_rate: 3.0000e-04
Epoch 4/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.4796 - dist_relevant_mae: 5.7217 - dist_relevant_mse: 63.2763 - loss: 1.6740 - prob_kld: 0.1531 - val_dist_dist_iou_metric: 0.3419 - val_dist_relevant_mae: 8.5012 - val_dist_relevant_mse: 150.5417 - val_loss: 2.3270 - val_prob_kld: 0.1116 - learning_rate: 3.0000e-04
Epoch 5/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.4964 - dist_relevant_mae: 5.4311 - dist_relevant_mse: 58.0395 - loss: 1.5534 - prob_kld: 0.1094 - val_dist_dist_iou_metric: 0.4269 - val_dist_relevant_mae: 7.2659 - val_dist_relevant_mse: 115.0903 - val_loss: 2.0640 - val_prob_kld: 0.0957 - learning_rate: 3.0000e-04
Epoch 6/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.5226 - dist_relevant_mae: 5.0477 - dist_relevant_mse: 50.8718 - loss: 1.4664 - prob_kld: 0.0900 - val_dist_dist_iou_metric: 0.4551 - val_dist_relevant_mae: 6.9290 - val_dist_relevant_mse: 101.6881 - val_loss: 2.0071 - val_prob_kld: 0.1061 - learning_rate: 3.0000e-04
Epoch 7/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.5544 - dist_relevant_mae: 4.5802 - dist_relevant_mse: 42.5514 - loss: 1.3897 - prob_kld: 0.1029 - val_dist_dist_iou_metric: 0.5209 - val_dist_relevant_mae: 6.0703 - val_dist_relevant_mse: 78.8316 - val_loss: 1.8247 - val_prob_kld: 0.0955 - learning_rate: 3.0000e-04
Epoch 8/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.5766 - dist_relevant_mae: 4.2513 - dist_relevant_mse: 37.2060 - loss: 1.2927 - prob_kld: 0.0768 - val_dist_dist_iou_metric: 0.5066 - val_dist_relevant_mae: 6.5110 - val_dist_relevant_mse: 83.5902 - val_loss: 1.9162 - val_prob_kld: 0.0988 - learning_rate: 3.0000e-04
Epoch 9/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.5964 - dist_relevant_mae: 3.9555 - dist_relevant_mse: 32.3517 - loss: 1.2139 - prob_kld: 0.0635 - val_dist_dist_iou_metric: 0.5198 - val_dist_relevant_mae: 6.0387 - val_dist_relevant_mse: 78.6560 - val_loss: 1.8324 - val_prob_kld: 0.1095 - learning_rate: 3.0000e-04
Epoch 10/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.6154 - dist_relevant_mae: 3.7544 - dist_relevant_mse: 29.8220 - loss: 1.1935 - prob_kld: 0.0693 - val_dist_dist_iou_metric: 0.5159 - val_dist_relevant_mae: 6.1397 - val_dist_relevant_mse: 78.4183 - val_loss: 1.8297 - val_prob_kld: 0.0866 - learning_rate: 3.0000e-04
Epoch 11/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.6303 - dist_relevant_mae: 3.5219 - dist_relevant_mse: 25.9147 - loss: 1.1289 - prob_kld: 0.0620 - val_dist_dist_iou_metric: 0.5337 - val_dist_relevant_mae: 5.9985 - val_dist_relevant_mse: 73.7102 - val_loss: 1.8037 - val_prob_kld: 0.0889 - learning_rate: 3.0000e-04
Epoch 12/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.6701 - dist_relevant_mae: 3.1137 - dist_relevant_mse: 21.1914 - loss: 1.0249 - prob_kld: 0.0500 - val_dist_dist_iou_metric: 0.5804 - val_dist_relevant_mae: 5.1865 - val_dist_relevant_mse: 56.3722 - val_loss: 1.6573 - val_prob_kld: 0.1048 - learning_rate: 3.0000e-04
Epoch 13/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.6927 - dist_relevant_mae: 2.8543 - dist_relevant_mse: 18.0229 - loss: 0.9833 - prob_kld: 0.0484 - val_dist_dist_iou_metric: 0.5919 - val_dist_relevant_mae: 4.8290 - val_dist_relevant_mse: 52.9113 - val_loss: 1.5534 - val_prob_kld: 0.0725 - learning_rate: 3.0000e-04
Epoch 14/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.7208 - dist_relevant_mae: 2.5489 - dist_relevant_mse: 14.6762 - loss: 0.9122 - prob_kld: 0.0389 - val_dist_dist_iou_metric: 0.6102 - val_dist_relevant_mae: 4.5877 - val_dist_relevant_mse: 46.3141 - val_loss: 1.5059 - val_prob_kld: 0.0733 - learning_rate: 3.0000e-04
Epoch 15/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.7260 - dist_relevant_mae: 2.5353 - dist_relevant_mse: 14.3680 - loss: 0.9121 - prob_kld: 0.0409 - val_dist_dist_iou_metric: 0.6668 - val_dist_relevant_mae: 3.9589 - val_dist_relevant_mse: 34.1588 - val_loss: 1.3918 - val_prob_kld: 0.0849 - learning_rate: 3.0000e-04
Epoch 16/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.7496 - dist_relevant_mae: 2.2154 - dist_relevant_mse: 11.3120 - loss: 0.8353 - prob_kld: 0.0332 - val_dist_dist_iou_metric: 0.6524 - val_dist_relevant_mae: 4.1353 - val_dist_relevant_mse: 36.5734 - val_loss: 1.4161 - val_prob_kld: 0.0738 - learning_rate: 3.0000e-04
Epoch 17/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.7559 - dist_relevant_mae: 2.1607 - dist_relevant_mse: 10.6495 - loss: 0.8206 - prob_kld: 0.0326 - val_dist_dist_iou_metric: 0.6968 - val_dist_relevant_mae: 3.4312 - val_dist_relevant_mse: 27.8910 - val_loss: 1.2603 - val_prob_kld: 0.0589 - learning_rate: 3.0000e-04
Epoch 18/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.7743 - dist_relevant_mae: 1.9756 - dist_relevant_mse: 9.0055 - loss: 0.7835 - prob_kld: 0.0273 - val_dist_dist_iou_metric: 0.6676 - val_dist_relevant_mae: 3.7073 - val_dist_relevant_mse: 32.4118 - val_loss: 1.3081 - val_prob_kld: 0.0515 - learning_rate: 3.0000e-04
Epoch 19/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.7801 - dist_relevant_mae: 1.9373 - dist_relevant_mse: 8.6066 - loss: 0.7925 - prob_kld: 0.0288 - val_dist_dist_iou_metric: 0.7096 - val_dist_relevant_mae: 3.3398 - val_dist_relevant_mse: 25.5800 - val_loss: 1.2308 - val_prob_kld: 0.0477 - learning_rate: 3.0000e-04
Epoch 20/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.7898 - dist_relevant_mae: 1.7966 - dist_relevant_mse: 7.5733 - loss: 0.7397 - prob_kld: 0.0243 - val_dist_dist_iou_metric: 0.6848 - val_dist_relevant_mae: 3.5183 - val_dist_relevant_mse: 27.7450 - val_loss: 1.2605 - val_prob_kld: 0.0417 - learning_rate: 3.0000e-04
Epoch 21/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8007 - dist_relevant_mae: 1.7248 - dist_relevant_mse: 6.8640 - loss: 0.7416 - prob_kld: 0.0229 - val_dist_dist_iou_metric: 0.7164 - val_dist_relevant_mae: 3.1955 - val_dist_relevant_mse: 24.6181 - val_loss: 1.1974 - val_prob_kld: 0.0431 - learning_rate: 3.0000e-04
Epoch 22/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.7981 - dist_relevant_mae: 1.7651 - dist_relevant_mse: 7.1687 - loss: 0.7431 - prob_kld: 0.0231 - val_dist_dist_iou_metric: 0.7391 - val_dist_relevant_mae: 2.8626 - val_dist_relevant_mse: 19.9051 - val_loss: 1.1292 - val_prob_kld: 0.0416 - learning_rate: 3.0000e-04
Epoch 23/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8095 - dist_relevant_mae: 1.6556 - dist_relevant_mse: 6.4781 - loss: 0.7112 - prob_kld: 0.0211 - val_dist_dist_iou_metric: 0.7070 - val_dist_relevant_mae: 3.3597 - val_dist_relevant_mse: 24.4202 - val_loss: 1.2262 - val_prob_kld: 0.0392 - learning_rate: 3.0000e-04
Epoch 24/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8073 - dist_relevant_mae: 1.6549 - dist_relevant_mse: 6.3752 - loss: 0.7193 - prob_kld: 0.0260 - val_dist_dist_iou_metric: 0.7331 - val_dist_relevant_mae: 2.9912 - val_dist_relevant_mse: 21.5523 - val_loss: 1.1533 - val_prob_kld: 0.0400 - learning_rate: 3.0000e-04
Epoch 25/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8192 - dist_relevant_mae: 1.5673 - dist_relevant_mse: 5.7448 - loss: 0.7052 - prob_kld: 0.0198 - val_dist_dist_iou_metric: 0.7158 - val_dist_relevant_mae: 3.0957 - val_dist_relevant_mse: 21.2644 - val_loss: 1.1705 - val_prob_kld: 0.0363 - learning_rate: 3.0000e-04
Epoch 26/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8181 - dist_relevant_mae: 1.5314 - dist_relevant_mse: 5.3555 - loss: 0.6812 - prob_kld: 0.0183 - val_dist_dist_iou_metric: 0.7675 - val_dist_relevant_mae: 2.6695 - val_dist_relevant_mse: 16.6302 - val_loss: 1.0867 - val_prob_kld: 0.0376 - learning_rate: 3.0000e-04
Epoch 27/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8267 - dist_relevant_mae: 1.4721 - dist_relevant_mse: 5.0032 - loss: 0.6698 - prob_kld: 0.0184 - val_dist_dist_iou_metric: 0.7744 - val_dist_relevant_mae: 2.5149 - val_dist_relevant_mse: 15.3713 - val_loss: 1.0627 - val_prob_kld: 0.0446 - learning_rate: 3.0000e-04
Epoch 28/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8323 - dist_relevant_mae: 1.4565 - dist_relevant_mse: 5.0432 - loss: 0.6730 - prob_kld: 0.0177 - val_dist_dist_iou_metric: 0.7882 - val_dist_relevant_mae: 2.3710 - val_dist_relevant_mse: 13.5670 - val_loss: 1.0208 - val_prob_kld: 0.0315 - learning_rate: 3.0000e-04
Epoch 29/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8364 - dist_relevant_mae: 1.4022 - dist_relevant_mse: 4.5580 - loss: 0.6603 - prob_kld: 0.0160 - val_dist_dist_iou_metric: 0.7743 - val_dist_relevant_mae: 2.5053 - val_dist_relevant_mse: 14.8030 - val_loss: 1.0629 - val_prob_kld: 0.0468 - learning_rate: 3.0000e-04
Epoch 30/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8319 - dist_relevant_mae: 1.4162 - dist_relevant_mse: 4.6274 - loss: 0.6602 - prob_kld: 0.0172 - val_dist_dist_iou_metric: 0.7847 - val_dist_relevant_mae: 2.4274 - val_dist_relevant_mse: 14.2825 - val_loss: 1.0331 - val_prob_kld: 0.0326 - learning_rate: 3.0000e-04
Epoch 31/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8411 - dist_relevant_mae: 1.3399 - dist_relevant_mse: 4.1295 - loss: 0.6408 - prob_kld: 0.0154 - val_dist_dist_iou_metric: 0.7835 - val_dist_relevant_mae: 2.3912 - val_dist_relevant_mse: 13.5362 - val_loss: 1.0260 - val_prob_kld: 0.0326 - learning_rate: 3.0000e-04
Epoch 32/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8368 - dist_relevant_mae: 1.3945 - dist_relevant_mse: 4.4654 - loss: 0.6580 - prob_kld: 0.0160 - val_dist_dist_iou_metric: 0.7885 - val_dist_relevant_mae: 2.3641 - val_dist_relevant_mse: 13.0076 - val_loss: 1.0199 - val_prob_kld: 0.0320 - learning_rate: 3.0000e-04
Epoch 33/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8495 - dist_relevant_mae: 1.2786 - dist_relevant_mse: 3.8238 - loss: 0.6437 - prob_kld: 0.0148 - val_dist_dist_iou_metric: 0.7986 - val_dist_relevant_mae: 2.2493 - val_dist_relevant_mse: 12.6429 - val_loss: 0.9920 - val_prob_kld: 0.0271 - learning_rate: 3.0000e-04
Epoch 34/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8458 - dist_relevant_mae: 1.3008 - dist_relevant_mse: 3.8937 - loss: 0.6299 - prob_kld: 0.0141 - val_dist_dist_iou_metric: 0.7972 - val_dist_relevant_mae: 2.3568 - val_dist_relevant_mse: 12.2796 - val_loss: 1.0354 - val_prob_kld: 0.0489 - learning_rate: 3.0000e-04
Epoch 35/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8394 - dist_relevant_mae: 1.3666 - dist_relevant_mse: 4.1944 - loss: 0.6489 - prob_kld: 0.0157 - val_dist_dist_iou_metric: 0.7661 - val_dist_relevant_mae: 2.5786 - val_dist_relevant_mse: 15.6307 - val_loss: 1.0624 - val_prob_kld: 0.0316 - learning_rate: 3.0000e-04
Epoch 36/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8477 - dist_relevant_mae: 1.2928 - dist_relevant_mse: 3.8171 - loss: 0.6218 - prob_kld: 0.0136 - val_dist_dist_iou_metric: 0.7675 - val_dist_relevant_mae: 2.4795 - val_dist_relevant_mse: 15.8205 - val_loss: 1.0362 - val_prob_kld: 0.0252 - learning_rate: 3.0000e-04
Epoch 37/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8545 - dist_relevant_mae: 1.2288 - dist_relevant_mse: 3.5292 - loss: 0.6226 - prob_kld: 0.0135 - val_dist_dist_iou_metric: 0.7708 - val_dist_relevant_mae: 2.5063 - val_dist_relevant_mse: 14.2598 - val_loss: 1.0464 - val_prob_kld: 0.0300 - learning_rate: 3.0000e-04
Epoch 38/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8560 - dist_relevant_mae: 1.2240 - dist_relevant_mse: 3.5227 - loss: 0.6250 - prob_kld: 0.0136 - val_dist_dist_iou_metric: 0.7957 - val_dist_relevant_mae: 2.2150 - val_dist_relevant_mse: 12.2545 - val_loss: 0.9840 - val_prob_kld: 0.0259 - learning_rate: 3.0000e-04
Epoch 39/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8529 - dist_relevant_mae: 1.2353 - dist_relevant_mse: 3.4448 - loss: 0.6201 - prob_kld: 0.0131 - val_dist_dist_iou_metric: 0.7810 - val_dist_relevant_mae: 2.3675 - val_dist_relevant_mse: 13.2876 - val_loss: 1.0194 - val_prob_kld: 0.0308 - learning_rate: 3.0000e-04
Epoch 40/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8589 - dist_relevant_mae: 1.1793 - dist_relevant_mse: 3.2282 - loss: 0.5973 - prob_kld: 0.0117 - val_dist_dist_iou_metric: 0.8206 - val_dist_relevant_mae: 2.0679 - val_dist_relevant_mse: 10.1906 - val_loss: 0.9656 - val_prob_kld: 0.0369 - learning_rate: 3.0000e-04
Epoch 41/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8558 - dist_relevant_mae: 1.2185 - dist_relevant_mse: 3.4173 - loss: 0.6251 - prob_kld: 0.0143 - val_dist_dist_iou_metric: 0.8050 - val_dist_relevant_mae: 2.1220 - val_dist_relevant_mse: 11.2767 - val_loss: 0.9669 - val_prob_kld: 0.0274 - learning_rate: 3.0000e-04
Epoch 42/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8543 - dist_relevant_mae: 1.2264 - dist_relevant_mse: 3.5027 - loss: 0.6326 - prob_kld: 0.0138 - val_dist_dist_iou_metric: 0.7958 - val_dist_relevant_mae: 2.2186 - val_dist_relevant_mse: 11.9002 - val_loss: 0.9869 - val_prob_kld: 0.0281 - learning_rate: 3.0000e-04
Epoch 43/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8388 - dist_relevant_mae: 1.3802 - dist_relevant_mse: 4.3860 - loss: 0.6590 - prob_kld: 0.0262 - val_dist_dist_iou_metric: 0.7985 - val_dist_relevant_mae: 2.1434 - val_dist_relevant_mse: 11.7529 - val_loss: 0.9767 - val_prob_kld: 0.0329 - learning_rate: 3.0000e-04
Epoch 44/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8618 - dist_relevant_mae: 1.1595 - dist_relevant_mse: 3.1363 - loss: 0.5917 - prob_kld: 0.0118 - val_dist_dist_iou_metric: 0.7808 - val_dist_relevant_mae: 2.3908 - val_dist_relevant_mse: 12.4723 - val_loss: 1.0204 - val_prob_kld: 0.0271 - learning_rate: 3.0000e-04
Epoch 45/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8633 - dist_relevant_mae: 1.1251 - dist_relevant_mse: 2.9773 - loss: 0.6039 - prob_kld: 0.0121 - val_dist_dist_iou_metric: 0.7835 - val_dist_relevant_mae: 2.3681 - val_dist_relevant_mse: 13.0038 - val_loss: 1.0147 - val_prob_kld: 0.0260 - learning_rate: 3.0000e-04
Epoch 46/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8616 - dist_relevant_mae: 1.1706 - dist_relevant_mse: 3.1789 - loss: 0.6027 - prob_kld: 0.0112 - val_dist_dist_iou_metric: 0.8068 - val_dist_relevant_mae: 2.1191 - val_dist_relevant_mse: 11.0040 - val_loss: 0.9635 - val_prob_kld: 0.0247 - learning_rate: 3.0000e-04
Epoch 47/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8666 - dist_relevant_mae: 1.0893 - dist_relevant_mse: 2.7429 - loss: 0.5905 - prob_kld: 0.0109 - val_dist_dist_iou_metric: 0.8002 - val_dist_relevant_mae: 2.1870 - val_dist_relevant_mse: 11.6650 - val_loss: 0.9779 - val_prob_kld: 0.0254 - learning_rate: 3.0000e-04
Epoch 48/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8663 - dist_relevant_mae: 1.1136 - dist_relevant_mse: 2.9507 - loss: 0.5978 - prob_kld: 0.0113 - val_dist_dist_iou_metric: 0.7925 - val_dist_relevant_mae: 2.2243 - val_dist_relevant_mse: 11.8270 - val_loss: 0.9892 - val_prob_kld: 0.0292 - learning_rate: 3.0000e-04
Epoch 49/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8699 - dist_relevant_mae: 1.0887 - dist_relevant_mse: 2.7067 - loss: 0.5933 - prob_kld: 0.0120 - val_dist_dist_iou_metric: 0.8143 - val_dist_relevant_mae: 2.2045 - val_dist_relevant_mse: 10.7234 - val_loss: 0.9857 - val_prob_kld: 0.0297 - learning_rate: 3.0000e-04
Epoch 50/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8529 - dist_relevant_mae: 1.2311 - dist_relevant_mse: 3.3261 - loss: 0.6218 - prob_kld: 0.0113 - val_dist_dist_iou_metric: 0.7772 - val_dist_relevant_mae: 2.4420 - val_dist_relevant_mse: 12.7571 - val_loss: 1.0289 - val_prob_kld: 0.0255 - learning_rate: 3.0000e-04
Epoch 51/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8690 - dist_relevant_mae: 1.0914 - dist_relevant_mse: 2.7604 - loss: 0.5835 - prob_kld: 0.0101 - val_dist_dist_iou_metric: 0.7885 - val_dist_relevant_mae: 2.2663 - val_dist_relevant_mse: 11.7542 - val_loss: 0.9929 - val_prob_kld: 0.0246 - learning_rate: 3.0000e-04
Epoch 52/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8759 - dist_relevant_mae: 1.0343 - dist_relevant_mse: 2.5800 - loss: 0.5739 - prob_kld: 0.0100 - val_dist_dist_iou_metric: 0.8180 - val_dist_relevant_mae: 2.1339 - val_dist_relevant_mse: 9.8030 - val_loss: 0.9774 - val_prob_kld: 0.0355 - learning_rate: 3.0000e-04
Epoch 53/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8676 - dist_relevant_mae: 1.1073 - dist_relevant_mse: 2.8418 - loss: 0.5953 - prob_kld: 0.0109 - val_dist_dist_iou_metric: 0.7881 - val_dist_relevant_mae: 2.2668 - val_dist_relevant_mse: 12.8638 - val_loss: 0.9956 - val_prob_kld: 0.0271 - learning_rate: 3.0000e-04
Epoch 54/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8732 - dist_relevant_mae: 1.0599 - dist_relevant_mse: 2.6168 - loss: 0.5956 - prob_kld: 0.0100 - val_dist_dist_iou_metric: 0.8326 - val_dist_relevant_mae: 1.8183 - val_dist_relevant_mse: 8.5732 - val_loss: 0.9008 - val_prob_kld: 0.0221 - learning_rate: 3.0000e-04
Epoch 55/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8786 - dist_relevant_mae: 1.0114 - dist_relevant_mse: 2.4010 - loss: 0.5791 - prob_kld: 0.0092 - val_dist_dist_iou_metric: 0.8114 - val_dist_relevant_mae: 2.0156 - val_dist_relevant_mse: 9.9104 - val_loss: 0.9453 - val_prob_kld: 0.0271 - learning_rate: 3.0000e-04
Epoch 56/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8748 - dist_relevant_mae: 1.0225 - dist_relevant_mse: 2.4288 - loss: 0.5714 - prob_kld: 0.0098 - val_dist_dist_iou_metric: 0.8224 - val_dist_relevant_mae: 1.9680 - val_dist_relevant_mse: 9.1461 - val_loss: 0.9304 - val_prob_kld: 0.0218 - learning_rate: 3.0000e-04
Epoch 57/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8746 - dist_relevant_mae: 1.0545 - dist_relevant_mse: 2.5706 - loss: 0.5883 - prob_kld: 0.0095 - val_dist_dist_iou_metric: 0.8035 - val_dist_relevant_mae: 2.1848 - val_dist_relevant_mse: 10.6596 - val_loss: 0.9759 - val_prob_kld: 0.0239 - learning_rate: 3.0000e-04
Epoch 58/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8733 - dist_relevant_mae: 1.0657 - dist_relevant_mse: 2.6244 - loss: 0.5921 - prob_kld: 0.0097 - val_dist_dist_iou_metric: 0.8045 - val_dist_relevant_mae: 2.1040 - val_dist_relevant_mse: 11.0023 - val_loss: 0.9559 - val_prob_kld: 0.0200 - learning_rate: 3.0000e-04
Epoch 59/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8676 - dist_relevant_mae: 1.1081 - dist_relevant_mse: 2.7716 - loss: 0.5904 - prob_kld: 0.0095 - val_dist_dist_iou_metric: 0.8061 - val_dist_relevant_mae: 2.0995 - val_dist_relevant_mse: 11.2092 - val_loss: 0.9547 - val_prob_kld: 0.0197 - learning_rate: 3.0000e-04
Epoch 60/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8822 - dist_relevant_mae: 0.9854 - dist_relevant_mse: 2.3185 - loss: 0.5670 - prob_kld: 0.0091 - val_dist_dist_iou_metric: 0.8149 - val_dist_relevant_mae: 2.0104 - val_dist_relevant_mse: 10.1264 - val_loss: 0.9403 - val_prob_kld: 0.0231 - learning_rate: 3.0000e-04
Epoch 61/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8828 - dist_relevant_mae: 0.9730 - dist_relevant_mse: 2.2271 - loss: 0.5711 - prob_kld: 0.0088 - val_dist_dist_iou_metric: 0.8069 - val_dist_relevant_mae: 2.0837 - val_dist_relevant_mse: 10.4478 - val_loss: 0.9594 - val_prob_kld: 0.0276 - learning_rate: 3.0000e-04
Epoch 62/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8829 - dist_relevant_mae: 0.9723 - dist_relevant_mse: 2.2452 - loss: 0.5630 - prob_kld: 0.0089 - val_dist_dist_iou_metric: 0.8069 - val_dist_relevant_mae: 2.0883 - val_dist_relevant_mse: 10.4899 - val_loss: 0.9558 - val_prob_kld: 0.0231 - learning_rate: 3.0000e-04
Epoch 63/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8845 - dist_relevant_mae: 0.9495 - dist_relevant_mse: 2.1449 - loss: 0.5746 - prob_kld: 0.0089 - val_dist_dist_iou_metric: 0.8107 - val_dist_relevant_mae: 2.0741 - val_dist_relevant_mse: 10.2863 - val_loss: 0.9540 - val_prob_kld: 0.0241 - learning_rate: 3.0000e-04
Epoch 64/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8797 - dist_relevant_mae: 1.0048 - dist_relevant_mse: 2.3087 - loss: 0.5739 - prob_kld: 0.0087 - val_dist_dist_iou_metric: 0.8203 - val_dist_relevant_mae: 1.9670 - val_dist_relevant_mse: 9.4934 - val_loss: 0.9316 - val_prob_kld: 0.0232 - learning_rate: 3.0000e-04
Epoch 65/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8756 - dist_relevant_mae: 1.0579 - dist_relevant_mse: 2.5431 - loss: 0.5923 - prob_kld: 0.0088 - val_dist_dist_iou_metric: 0.8184 - val_dist_relevant_mae: 1.9719 - val_dist_relevant_mse: 9.6841 - val_loss: 0.9296 - val_prob_kld: 0.0201 - learning_rate: 3.0000e-04
Epoch 66/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8796 - dist_relevant_mae: 1.0009 - dist_relevant_mse: 2.2916 - loss: 0.5705 - prob_kld: 0.0084 - val_dist_dist_iou_metric: 0.8049 - val_dist_relevant_mae: 2.0939 - val_dist_relevant_mse: 9.9834 - val_loss: 0.9558 - val_prob_kld: 0.0220 - learning_rate: 3.0000e-04
Epoch 67/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8810 - dist_relevant_mae: 0.9665 - dist_relevant_mse: 2.2428 - loss: 0.5478 - prob_kld: 0.0081 - val_dist_dist_iou_metric: 0.7967 - val_dist_relevant_mae: 2.1839 - val_dist_relevant_mse: 10.6883 - val_loss: 0.9734 - val_prob_kld: 0.0216 - learning_rate: 3.0000e-04
Epoch 68/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8803 - dist_relevant_mae: 1.0044 - dist_relevant_mse: 2.3272 - loss: 0.5741 - prob_kld: 0.0081 - val_dist_dist_iou_metric: 0.8186 - val_dist_relevant_mae: 1.9112 - val_dist_relevant_mse: 9.1689 - val_loss: 0.9209 - val_prob_kld: 0.0236 - learning_rate: 3.0000e-04
Epoch 69/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8854 - dist_relevant_mae: 0.9516 - dist_relevant_mse: 2.1716 - loss: 0.5593 - prob_kld: 0.0082 - val_dist_dist_iou_metric: 0.8188 - val_dist_relevant_mae: 1.9139 - val_dist_relevant_mse: 9.3104 - val_loss: 0.9259 - val_prob_kld: 0.0280 - learning_rate: 3.0000e-04
Epoch 70/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8854 - dist_relevant_mae: 0.9423 - dist_relevant_mse: 2.1136 - loss: 0.5633 - prob_kld: 0.0083 - val_dist_dist_iou_metric: 0.8239 - val_dist_relevant_mae: 1.9348 - val_dist_relevant_mse: 8.3788 - val_loss: 0.9210 - val_prob_kld: 0.0190 - learning_rate: 3.0000e-04
Epoch 71/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8866 - dist_relevant_mae: 0.9448 - dist_relevant_mse: 2.1244 - loss: 0.5574 - prob_kld: 0.0076 - val_dist_dist_iou_metric: 0.8144 - val_dist_relevant_mae: 1.9978 - val_dist_relevant_mse: 9.4325 - val_loss: 0.9366 - val_prob_kld: 0.0220 - learning_rate: 3.0000e-04
Epoch 72/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8890 - dist_relevant_mae: 0.9032 - dist_relevant_mse: 1.9515 - loss: 0.5566 - prob_kld: 0.0076 - val_dist_dist_iou_metric: 0.8327 - val_dist_relevant_mae: 1.8737 - val_dist_relevant_mse: 8.3486 - val_loss: 0.9087 - val_prob_kld: 0.0189 - learning_rate: 3.0000e-04
Epoch 73/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8912 - dist_relevant_mae: 0.9089 - dist_relevant_mse: 1.9219 - loss: 0.5566 - prob_kld: 0.0073 - val_dist_dist_iou_metric: 0.7986 - val_dist_relevant_mae: 2.2417 - val_dist_relevant_mse: 11.3714 - val_loss: 0.9892 - val_prob_kld: 0.0258 - learning_rate: 3.0000e-04
Epoch 74/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8823 - dist_relevant_mae: 0.9645 - dist_relevant_mse: 2.2244 - loss: 0.5672 - prob_kld: 0.0086 - val_dist_dist_iou_metric: 0.8407 - val_dist_relevant_mae: 1.7768 - val_dist_relevant_mse: 7.2989 - val_loss: 0.8898 - val_prob_kld: 0.0194 - learning_rate: 3.0000e-04
Epoch 75/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8856 - dist_relevant_mae: 0.9493 - dist_relevant_mse: 2.0915 - loss: 0.5580 - prob_kld: 0.0082 - val_dist_dist_iou_metric: 0.8161 - val_dist_relevant_mae: 1.9709 - val_dist_relevant_mse: 9.1692 - val_loss: 0.9291 - val_prob_kld: 0.0199 - learning_rate: 3.0000e-04
Epoch 76/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8855 - dist_relevant_mae: 0.9305 - dist_relevant_mse: 2.0474 - loss: 0.5632 - prob_kld: 0.0083 - val_dist_dist_iou_metric: 0.8261 - val_dist_relevant_mae: 1.8666 - val_dist_relevant_mse: 8.3851 - val_loss: 0.9052 - val_prob_kld: 0.0168 - learning_rate: 3.0000e-04
Epoch 77/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8894 - dist_relevant_mae: 0.9106 - dist_relevant_mse: 1.9723 - loss: 0.5540 - prob_kld: 0.0077 - val_dist_dist_iou_metric: 0.8148 - val_dist_relevant_mae: 2.0064 - val_dist_relevant_mse: 9.7343 - val_loss: 0.9370 - val_prob_kld: 0.0207 - learning_rate: 3.0000e-04
Epoch 78/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8898 - dist_relevant_mae: 0.8997 - dist_relevant_mse: 1.9436 - loss: 0.5618 - prob_kld: 0.0078 - val_dist_dist_iou_metric: 0.8210 - val_dist_relevant_mae: 1.8905 - val_dist_relevant_mse: 9.2179 - val_loss: 0.9097 - val_prob_kld: 0.0165 - learning_rate: 3.0000e-04
Epoch 79/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8910 - dist_relevant_mae: 0.8989 - dist_relevant_mse: 1.9377 - loss: 0.5635 - prob_kld: 0.0081 - val_dist_dist_iou_metric: 0.8023 - val_dist_relevant_mae: 2.0848 - val_dist_relevant_mse: 10.2506 - val_loss: 0.9504 - val_prob_kld: 0.0184 - learning_rate: 3.0000e-04
Epoch 80/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8864 - dist_relevant_mae: 0.9494 - dist_relevant_mse: 2.0874 - loss: 0.5700 - prob_kld: 0.0106 - val_dist_dist_iou_metric: 0.8269 - val_dist_relevant_mae: 1.9125 - val_dist_relevant_mse: 8.1546 - val_loss: 0.9151 - val_prob_kld: 0.0176 - learning_rate: 3.0000e-04
Epoch 81/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8909 - dist_relevant_mae: 0.9019 - dist_relevant_mse: 1.9108 - loss: 0.5591 - prob_kld: 0.0070 - val_dist_dist_iou_metric: 0.7990 - val_dist_relevant_mae: 2.1812 - val_dist_relevant_mse: 10.1248 - val_loss: 0.9696 - val_prob_kld: 0.0183 - learning_rate: 3.0000e-04
Epoch 82/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8820 - dist_relevant_mae: 0.9902 - dist_relevant_mse: 2.2183 - loss: 0.5644 - prob_kld: 0.0073 - val_dist_dist_iou_metric: 0.8188 - val_dist_relevant_mae: 1.9206 - val_dist_relevant_mse: 8.7484 - val_loss: 0.9177 - val_prob_kld: 0.0186 - learning_rate: 3.0000e-04
Epoch 83/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8931 - dist_relevant_mae: 0.8653 - dist_relevant_mse: 1.8359 - loss: 0.5407 - prob_kld: 0.0070 - val_dist_dist_iou_metric: 0.8261 - val_dist_relevant_mae: 1.8441 - val_dist_relevant_mse: 8.5659 - val_loss: 0.9049 - val_prob_kld: 0.0210 - learning_rate: 3.0000e-04
Epoch 84/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8937 - dist_relevant_mae: 0.8800 - dist_relevant_mse: 1.8426 - loss: 0.5398 - prob_kld: 0.0065 - val_dist_dist_iou_metric: 0.8256 - val_dist_relevant_mae: 1.8531 - val_dist_relevant_mse: 8.3943 - val_loss: 0.9045 - val_prob_kld: 0.0188 - learning_rate: 3.0000e-04
Epoch 85/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8924 - dist_relevant_mae: 0.8714 - dist_relevant_mse: 1.8484 - loss: 0.5582 - prob_kld: 0.0073 - val_dist_dist_iou_metric: 0.8312 - val_dist_relevant_mae: 1.8642 - val_dist_relevant_mse: 8.1085 - val_loss: 0.9109 - val_prob_kld: 0.0230 - learning_rate: 3.0000e-04
Epoch 86/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8905 - dist_relevant_mae: 0.9153 - dist_relevant_mse: 1.9277 - loss: 0.5526 - prob_kld: 0.0069 - val_dist_dist_iou_metric: 0.8293 - val_dist_relevant_mae: 1.8420 - val_dist_relevant_mse: 8.1654 - val_loss: 0.9047 - val_prob_kld: 0.0213 - learning_rate: 3.0000e-04
Epoch 87/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8855 - dist_relevant_mae: 0.9450 - dist_relevant_mse: 2.0298 - loss: 0.5551 - prob_kld: 0.0071 - val_dist_dist_iou_metric: 0.8212 - val_dist_relevant_mae: 1.9040 - val_dist_relevant_mse: 8.6055 - val_loss: 0.9182 - val_prob_kld: 0.0224 - learning_rate: 3.0000e-04
Epoch 88/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8885 - dist_relevant_mae: 0.9442 - dist_relevant_mse: 2.0478 - loss: 0.5552 - prob_kld: 0.0070 - val_dist_dist_iou_metric: 0.8300 - val_dist_relevant_mae: 1.8551 - val_dist_relevant_mse: 7.8768 - val_loss: 0.9052 - val_prob_kld: 0.0191 - learning_rate: 3.0000e-04
Epoch 89/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8960 - dist_relevant_mae: 0.8505 - dist_relevant_mse: 1.7216 - loss: 0.5520 - prob_kld: 0.0070 - val_dist_dist_iou_metric: 0.8327 - val_dist_relevant_mae: 1.7891 - val_dist_relevant_mse: 7.5541 - val_loss: 0.8897 - val_prob_kld: 0.0169 - learning_rate: 3.0000e-04
Epoch 90/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8937 - dist_relevant_mae: 0.8543 - dist_relevant_mse: 1.7233 - loss: 0.5458 - prob_kld: 0.0073 - val_dist_dist_iou_metric: 0.8185 - val_dist_relevant_mae: 1.9379 - val_dist_relevant_mse: 8.7299 - val_loss: 0.9228 - val_prob_kld: 0.0201 - learning_rate: 3.0000e-04
Epoch 91/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8932 - dist_relevant_mae: 0.8815 - dist_relevant_mse: 1.8382 - loss: 0.5455 - prob_kld: 0.0073 - val_dist_dist_iou_metric: 0.8246 - val_dist_relevant_mae: 1.8857 - val_dist_relevant_mse: 7.9738 - val_loss: 0.9097 - val_prob_kld: 0.0175 - learning_rate: 3.0000e-04
Epoch 92/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8983 - dist_relevant_mae: 0.8384 - dist_relevant_mse: 1.6902 - loss: 0.5312 - prob_kld: 0.0064 - val_dist_dist_iou_metric: 0.8260 - val_dist_relevant_mae: 1.8503 - val_dist_relevant_mse: 8.6319 - val_loss: 0.9031 - val_prob_kld: 0.0180 - learning_rate: 3.0000e-04
Epoch 93/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8964 - dist_relevant_mae: 0.8549 - dist_relevant_mse: 1.7197 - loss: 0.5545 - prob_kld: 0.0065 - val_dist_dist_iou_metric: 0.8246 - val_dist_relevant_mae: 1.8604 - val_dist_relevant_mse: 8.4963 - val_loss: 0.9028 - val_prob_kld: 0.0156 - learning_rate: 3.0000e-04
Epoch 94/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8965 - dist_relevant_mae: 0.8514 - dist_relevant_mse: 1.7799 - loss: 0.5558 - prob_kld: 0.0068 - val_dist_dist_iou_metric: 0.8298 - val_dist_relevant_mae: 1.8145 - val_dist_relevant_mse: 7.6417 - val_loss: 0.8973 - val_prob_kld: 0.0194 - learning_rate: 3.0000e-04
Epoch 95/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8971 - dist_relevant_mae: 0.8551 - dist_relevant_mse: 1.6757 - loss: 0.5461 - prob_kld: 0.0065 - val_dist_dist_iou_metric: 0.8327 - val_dist_relevant_mae: 1.7827 - val_dist_relevant_mse: 7.6561 - val_loss: 0.8900 - val_prob_kld: 0.0184 - learning_rate: 3.0000e-04
Epoch 96/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8995 - dist_relevant_mae: 0.8197 - dist_relevant_mse: 1.6229 - loss: 0.5272 - prob_kld: 0.0061 - val_dist_dist_iou_metric: 0.8417 - val_dist_relevant_mae: 1.6945 - val_dist_relevant_mse: 6.9273 - val_loss: 0.8698 - val_prob_kld: 0.0158 - learning_rate: 3.0000e-04
Epoch 97/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8974 - dist_relevant_mae: 0.8392 - dist_relevant_mse: 1.6990 - loss: 0.5325 - prob_kld: 0.0064 - val_dist_dist_iou_metric: 0.8071 - val_dist_relevant_mae: 2.0612 - val_dist_relevant_mse: 10.1184 - val_loss: 0.9476 - val_prob_kld: 0.0203 - learning_rate: 3.0000e-04
Epoch 98/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8933 - dist_relevant_mae: 0.8794 - dist_relevant_mse: 1.7812 - loss: 0.5468 - prob_kld: 0.0063 - val_dist_dist_iou_metric: 0.8248 - val_dist_relevant_mae: 1.8594 - val_dist_relevant_mse: 8.3057 - val_loss: 0.9080 - val_prob_kld: 0.0210 - learning_rate: 3.0000e-04
Epoch 99/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8972 - dist_relevant_mae: 0.8299 - dist_relevant_mse: 1.6516 - loss: 0.5324 - prob_kld: 0.0065 - val_dist_dist_iou_metric: 0.8127 - val_dist_relevant_mae: 2.0104 - val_dist_relevant_mse: 9.1333 - val_loss: 0.9341 - val_prob_kld: 0.0170 - learning_rate: 3.0000e-04
Epoch 100/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9016 - dist_relevant_mae: 0.8146 - dist_relevant_mse: 1.6210 - loss: 0.5338 - prob_kld: 0.0059 - val_dist_dist_iou_metric: 0.8075 - val_dist_relevant_mae: 2.0076 - val_dist_relevant_mse: 9.7211 - val_loss: 0.9439 - val_prob_kld: 0.0273 - learning_rate: 3.0000e-04
Epoch 101/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8932 - dist_relevant_mae: 0.8895 - dist_relevant_mse: 1.8596 - loss: 0.5390 - prob_kld: 0.0071 - val_dist_dist_iou_metric: 0.7920 - val_dist_relevant_mae: 2.1817 - val_dist_relevant_mse: 11.1619 - val_loss: 0.9678 - val_prob_kld: 0.0165 - learning_rate: 3.0000e-04
Epoch 102/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8934 - dist_relevant_mae: 0.8673 - dist_relevant_mse: 1.7496 - loss: 0.5441 - prob_kld: 0.0064 - val_dist_dist_iou_metric: 0.8018 - val_dist_relevant_mae: 2.1007 - val_dist_relevant_mse: 10.5285 - val_loss: 0.9529 - val_prob_kld: 0.0177 - learning_rate: 3.0000e-04
Epoch 103/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8980 - dist_relevant_mae: 0.8665 - dist_relevant_mse: 1.7362 - loss: 0.5502 - prob_kld: 0.0062 - val_dist_dist_iou_metric: 0.8313 - val_dist_relevant_mae: 1.7793 - val_dist_relevant_mse: 7.8334 - val_loss: 0.8856 - val_prob_kld: 0.0147 - learning_rate: 3.0000e-04
Epoch 104/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8956 - dist_relevant_mae: 0.8587 - dist_relevant_mse: 1.7302 - loss: 0.5305 - prob_kld: 0.0072 - val_dist_dist_iou_metric: 0.8239 - val_dist_relevant_mae: 1.8606 - val_dist_relevant_mse: 8.1421 - val_loss: 0.9044 - val_prob_kld: 0.0173 - learning_rate: 3.0000e-04
Epoch 105/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8942 - dist_relevant_mae: 0.8635 - dist_relevant_mse: 1.7175 - loss: 0.5448 - prob_kld: 0.0063 - val_dist_dist_iou_metric: 0.8189 - val_dist_relevant_mae: 1.9153 - val_dist_relevant_mse: 8.3274 - val_loss: 0.9139 - val_prob_kld: 0.0158 - learning_rate: 3.0000e-04
Epoch 106/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8991 - dist_relevant_mae: 0.8544 - dist_relevant_mse: 1.6972 - loss: 0.5358 - prob_kld: 0.0058 - val_dist_dist_iou_metric: 0.8393 - val_dist_relevant_mae: 1.6839 - val_dist_relevant_mse: 7.1026 - val_loss: 0.8660 - val_prob_kld: 0.0142 - learning_rate: 3.0000e-04
Epoch 107/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9016 - dist_relevant_mae: 0.8036 - dist_relevant_mse: 1.5630 - loss: 0.5285 - prob_kld: 0.0059 - val_dist_dist_iou_metric: 0.8498 - val_dist_relevant_mae: 1.6182 - val_dist_relevant_mse: 6.3319 - val_loss: 0.8575 - val_prob_kld: 0.0188 - learning_rate: 3.0000e-04
Epoch 108/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9023 - dist_relevant_mae: 0.8042 - dist_relevant_mse: 1.5461 - loss: 0.5324 - prob_kld: 0.0056 - val_dist_dist_iou_metric: 0.8209 - val_dist_relevant_mae: 1.8843 - val_dist_relevant_mse: 8.1534 - val_loss: 0.9066 - val_prob_kld: 0.0146 - learning_rate: 3.0000e-04
Epoch 109/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9016 - dist_relevant_mae: 0.8048 - dist_relevant_mse: 1.5494 - loss: 0.5346 - prob_kld: 0.0060 - val_dist_dist_iou_metric: 0.8406 - val_dist_relevant_mae: 1.6932 - val_dist_relevant_mse: 6.8232 - val_loss: 0.8729 - val_prob_kld: 0.0192 - learning_rate: 3.0000e-04
Epoch 110/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8869 - dist_relevant_mae: 0.9458 - dist_relevant_mse: 1.9648 - loss: 0.5704 - prob_kld: 0.0070 - val_dist_dist_iou_metric: 0.8109 - val_dist_relevant_mae: 1.9647 - val_dist_relevant_mse: 9.5583 - val_loss: 0.9229 - val_prob_kld: 0.0149 - learning_rate: 3.0000e-04
Epoch 111/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8962 - dist_relevant_mae: 0.8415 - dist_relevant_mse: 1.7203 - loss: 0.5286 - prob_kld: 0.0064 - val_dist_dist_iou_metric: 0.8308 - val_dist_relevant_mae: 1.7700 - val_dist_relevant_mse: 7.9971 - val_loss: 0.8844 - val_prob_kld: 0.0153 - learning_rate: 3.0000e-04
Epoch 112/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9016 - dist_relevant_mae: 0.7851 - dist_relevant_mse: 1.4677 - loss: 0.5041 - prob_kld: 0.0054 - val_dist_dist_iou_metric: 0.8262 - val_dist_relevant_mae: 1.8407 - val_dist_relevant_mse: 7.8685 - val_loss: 0.9018 - val_prob_kld: 0.0186 - learning_rate: 3.0000e-04
Epoch 113/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9005 - dist_relevant_mae: 0.8244 - dist_relevant_mse: 1.6404 - loss: 0.5554 - prob_kld: 0.0063 - val_dist_dist_iou_metric: 0.8417 - val_dist_relevant_mae: 1.6555 - val_dist_relevant_mse: 6.8548 - val_loss: 0.8605 - val_prob_kld: 0.0143 - learning_rate: 3.0000e-04
Epoch 114/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9026 - dist_relevant_mae: 0.7937 - dist_relevant_mse: 1.5428 - loss: 0.5419 - prob_kld: 0.0061 - val_dist_dist_iou_metric: 0.8278 - val_dist_relevant_mae: 1.8374 - val_dist_relevant_mse: 8.0890 - val_loss: 0.9060 - val_prob_kld: 0.0235 - learning_rate: 3.0000e-04
Epoch 115/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9077 - dist_relevant_mae: 0.7548 - dist_relevant_mse: 1.3914 - loss: 0.5194 - prob_kld: 0.0056 - val_dist_dist_iou_metric: 0.8228 - val_dist_relevant_mae: 1.8597 - val_dist_relevant_mse: 8.5917 - val_loss: 0.9025 - val_prob_kld: 0.0155 - learning_rate: 3.0000e-04
Epoch 116/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9048 - dist_relevant_mae: 0.7924 - dist_relevant_mse: 1.5326 - loss: 0.5263 - prob_kld: 0.0055 - val_dist_dist_iou_metric: 0.8346 - val_dist_relevant_mae: 1.7310 - val_dist_relevant_mse: 7.5259 - val_loss: 0.8769 - val_prob_kld: 0.0156 - learning_rate: 3.0000e-04
Epoch 117/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9053 - dist_relevant_mae: 0.7856 - dist_relevant_mse: 1.4889 - loss: 0.5292 - prob_kld: 0.0056 - val_dist_dist_iou_metric: 0.8281 - val_dist_relevant_mae: 1.7826 - val_dist_relevant_mse: 7.8789 - val_loss: 0.8864 - val_prob_kld: 0.0148 - learning_rate: 3.0000e-04
Epoch 118/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9030 - dist_relevant_mae: 0.7949 - dist_relevant_mse: 1.5215 - loss: 0.5211 - prob_kld: 0.0057 - val_dist_dist_iou_metric: 0.8269 - val_dist_relevant_mae: 1.8440 - val_dist_relevant_mse: 7.7509 - val_loss: 0.8988 - val_prob_kld: 0.0150 - learning_rate: 3.0000e-04
Epoch 119/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9022 - dist_relevant_mae: 0.7963 - dist_relevant_mse: 1.4768 - loss: 0.5190 - prob_kld: 0.0068 - val_dist_dist_iou_metric: 0.8257 - val_dist_relevant_mae: 1.8665 - val_dist_relevant_mse: 7.9502 - val_loss: 0.9054 - val_prob_kld: 0.0170 - learning_rate: 3.0000e-04
Epoch 120/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9025 - dist_relevant_mae: 0.7996 - dist_relevant_mse: 1.5685 - loss: 0.5181 - prob_kld: 0.0055 - val_dist_dist_iou_metric: 0.8475 - val_dist_relevant_mae: 1.6441 - val_dist_relevant_mse: 6.5625 - val_loss: 0.8617 - val_prob_kld: 0.0178 - learning_rate: 3.0000e-04
Epoch 121/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.8973 - dist_relevant_mae: 0.8456 - dist_relevant_mse: 1.6618 - loss: 0.5417 - prob_kld: 0.0063 - val_dist_dist_iou_metric: 0.8180 - val_dist_relevant_mae: 1.9223 - val_dist_relevant_mse: 8.4273 - val_loss: 0.9153 - val_prob_kld: 0.0158 - learning_rate: 3.0000e-04
Epoch 122/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9001 - dist_relevant_mae: 0.8152 - dist_relevant_mse: 1.5780 - loss: 0.5357 - prob_kld: 0.0060 - val_dist_dist_iou_metric: 0.8243 - val_dist_relevant_mae: 1.8206 - val_dist_relevant_mse: 8.0909 - val_loss: 0.8938 - val_prob_kld: 0.0146 - learning_rate: 3.0000e-04
Epoch 123/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9035 - dist_relevant_mae: 0.7761 - dist_relevant_mse: 1.4446 - loss: 0.5263 - prob_kld: 0.0058 - val_dist_dist_iou_metric: 0.8399 - val_dist_relevant_mae: 1.6580 - val_dist_relevant_mse: 6.8095 - val_loss: 0.8615 - val_prob_kld: 0.0149 - learning_rate: 3.0000e-04
Epoch 124/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9037 - dist_relevant_mae: 0.7802 - dist_relevant_mse: 1.4638 - loss: 0.5141 - prob_kld: 0.0054 - val_dist_dist_iou_metric: 0.8100 - val_dist_relevant_mae: 2.0290 - val_dist_relevant_mse: 9.5887 - val_loss: 0.9393 - val_prob_kld: 0.0184 - learning_rate: 3.0000e-04
Epoch 125/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9026 - dist_relevant_mae: 0.7935 - dist_relevant_mse: 1.5271 - loss: 0.5266 - prob_kld: 0.0056 - val_dist_dist_iou_metric: 0.8422 - val_dist_relevant_mae: 1.6611 - val_dist_relevant_mse: 6.3502 - val_loss: 0.8600 - val_prob_kld: 0.0127 - learning_rate: 3.0000e-04
Epoch 126/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9066 - dist_relevant_mae: 0.7712 - dist_relevant_mse: 1.3887 - loss: 0.5274 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8235 - val_dist_relevant_mae: 1.8572 - val_dist_relevant_mse: 8.2035 - val_loss: 0.9016 - val_prob_kld: 0.0151 - learning_rate: 3.0000e-04
Epoch 127/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9054 - dist_relevant_mae: 0.7729 - dist_relevant_mse: 1.4028 - loss: 0.5301 - prob_kld: 0.0053 - val_dist_dist_iou_metric: 0.8036 - val_dist_relevant_mae: 2.0771 - val_dist_relevant_mse: 9.8563 - val_loss: 0.9455 - val_prob_kld: 0.0150 - learning_rate: 3.0000e-04
Epoch 128/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9058 - dist_relevant_mae: 0.7808 - dist_relevant_mse: 1.4484 - loss: 0.5165 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8423 - val_dist_relevant_mae: 1.7422 - val_dist_relevant_mse: 6.6275 - val_loss: 0.8815 - val_prob_kld: 0.0180 - learning_rate: 3.0000e-04
Epoch 129/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.8988 - dist_relevant_mae: 0.8329 - dist_relevant_mse: 1.6011 - loss: 0.5345 - prob_kld: 0.0060 - val_dist_dist_iou_metric: 0.8025 - val_dist_relevant_mae: 2.1051 - val_dist_relevant_mse: 10.0849 - val_loss: 0.9539 - val_prob_kld: 0.0178 - learning_rate: 3.0000e-04
Epoch 130/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9062 - dist_relevant_mae: 0.7707 - dist_relevant_mse: 1.4069 - loss: 0.5351 - prob_kld: 0.0055 - val_dist_dist_iou_metric: 0.8217 - val_dist_relevant_mae: 1.8289 - val_dist_relevant_mse: 8.2983 - val_loss: 0.8979 - val_prob_kld: 0.0170 - learning_rate: 3.0000e-04
Epoch 131/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9059 - dist_relevant_mae: 0.7978 - dist_relevant_mse: 1.5137 - loss: 0.5342 - prob_kld: 0.0052 - val_dist_dist_iou_metric: 0.8379 - val_dist_relevant_mae: 1.6944 - val_dist_relevant_mse: 6.6667 - val_loss: 0.8681 - val_prob_kld: 0.0142 - learning_rate: 3.0000e-04
Epoch 132/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9081 - dist_relevant_mae: 0.7517 - dist_relevant_mse: 1.3565 - loss: 0.5416 - prob_kld: 0.0054 - val_dist_dist_iou_metric: 0.8439 - val_dist_relevant_mae: 1.6638 - val_dist_relevant_mse: 6.7406 - val_loss: 0.8632 - val_prob_kld: 0.0154 - learning_rate: 3.0000e-04
Epoch 133/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9089 - dist_relevant_mae: 0.7439 - dist_relevant_mse: 1.3702 - loss: 0.5081 - prob_kld: 0.0048 - val_dist_dist_iou_metric: 0.8378 - val_dist_relevant_mae: 1.6922 - val_dist_relevant_mse: 6.9616 - val_loss: 0.8727 - val_prob_kld: 0.0192 - learning_rate: 3.0000e-04
Epoch 134/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9057 - dist_relevant_mae: 0.7798 - dist_relevant_mse: 1.4471 - loss: 0.5270 - prob_kld: 0.0054 - val_dist_dist_iou_metric: 0.8426 - val_dist_relevant_mae: 1.6526 - val_dist_relevant_mse: 6.6876 - val_loss: 0.8616 - val_prob_kld: 0.0160 - learning_rate: 3.0000e-04
Epoch 135/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9057 - dist_relevant_mae: 0.7829 - dist_relevant_mse: 1.4706 - loss: 0.5356 - prob_kld: 0.0053 - val_dist_dist_iou_metric: 0.8436 - val_dist_relevant_mae: 1.6229 - val_dist_relevant_mse: 6.6097 - val_loss: 0.8564 - val_prob_kld: 0.0167 - learning_rate: 3.0000e-04
Epoch 136/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9025 - dist_relevant_mae: 0.7915 - dist_relevant_mse: 1.4715 - loss: 0.5241 - prob_kld: 0.0052 - val_dist_dist_iou_metric: 0.8322 - val_dist_relevant_mae: 1.7903 - val_dist_relevant_mse: 7.4200 - val_loss: 0.8913 - val_prob_kld: 0.0181 - learning_rate: 3.0000e-04
Epoch 137/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9088 - dist_relevant_mae: 0.7472 - dist_relevant_mse: 1.3644 - loss: 0.5231 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8227 - val_dist_relevant_mae: 1.8758 - val_dist_relevant_mse: 8.0480 - val_loss: 0.9063 - val_prob_kld: 0.0161 - learning_rate: 3.0000e-04
Epoch 138/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9056 - dist_relevant_mae: 0.7665 - dist_relevant_mse: 1.4244 - loss: 0.5177 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8226 - val_dist_relevant_mae: 1.8885 - val_dist_relevant_mse: 8.1464 - val_loss: 0.9081 - val_prob_kld: 0.0153 - learning_rate: 3.0000e-04
Epoch 139/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9101 - dist_relevant_mae: 0.7312 - dist_relevant_mse: 1.3242 - loss: 0.5275 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8423 - val_dist_relevant_mae: 1.6518 - val_dist_relevant_mse: 6.9242 - val_loss: 0.8613 - val_prob_kld: 0.0159 - learning_rate: 3.0000e-04
Epoch 140/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9118 - dist_relevant_mae: 0.7200 - dist_relevant_mse: 1.3022 - loss: 0.5144 - prob_kld: 0.0047 - val_dist_dist_iou_metric: 0.8202 - val_dist_relevant_mae: 1.8739 - val_dist_relevant_mse: 8.4696 - val_loss: 0.9081 - val_prob_kld: 0.0183 - learning_rate: 3.0000e-04
Epoch 141/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9094 - dist_relevant_mae: 0.7491 - dist_relevant_mse: 1.3265 - loss: 0.5256 - prob_kld: 0.0051 - val_dist_dist_iou_metric: 0.8528 - val_dist_relevant_mae: 1.5574 - val_dist_relevant_mse: 5.9991 - val_loss: 0.8429 - val_prob_kld: 0.0164 - learning_rate: 3.0000e-04
Epoch 142/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9032 - dist_relevant_mae: 0.8021 - dist_relevant_mse: 1.4825 - loss: 0.5275 - prob_kld: 0.0052 - val_dist_dist_iou_metric: 0.8401 - val_dist_relevant_mae: 1.6982 - val_dist_relevant_mse: 6.6420 - val_loss: 0.8701 - val_prob_kld: 0.0154 - learning_rate: 3.0000e-04
Epoch 143/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9097 - dist_relevant_mae: 0.7301 - dist_relevant_mse: 1.3134 - loss: 0.5222 - prob_kld: 0.0049 - val_dist_dist_iou_metric: 0.8444 - val_dist_relevant_mae: 1.6647 - val_dist_relevant_mse: 6.3154 - val_loss: 0.8626 - val_prob_kld: 0.0146 - learning_rate: 3.0000e-04
Epoch 144/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9059 - dist_relevant_mae: 0.7689 - dist_relevant_mse: 1.3743 - loss: 0.5238 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8342 - val_dist_relevant_mae: 1.7684 - val_dist_relevant_mse: 7.0921 - val_loss: 0.8871 - val_prob_kld: 0.0184 - learning_rate: 3.0000e-04
Epoch 145/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9094 - dist_relevant_mae: 0.7517 - dist_relevant_mse: 1.3631 - loss: 0.5044 - prob_kld: 0.0047 - val_dist_dist_iou_metric: 0.8465 - val_dist_relevant_mae: 1.6301 - val_dist_relevant_mse: 6.3173 - val_loss: 0.8558 - val_prob_kld: 0.0147 - learning_rate: 3.0000e-04
Epoch 146/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9090 - dist_relevant_mae: 0.7451 - dist_relevant_mse: 1.3310 - loss: 0.5192 - prob_kld: 0.0049 - val_dist_dist_iou_metric: 0.8204 - val_dist_relevant_mae: 1.9068 - val_dist_relevant_mse: 7.9470 - val_loss: 0.9112 - val_prob_kld: 0.0148 - learning_rate: 3.0000e-04
Epoch 147/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9082 - dist_relevant_mae: 0.7574 - dist_relevant_mse: 1.3719 - loss: 0.5116 - prob_kld: 0.0047 - val_dist_dist_iou_metric: 0.8330 - val_dist_relevant_mae: 1.7783 - val_dist_relevant_mse: 6.9327 - val_loss: 0.8862 - val_prob_kld: 0.0155 - learning_rate: 3.0000e-04
Epoch 148/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9084 - dist_relevant_mae: 0.7404 - dist_relevant_mse: 1.3145 - loss: 0.5183 - prob_kld: 0.0051 - val_dist_dist_iou_metric: 0.8445 - val_dist_relevant_mae: 1.6114 - val_dist_relevant_mse: 6.2765 - val_loss: 0.8521 - val_prob_kld: 0.0147 - learning_rate: 3.0000e-04
Epoch 149/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9087 - dist_relevant_mae: 0.7456 - dist_relevant_mse: 1.3370 - loss: 0.5170 - prob_kld: 0.0051 - val_dist_dist_iou_metric: 0.8161 - val_dist_relevant_mae: 1.9218 - val_dist_relevant_mse: 8.4209 - val_loss: 0.9307 - val_prob_kld: 0.0313 - learning_rate: 3.0000e-04
Epoch 150/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9104 - dist_relevant_mae: 0.7368 - dist_relevant_mse: 1.3154 - loss: 0.5130 - prob_kld: 0.0060 - val_dist_dist_iou_metric: 0.8166 - val_dist_relevant_mae: 1.9235 - val_dist_relevant_mse: 8.2803 - val_loss: 0.9139 - val_prob_kld: 0.0142 - learning_rate: 3.0000e-04
Epoch 151/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9106 - dist_relevant_mae: 0.7322 - dist_relevant_mse: 1.3087 - loss: 0.5090 - prob_kld: 0.0046 - val_dist_dist_iou_metric: 0.8388 - val_dist_relevant_mae: 1.7108 - val_dist_relevant_mse: 6.9121 - val_loss: 0.8729 - val_prob_kld: 0.0157 - learning_rate: 3.0000e-04
Epoch 152/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9122 - dist_relevant_mae: 0.7306 - dist_relevant_mse: 1.2595 - loss: 0.5262 - prob_kld: 0.0045 - val_dist_dist_iou_metric: 0.8144 - val_dist_relevant_mae: 1.9230 - val_dist_relevant_mse: 9.0382 - val_loss: 0.9164 - val_prob_kld: 0.0167 - learning_rate: 3.0000e-04
Epoch 153/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9116 - dist_relevant_mae: 0.7209 - dist_relevant_mse: 1.2792 - loss: 0.5227 - prob_kld: 0.0049 - val_dist_dist_iou_metric: 0.8347 - val_dist_relevant_mae: 1.7179 - val_dist_relevant_mse: 7.0757 - val_loss: 0.8721 - val_prob_kld: 0.0135 - learning_rate: 3.0000e-04
Epoch 154/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9084 - dist_relevant_mae: 0.7598 - dist_relevant_mse: 1.3750 - loss: 0.5102 - prob_kld: 0.0047 - val_dist_dist_iou_metric: 0.8181 - val_dist_relevant_mae: 1.9102 - val_dist_relevant_mse: 8.2477 - val_loss: 0.9137 - val_prob_kld: 0.0166 - learning_rate: 3.0000e-04
Epoch 155/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9069 - dist_relevant_mae: 0.7636 - dist_relevant_mse: 1.4145 - loss: 0.5208 - prob_kld: 0.0049 - val_dist_dist_iou_metric: 0.8361 - val_dist_relevant_mae: 1.7307 - val_dist_relevant_mse: 6.7466 - val_loss: 0.8764 - val_prob_kld: 0.0152 - learning_rate: 3.0000e-04
Epoch 156/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9116 - dist_relevant_mae: 0.7232 - dist_relevant_mse: 1.2546 - loss: 0.5164 - prob_kld: 0.0045 - val_dist_dist_iou_metric: 0.8499 - val_dist_relevant_mae: 1.5663 - val_dist_relevant_mse: 6.2010 - val_loss: 0.8431 - val_prob_kld: 0.0148 - learning_rate: 3.0000e-04
Epoch 157/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9138 - dist_relevant_mae: 0.7091 - dist_relevant_mse: 1.2196 - loss: 0.5168 - prob_kld: 0.0044 - val_dist_dist_iou_metric: 0.8479 - val_dist_relevant_mae: 1.5486 - val_dist_relevant_mse: 6.3108 - val_loss: 0.8394 - val_prob_kld: 0.0146 - learning_rate: 3.0000e-04
Epoch 158/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9088 - dist_relevant_mae: 0.7280 - dist_relevant_mse: 1.2651 - loss: 0.5143 - prob_kld: 0.0046 - val_dist_dist_iou_metric: 0.7912 - val_dist_relevant_mae: 2.1801 - val_dist_relevant_mse: 10.9643 - val_loss: 0.9653 - val_prob_kld: 0.0142 - learning_rate: 3.0000e-04
Epoch 159/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9031 - dist_relevant_mae: 0.8026 - dist_relevant_mse: 1.4838 - loss: 0.5338 - prob_kld: 0.0049 - val_dist_dist_iou_metric: 0.8228 - val_dist_relevant_mae: 1.8448 - val_dist_relevant_mse: 7.6440 - val_loss: 0.8977 - val_prob_kld: 0.0137 - learning_rate: 3.0000e-04
Epoch 160/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9125 - dist_relevant_mae: 0.7216 - dist_relevant_mse: 1.2137 - loss: 0.5026 - prob_kld: 0.0045 - val_dist_dist_iou_metric: 0.8280 - val_dist_relevant_mae: 1.7920 - val_dist_relevant_mse: 7.8261 - val_loss: 0.8877 - val_prob_kld: 0.0142 - learning_rate: 3.0000e-04
Epoch 161/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9099 - dist_relevant_mae: 0.7385 - dist_relevant_mse: 1.3568 - loss: 0.5218 - prob_kld: 0.0047 - val_dist_dist_iou_metric: 0.8263 - val_dist_relevant_mae: 1.7915 - val_dist_relevant_mse: 8.0138 - val_loss: 0.8930 - val_prob_kld: 0.0197 - learning_rate: 3.0000e-04
Epoch 162/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9133 - dist_relevant_mae: 0.7146 - dist_relevant_mse: 1.2434 - loss: 0.5088 - prob_kld: 0.0042 - val_dist_dist_iou_metric: 0.8246 - val_dist_relevant_mae: 1.8034 - val_dist_relevant_mse: 7.3346 - val_loss: 0.8896 - val_prob_kld: 0.0138 - learning_rate: 3.0000e-04
Epoch 163/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9141 - dist_relevant_mae: 0.7024 - dist_relevant_mse: 1.2010 - loss: 0.5074 - prob_kld: 0.0043 - val_dist_dist_iou_metric: 0.8340 - val_dist_relevant_mae: 1.7334 - val_dist_relevant_mse: 7.3834 - val_loss: 0.8808 - val_prob_kld: 0.0191 - learning_rate: 3.0000e-04
Epoch 164/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9128 - dist_relevant_mae: 0.7231 - dist_relevant_mse: 1.2964 - loss: 0.5317 - prob_kld: 0.0050 - val_dist_dist_iou_metric: 0.8458 - val_dist_relevant_mae: 1.6088 - val_dist_relevant_mse: 6.3382 - val_loss: 0.8517 - val_prob_kld: 0.0148 - learning_rate: 3.0000e-04
Epoch 165/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9143 - dist_relevant_mae: 0.7054 - dist_relevant_mse: 1.2086 - loss: 0.5014 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8184 - val_dist_relevant_mae: 1.8950 - val_dist_relevant_mse: 8.6144 - val_loss: 0.9087 - val_prob_kld: 0.0147 - learning_rate: 3.0000e-04
Epoch 166/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9124 - dist_relevant_mae: 0.7048 - dist_relevant_mse: 1.2019 - loss: 0.5088 - prob_kld: 0.0044 - val_dist_dist_iou_metric: 0.8488 - val_dist_relevant_mae: 1.5648 - val_dist_relevant_mse: 6.3127 - val_loss: 0.8416 - val_prob_kld: 0.0136 - learning_rate: 3.0000e-04
Epoch 167/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9077 - dist_relevant_mae: 0.7804 - dist_relevant_mse: 1.4122 - loss: 0.5330 - prob_kld: 0.0047 - val_dist_dist_iou_metric: 0.8453 - val_dist_relevant_mae: 1.6147 - val_dist_relevant_mse: 6.3674 - val_loss: 0.8530 - val_prob_kld: 0.0150 - learning_rate: 3.0000e-04
Epoch 168/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9102 - dist_relevant_mae: 0.7424 - dist_relevant_mse: 1.3139 - loss: 0.5159 - prob_kld: 0.0053 - val_dist_dist_iou_metric: 0.8104 - val_dist_relevant_mae: 2.0032 - val_dist_relevant_mse: 9.0759 - val_loss: 0.9309 - val_prob_kld: 0.0152 - learning_rate: 3.0000e-04
Epoch 169/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9146 - dist_relevant_mae: 0.7066 - dist_relevant_mse: 1.2175 - loss: 0.5189 - prob_kld: 0.0043 - val_dist_dist_iou_metric: 0.8473 - val_dist_relevant_mae: 1.5737 - val_dist_relevant_mse: 6.1386 - val_loss: 0.8440 - val_prob_kld: 0.0142 - learning_rate: 3.0000e-04
Epoch 170/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9054 - dist_relevant_mae: 0.7942 - dist_relevant_mse: 1.4439 - loss: 0.5206 - prob_kld: 0.0044 - val_dist_dist_iou_metric: 0.8193 - val_dist_relevant_mae: 1.8928 - val_dist_relevant_mse: 8.5140 - val_loss: 0.9097 - val_prob_kld: 0.0161 - learning_rate: 3.0000e-04
Epoch 171/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9148 - dist_relevant_mae: 0.7083 - dist_relevant_mse: 1.2297 - loss: 0.5121 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8241 - val_dist_relevant_mae: 1.8203 - val_dist_relevant_mse: 8.0476 - val_loss: 0.8936 - val_prob_kld: 0.0144 - learning_rate: 3.0000e-04
Epoch 172/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9124 - dist_relevant_mae: 0.7261 - dist_relevant_mse: 1.2576 - loss: 0.5246 - prob_kld: 0.0048 - val_dist_dist_iou_metric: 0.8552 - val_dist_relevant_mae: 1.5410 - val_dist_relevant_mse: 5.7439 - val_loss: 0.8391 - val_prob_kld: 0.0158 - learning_rate: 3.0000e-04
Epoch 173/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9054 - dist_relevant_mae: 0.7783 - dist_relevant_mse: 1.4059 - loss: 0.5300 - prob_kld: 0.0052 - val_dist_dist_iou_metric: 0.8334 - val_dist_relevant_mae: 1.7242 - val_dist_relevant_mse: 7.2322 - val_loss: 0.8751 - val_prob_kld: 0.0152 - learning_rate: 3.0000e-04
Epoch 174/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9154 - dist_relevant_mae: 0.6979 - dist_relevant_mse: 1.1855 - loss: 0.5061 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8392 - val_dist_relevant_mae: 1.6239 - val_dist_relevant_mse: 6.8643 - val_loss: 0.8556 - val_prob_kld: 0.0157 - learning_rate: 3.0000e-04
Epoch 175/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9137 - dist_relevant_mae: 0.7045 - dist_relevant_mse: 1.2078 - loss: 0.5049 - prob_kld: 0.0048 - val_dist_dist_iou_metric: 0.8447 - val_dist_relevant_mae: 1.6127 - val_dist_relevant_mse: 6.5482 - val_loss: 0.8539 - val_prob_kld: 0.0163 - learning_rate: 3.0000e-04
Epoch 176/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9139 - dist_relevant_mae: 0.7021 - dist_relevant_mse: 1.1803 - loss: 0.5106 - prob_kld: 0.0042 - val_dist_dist_iou_metric: 0.8416 - val_dist_relevant_mae: 1.6503 - val_dist_relevant_mse: 6.7175 - val_loss: 0.8601 - val_prob_kld: 0.0149 - learning_rate: 3.0000e-04
Epoch 177/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9104 - dist_relevant_mae: 0.7497 - dist_relevant_mse: 1.3517 - loss: 0.5227 - prob_kld: 0.0043 - val_dist_dist_iou_metric: 0.8225 - val_dist_relevant_mae: 1.8289 - val_dist_relevant_mse: 8.5254 - val_loss: 0.9097 - val_prob_kld: 0.0288 - learning_rate: 3.0000e-04
Epoch 178/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9149 - dist_relevant_mae: 0.6848 - dist_relevant_mse: 1.1571 - loss: 0.5246 - prob_kld: 0.0054 - val_dist_dist_iou_metric: 0.8572 - val_dist_relevant_mae: 1.4947 - val_dist_relevant_mse: 5.6519 - val_loss: 0.8382 - val_prob_kld: 0.0242 - learning_rate: 3.0000e-04
Epoch 179/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9113 - dist_relevant_mae: 0.7183 - dist_relevant_mse: 1.2721 - loss: 0.5057 - prob_kld: 0.0048 - val_dist_dist_iou_metric: 0.8456 - val_dist_relevant_mae: 1.5543 - val_dist_relevant_mse: 6.5966 - val_loss: 0.8411 - val_prob_kld: 0.0152 - learning_rate: 3.0000e-04
Epoch 180/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9171 - dist_relevant_mae: 0.6868 - dist_relevant_mse: 1.1563 - loss: 0.4977 - prob_kld: 0.0038 - val_dist_dist_iou_metric: 0.8373 - val_dist_relevant_mae: 1.7145 - val_dist_relevant_mse: 7.0782 - val_loss: 0.8735 - val_prob_kld: 0.0156 - learning_rate: 3.0000e-04
Epoch 181/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9122 - dist_relevant_mae: 0.7245 - dist_relevant_mse: 1.2475 - loss: 0.5135 - prob_kld: 0.0044 - val_dist_dist_iou_metric: 0.8388 - val_dist_relevant_mae: 1.6563 - val_dist_relevant_mse: 6.8313 - val_loss: 0.8604 - val_prob_kld: 0.0141 - learning_rate: 3.0000e-04
Epoch 182/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9107 - dist_relevant_mae: 0.7292 - dist_relevant_mse: 1.2862 - loss: 0.5092 - prob_kld: 0.0046 - val_dist_dist_iou_metric: 0.8464 - val_dist_relevant_mae: 1.6106 - val_dist_relevant_mse: 6.1444 - val_loss: 0.8515 - val_prob_kld: 0.0143 - learning_rate: 3.0000e-04
Epoch 183/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9145 - dist_relevant_mae: 0.7027 - dist_relevant_mse: 1.1973 - loss: 0.5110 - prob_kld: 0.0039 - val_dist_dist_iou_metric: 0.8344 - val_dist_relevant_mae: 1.7026 - val_dist_relevant_mse: 7.2084 - val_loss: 0.8703 - val_prob_kld: 0.0147 - learning_rate: 3.0000e-04
Epoch 184/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9154 - dist_relevant_mae: 0.6889 - dist_relevant_mse: 1.1611 - loss: 0.5094 - prob_kld: 0.0042 - val_dist_dist_iou_metric: 0.8422 - val_dist_relevant_mae: 1.6499 - val_dist_relevant_mse: 6.5887 - val_loss: 0.8585 - val_prob_kld: 0.0135 - learning_rate: 3.0000e-04
Epoch 185/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9144 - dist_relevant_mae: 0.7036 - dist_relevant_mse: 1.2134 - loss: 0.5057 - prob_kld: 0.0042 - val_dist_dist_iou_metric: 0.7989 - val_dist_relevant_mae: 2.1673 - val_dist_relevant_mse: 9.6188 - val_loss: 0.9614 - val_prob_kld: 0.0129 - learning_rate: 3.0000e-04
Epoch 186/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9049 - dist_relevant_mae: 0.7896 - dist_relevant_mse: 1.4192 - loss: 0.5158 - prob_kld: 0.0052 - val_dist_dist_iou_metric: 0.8273 - val_dist_relevant_mae: 1.7849 - val_dist_relevant_mse: 7.6914 - val_loss: 0.8858 - val_prob_kld: 0.0138 - learning_rate: 3.0000e-04
Epoch 187/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9195 - dist_relevant_mae: 0.6627 - dist_relevant_mse: 1.0657 - loss: 0.5092 - prob_kld: 0.0038 - val_dist_dist_iou_metric: 0.8458 - val_dist_relevant_mae: 1.6073 - val_dist_relevant_mse: 6.4168 - val_loss: 0.8501 - val_prob_kld: 0.0136 - learning_rate: 3.0000e-04
Epoch 188/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9163 - dist_relevant_mae: 0.6794 - dist_relevant_mse: 1.1186 - loss: 0.4949 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8318 - val_dist_relevant_mae: 1.7375 - val_dist_relevant_mse: 7.2679 - val_loss: 0.8762 - val_prob_kld: 0.0137 - learning_rate: 3.0000e-04
Epoch 189/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9173 - dist_relevant_mae: 0.6778 - dist_relevant_mse: 1.1198 - loss: 0.4983 - prob_kld: 0.0037 - val_dist_dist_iou_metric: 0.8199 - val_dist_relevant_mae: 1.8595 - val_dist_relevant_mse: 7.8043 - val_loss: 0.9019 - val_prob_kld: 0.0150 - learning_rate: 3.0000e-04
Epoch 190/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9116 - dist_relevant_mae: 0.7141 - dist_relevant_mse: 1.2508 - loss: 0.5107 - prob_kld: 0.0044 - val_dist_dist_iou_metric: 0.8168 - val_dist_relevant_mae: 1.8802 - val_dist_relevant_mse: 8.3065 - val_loss: 0.9065 - val_prob_kld: 0.0154 - learning_rate: 3.0000e-04
Epoch 191/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9143 - dist_relevant_mae: 0.7257 - dist_relevant_mse: 1.2509 - loss: 0.5211 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8338 - val_dist_relevant_mae: 1.6963 - val_dist_relevant_mse: 7.2549 - val_loss: 0.8724 - val_prob_kld: 0.0181 - learning_rate: 3.0000e-04
Epoch 192/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9127 - dist_relevant_mae: 0.7212 - dist_relevant_mse: 1.2601 - loss: 0.5149 - prob_kld: 0.0043 - val_dist_dist_iou_metric: 0.8188 - val_dist_relevant_mae: 1.8386 - val_dist_relevant_mse: 8.1073 - val_loss: 0.8967 - val_prob_kld: 0.0139 - learning_rate: 3.0000e-04
Epoch 193/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9159 - dist_relevant_mae: 0.6861 - dist_relevant_mse: 1.1788 - loss: 0.5018 - prob_kld: 0.0042 - val_dist_dist_iou_metric: 0.8385 - val_dist_relevant_mae: 1.6256 - val_dist_relevant_mse: 6.8048 - val_loss: 0.8525 - val_prob_kld: 0.0123 - learning_rate: 3.0000e-04
Epoch 194/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9168 - dist_relevant_mae: 0.6767 - dist_relevant_mse: 1.1258 - loss: 0.5001 - prob_kld: 0.0037 - val_dist_dist_iou_metric: 0.8189 - val_dist_relevant_mae: 1.8989 - val_dist_relevant_mse: 8.6153 - val_loss: 0.9089 - val_prob_kld: 0.0141 - learning_rate: 3.0000e-04
Epoch 195/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9213 - dist_relevant_mae: 0.6421 - dist_relevant_mse: 1.0067 - loss: 0.5044 - prob_kld: 0.0036 - val_dist_dist_iou_metric: 0.8248 - val_dist_relevant_mae: 1.8043 - val_dist_relevant_mse: 8.0842 - val_loss: 0.8899 - val_prob_kld: 0.0140 - learning_rate: 3.0000e-04
Epoch 196/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9136 - dist_relevant_mae: 0.7199 - dist_relevant_mse: 1.2778 - loss: 0.5077 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8045 - val_dist_relevant_mae: 2.0171 - val_dist_relevant_mse: 9.6876 - val_loss: 0.9305 - val_prob_kld: 0.0120 - learning_rate: 3.0000e-04
Epoch 197/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 46ms/step - dist_dist_iou_metric: 0.9169 - dist_relevant_mae: 0.6920 - dist_relevant_mse: 1.1656 - loss: 0.4991 - prob_kld: 0.0040 - val_dist_dist_iou_metric: 0.8268 - val_dist_relevant_mae: 1.7723 - val_dist_relevant_mse: 7.6855 - val_loss: 0.8838 - val_prob_kld: 0.0143 - learning_rate: 3.0000e-04
Epoch 198/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9177 - dist_relevant_mae: 0.6825 - dist_relevant_mse: 1.1306 - loss: 0.5199 - prob_kld: 0.0041 - val_dist_dist_iou_metric: 0.8301 - val_dist_relevant_mae: 1.7486 - val_dist_relevant_mse: 7.3297 - val_loss: 0.8776 - val_prob_kld: 0.0128 - learning_rate: 3.0000e-04
Epoch 199/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9189 - dist_relevant_mae: 0.6534 - dist_relevant_mse: 1.0539 - loss: 0.5090 - prob_kld: 0.0039 - val_dist_dist_iou_metric: 0.8473 - val_dist_relevant_mae: 1.5366 - val_dist_relevant_mse: 6.2195 - val_loss: 0.8388 - val_prob_kld: 0.0164 - learning_rate: 3.0000e-04
Epoch 200/200
200/200 ━━━━━━━━━━━━━━━━━━━━ 9s 45ms/step - dist_dist_iou_metric: 0.9168 - dist_relevant_mae: 0.6929 - dist_relevant_mse: 1.1733 - loss: 0.5157 - prob_kld: 0.0045 - val_dist_dist_iou_metric: 0.8394 - val_dist_relevant_mae: 1.6402 - val_dist_relevant_mse: 6.8935 - val_loss: 0.8620 - val_prob_kld: 0.0189 - learning_rate: 3.0000e-04

Loading network weights from 'weights_best.h5'.

<keras.src.callbacks.history.History at 0x7de4ccb7b050>

# Not sure optimize threshold works with sparse labeling
#model.optimize_thresholds(X_train, Y_train)

Test network on one of the training images (self prediction)

Self prediction is not a good way to evaluate the ‘real’ performance of the NN. However it is good sanity test. If the self prediction looks wrong something really went bad.

for n in range(10):
    labels, details = model.predict_instances(X_train[n], prob_thresh=0.5, nms_thresh=0.9)
    fig = imshow_multi2d([X_train[n],Y_train[n],labels],['input','ground truth','predicted instances'],1,3)

Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers). Got range [-0.007905139..1.0].
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers). Got range [-0.024096385..1.0].
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers). Got range [-0.007905139..1.0].
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers). Got range [-0.007905139..1.0].