Mask_rcnn: No 'instances_train2014.json' File Found Error Issue

[msson]

Hi,

I am trying to train for coco2014 dataset. (I just followed this repo's training example)

I put coco dataset on
home/mson/download/Mask_RCNN/coco/train2014/(image files..)
home/mson/download/Mask_RCNN/coco/val2014/(image files..)
home/mson/download/Mask_RCNN/coco/annotations/(.json files including 'instances_train2014.json')

The command I am using is: python coco.py train --dataset=/path/to/coco/ --model=coco

The result I can see is that
FileNotFoundError: [Errno 2] No such file or directory: '/path/to/coco/annotations /instances_train2014.json'

Can anyone please help? Please see the logs below.

/home/mson/anaconda3/envs/tf/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type.
from ._conv import register_converters as _register_converters
Using TensorFlow backend.
Command: train
Model: coco
Dataset: /path/to/coco/
Year: 2014
Logs: /home/mson/download/Mask_RCNN/logs
Auto Download: False

Configurations:
BACKBONE_SHAPES [[256 256]
[128 128]
[ 64 64]
[ 32 32]
[ 16 16]]
BACKBONE_STRIDES [4, 8, 16, 32, 64]
BATCH_SIZE 2
BBOX_STD_DEV [0.1 0.1 0.2 0.2]
DETECTION_MAX_INSTANCES 100
DETECTION_MIN_CONFIDENCE 0.7
DETECTION_NMS_THRESHOLD 0.3
GPU_COUNT 1
IMAGES_PER_GPU 2
IMAGE_MAX_DIM 1024
IMAGE_MIN_DIM 800
IMAGE_PADDING True
IMAGE_SHAPE [1024 1024 3]
LEARNING_MOMENTUM 0.9
LEARNING_RATE 0.001
MASK_POOL_SIZE 14
MASK_SHAPE [28, 28]
MAX_GT_INSTANCES 100
MEAN_PIXEL [123.7 116.8 103.9]
MINI_MASK_SHAPE (56, 56)
NAME coco
NUM_CLASSES 81
POOL_SIZE 7
POST_NMS_ROIS_INFERENCE 1000
POST_NMS_ROIS_TRAINING 2000
ROI_POSITIVE_RATIO 0.33
RPN_ANCHOR_RATIOS [0.5, 1, 2]
RPN_ANCHOR_SCALES (32, 64, 128, 256, 512)
RPN_ANCHOR_STRIDE 1
RPN_BBOX_STD_DEV [0.1 0.1 0.2 0.2]
RPN_NMS_THRESHOLD 0.7
RPN_TRAIN_ANCHORS_PER_IMAGE 256
STEPS_PER_EPOCH 1000
TRAIN_ROIS_PER_IMAGE 200
USE_MINI_MASK True
USE_RPN_ROIS True
VALIDATION_STEPS 50
WEIGHT_DECAY 0.0001

Loading weights /home/mson/download/Mask_RCNN/mask_rcnn_coco.h5
2018-02-09 16:07:30.877009: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2018-02-09 16:07:31.080556: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:895] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2018-02-09 16:07:31.080894: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1105] Found device 0 with properties:
name: Quadro K5200 major: 3 minor: 5 memoryClockRate(GHz): 0.771
pciBusID: 0000:02:00.0
totalMemory: 7.93GiB freeMemory: 7.51GiB
2018-02-09 16:07:31.080922: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1195] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: Quadro K5200, pci bus id: 0000:02:00.0, compute capability: 3.5)
loading annotations into memory...
Traceback (most recent call last):
File "coco.py", line 476, in
dataset_train.load_coco(args.dataset, "train", year=args.year, auto_download=args.download)
File "coco.py", line 108, in load_coco
coco = COCO("{}annotations/instances_{}{}.json".format(dataset_dir, subset, year))
File "/home/mson/anaconda3/envs/tf/lib/python3.6/site-packages/pycocotools/coco.py", line 79, in __init__
dataset = json.load(open(annotation_file, 'r'))
FileNotFoundError: [Errno 2] No such file or directory: '/path/to/coco/annotations/instances_train2014.json'

[msson]

I tried to solve this prob for two days please help me thanks!

[monomon]

You need to pass the path to coco as the --dataset argument,
/path/to/coco/ has to become /home/mson/download/Mask_RCNN/coco, which gives you

python coco.py train --dataset=/home/mson/download/Mask_RCNN/coco --model=coco

You do not think coco will magically find your dataset, right?

[Alvin1994]

@msson ,
As @monomon said, that is the right way to pass coco data set, if it still failed. You need to remove the json file manually and let the script download again.

[sumodm]

Another potential issue if you give --dataset=~/<path>, it won't work as it is unable to expand '~' to /home/user/.