Closed andreqwert closed 4 years ago
Hi! Thanks for trying my code out! I'll try my best to solve the issue.
the models
directory which stores the pretrained object detection models is different to object_detection.models
. My directory tree structure looks like this:
Folder PATH listing
Volume serial number is DA74-F750
F:.
ÃÄÄÄ.dist
ÃÄÄÄ.idea
³ ÀÄÄÄinspectionProfiles
ÃÄÄÄ.vscode
ÃÄÄÄasset
ÃÄÄÄfunctions
³ ÀÄÄÄ__pycache__
ÃÄÄÄlabel
ÃÄÄÄmodels
³ ÃÄÄÄcenternet_hg104_1024x1024_kpts_coco17_tpu-32
³ ³ ÃÄÄÄcheckpoint
³ ³ ÀÄÄÄsaved_model
³ ³ ÀÄÄÄvariables
³ ÃÄÄÄefficientdet_d0_coco17_tpu-32
³ ³ ÃÄÄÄcheckpoint
³ ³ ÀÄÄÄsaved_model
³ ³ ÃÄÄÄassets
³ ³ ÀÄÄÄvariables
³ ÃÄÄÄfaster_rcnn_inception_resnet_v2_1024x1024_coco17_tpu-8
³ ³ ÃÄÄÄcheckpoint
³ ³ ÀÄÄÄsaved_model
³ ³ ÀÄÄÄvariables
³ ÃÄÄÄfaster_rcnn_resnet101_v1_1024x1024_coco17_tpu-8
³ ³ ÃÄÄÄcheckpoint
³ ³ ÀÄÄÄsaved_model
³ ³ ÀÄÄÄvariables
³ ÃÄÄÄfaster_rcnn_resnet50_coco_2018_01_28
³ ³ ÀÄÄÄsaved_model
³ ³ ÀÄÄÄvariables
³ ÃÄÄÄssd_inception_v2_coco_2017_11_17
³ ÃÄÄÄssd_mobilenet_v1_ppn_2018_07_03
³ ³ ÀÄÄÄsaved_model
³ ³ ÀÄÄÄvariables
³ ÃÄÄÄssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8
³ ³ ÃÄÄÄcheckpoint
³ ³ ÀÄÄÄsaved_model
³ ³ ÀÄÄÄvariables
³ ÀÄÄÄssd_resnet50_v1_fpn_640x640_coco17_tpu-8
³ ÃÄÄÄcheckpoint
³ ÀÄÄÄsaved_model
³ ÃÄÄÄassets
³ ÀÄÄÄvariables
ÃÄÄÄobject_detection
³ ÃÄÄÄanchor_generators
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄbox_coders
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄbuilders
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄcolab_tutorials
³ ÃÄÄÄconfigs
³ ³ ÀÄÄÄtf2
³ ÃÄÄÄcore
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄdata
³ ÃÄÄÄdataset_tools
³ ³ ÃÄÄÄcontext_rcnn
³ ³ ÀÄÄÄdensepose
³ ÃÄÄÄdata_decoders
³ ÃÄÄÄdockerfiles
³ ³ ÃÄÄÄandroid
³ ³ ÃÄÄÄtf1
³ ³ ÀÄÄÄtf2
³ ÃÄÄÄg3doc
³ ³ ÀÄÄÄimg
³ ÃÄÄÄinference
³ ÃÄÄÄlegacy
³ ÃÄÄÄmatchers
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄmeta_architectures
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄmetrics
³ ÃÄÄÄmodels
³ ³ ÃÄÄÄkeras_models
³ ³ ³ ÃÄÄÄbase_models
³ ³ ³ ÀÄÄÄ__pycache__
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄpackages
³ ³ ÃÄÄÄtf1
³ ³ ÀÄÄÄtf2
³ ÃÄÄÄpredictors
³ ³ ÃÄÄÄheads
³ ³ ³ ÀÄÄÄ__pycache__
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄprotos
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄsamples
³ ³ ÃÄÄÄcloud
³ ³ ÀÄÄÄconfigs
³ ÃÄÄÄtest_data
³ ÃÄÄÄtest_images
³ ³ ÃÄÄÄducky
³ ³ ³ ÃÄÄÄtest
³ ³ ³ ÀÄÄÄtrain
³ ³ ÀÄÄÄsnapshot_serengeti
³ ÃÄÄÄtpu_exporters
³ ³ ÀÄÄÄtestdata
³ ³ ÃÄÄÄfaster_rcnn
³ ³ ÀÄÄÄssd
³ ÃÄÄÄutils
³ ³ ÀÄÄÄ__pycache__
³ ÀÄÄÄ__pycache__
ÃÄÄÄofficial
³ ÃÄÄÄbenchmark
³ ³ ÃÄÄÄdatastore
³ ³ ³ ÀÄÄÄschema
³ ³ ÀÄÄÄmodels
³ ³ ÀÄÄÄshakespeare
³ ÃÄÄÄcolab
³ ³ ÀÄÄÄnlp
³ ÃÄÄÄcore
³ ÃÄÄÄmodeling
³ ³ ÃÄÄÄactivations
³ ³ ³ ÀÄÄÄ__pycache__
³ ³ ÃÄÄÄhyperparams
³ ³ ³ ÀÄÄÄ__pycache__
³ ³ ÃÄÄÄoptimization
³ ³ ³ ÀÄÄÄconfigs
³ ³ ÃÄÄÄtraining
³ ³ ÀÄÄÄ__pycache__
³ ÃÄÄÄnlp
³ ³ ÃÄÄÄalbert
³ ³ ÃÄÄÄbert
³ ³ ÃÄÄÄconfigs
³ ³ ÃÄÄÄdata
³ ³ ÃÄÄÄmodeling
³ ³ ³ ÃÄÄÄlayers
³ ³ ³ ÃÄÄÄlosses
³ ³ ³ ÃÄÄÄmodels
³ ³ ³ ÃÄÄÄnetworks
³ ³ ³ ÀÄÄÄops
³ ³ ÃÄÄÄnhnet
³ ³ ³ ÀÄÄÄtestdata
³ ³ ³ ÀÄÄÄcrawled_articles
³ ³ ³ ÃÄÄÄdomain_0.com
³ ³ ³ ÀÄÄÄdomain_1.com
³ ³ ÃÄÄÄtasks
³ ³ ÃÄÄÄtransformer
³ ³ ³ ÀÄÄÄutils
³ ³ ÀÄÄÄxlnet
³ ÃÄÄÄpip_package
³ ÃÄÄÄrecommendation
³ ÃÄÄÄstaging
³ ³ ÀÄÄÄtraining
³ ÃÄÄÄutils
³ ³ ÃÄÄÄflags
³ ³ ÃÄÄÄmisc
³ ³ ÀÄÄÄtesting
³ ³ ÀÄÄÄscripts
³ ÃÄÄÄvision
³ ³ ÃÄÄÄdetection
³ ³ ³ ÃÄÄÄconfigs
³ ³ ³ ÃÄÄÄdataloader
³ ³ ³ ÃÄÄÄevaluation
³ ³ ³ ÃÄÄÄexecutor
³ ³ ³ ÃÄÄÄmodeling
³ ³ ³ ³ ÀÄÄÄarchitecture
³ ³ ³ ÃÄÄÄops
³ ³ ³ ÀÄÄÄutils
³ ³ ³ ÀÄÄÄobject_detection
³ ³ ÃÄÄÄimage_classification
³ ³ ³ ÃÄÄÄconfigs
³ ³ ³ ³ ÀÄÄÄexamples
³ ³ ³ ³ ÃÄÄÄefficientnet
³ ³ ³ ³ ³ ÀÄÄÄimagenet
³ ³ ³ ³ ÀÄÄÄresnet
³ ³ ³ ³ ÀÄÄÄimagenet
³ ³ ³ ÃÄÄÄefficientnet
³ ³ ³ ³ ÀÄÄÄ__pycache__
³ ³ ³ ÃÄÄÄresnet
³ ³ ³ ÀÄÄÄ__pycache__
³ ³ ÀÄÄÄ__pycache__
³ ÀÄÄÄ__pycache__
ÃÄÄÄvideos
ÀÄÄÄ__pycache__
I don't really know the cause now, but maybe you can try this:
object_detection
and official
folder from people-tracker-and-counter
. Then go to TensorFlow 2 Object Detection API and install their environment on another directory. Follow the instructions hereobject_detection
from models/research
and official
from models
in TF2 project directory, then paste it under people-tracker-and-counter
.If the problem still persists, do you mind to copy the error log, line, and the code in the problematic line? I'll try to think the solution as soon as possible. Thank you!
Oh, one more thing. Could to send me the directory tree structure of your project? I haven't tested this on different computer so I'm curious about it.
Now it works! Thanks a lot!
tree -d -L 3
shows:
├── __pycache__
├── asset
├── functions
│ └── __pycache__
├── label
├── models
│ ├── community
│ ├── efficientdet_d0_coco17_tpu-32
│ │ ├── checkpoint
│ │ └── saved_model
│ ├── orbit
│ └── research
│ ├── a3c_blogpost
│ ├── adversarial_text
│ ├── attention_ocr
│ ├── audioset
│ ├── autoaugment
│ ├── cognitive_planning
│ ├── cvt_text
│ ├── deep_speech
│ ├── deeplab
│ ├── delf
│ ├── efficient-hrl
│ ├── lfads
│ ├── lstm_object_detection
│ ├── marco
│ ├── nst_blogpost
│ ├── pcl_rl
│ ├── rebar
│ ├── slim
│ └── vid2depth
├── object_detection
│ ├── __pycache__
│ ├── anchor_generators
│ │ └── __pycache__
│ ├── box_coders
│ │ └── __pycache__
│ ├── builders
│ │ └── __pycache__
│ ├── colab_tutorials
│ ├── configs
│ │ └── tf2
│ ├── core
│ │ └── __pycache__
│ ├── data
│ ├── data_decoders
│ ├── dataset_tools
│ │ ├── context_rcnn
│ │ └── densepose
│ ├── dockerfiles
│ │ ├── android
│ │ ├── tf1
│ │ └── tf2
│ ├── g3doc
│ │ └── img
│ ├── inference
│ ├── legacy
│ ├── matchers
│ │ └── __pycache__
│ ├── meta_architectures
│ │ └── __pycache__
│ ├── metrics
│ ├── models
│ │ ├── __pycache__
│ │ └── keras_models
│ ├── packages
│ │ ├── tf1
│ │ └── tf2
│ ├── predictors
│ │ ├── __pycache__
│ │ └── heads
│ ├── protos
│ │ └── __pycache__
│ ├── samples
│ │ ├── cloud
│ │ └── configs
│ ├── test_data
│ ├── test_images
│ │ ├── ducky
│ │ └── snapshot_serengeti
│ ├── tpu_exporters
│ │ └── testdata
│ └── utils
│ └── __pycache__
├── official
│ ├── __pycache__
│ ├── colab
│ │ └── nlp
│ ├── core
│ ├── modeling
│ │ ├── __pycache__
│ │ ├── activations
│ │ ├── hyperparams
│ │ ├── optimization
│ │ └── training
│ ├── nlp
│ │ ├── albert
│ │ ├── bert
│ │ ├── configs
│ │ ├── data
│ │ ├── modeling
│ │ ├── nhnet
│ │ ├── tasks
│ │ ├── transformer
│ │ └── xlnet
│ ├── pip_package
│ ├── recommendation
│ ├── staging
│ │ └── training
│ ├── utils
│ │ ├── flags
│ │ ├── misc
│ │ └── testing
│ └── vision
│ ├── __pycache__
│ ├── detection
│ └── image_classification
└── videos
I work under Ubuntu 18.01 x64
Hello, Thanks for sharing your code! I tried to launch your repo with the efficientdet model provided in additional requirements. All the described steps in README were completed but I got the following error:
ModuleNotFoundError: No module named 'object_detection.models'
. After that I made a replacement in a way that directorymodels
was in directoryobject_detection
. Since I have the error:ImportError: cannot import name 'center_net_hourglass_feature_extractor'
. Could you please describe how can I solve this problem to launch the script?