[failed!] proses pembuatan model tensorflow keras google colab

Опубликовано: 05 Август 2026
на канале: faiq hatta

137

percobaan pembuatan model sedang dalam masalah kestabilan jaringan internet.
Tensorflow-1.5.0
python 3.6
object detection api 0.1
MSVC 2019
Bazel 3.7.2

bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package

pilih antara train dan test
menjadi satu folder yang berbeda

urutkan penamaan sehingga tidak
numpuk nama filenya (name.py)

ubuah file test ke xml
(pake app labeling)

install model tf yang lama
(model archeive/research/cmd)
python setup.py build
(python setup.py install)

ganti file (generate_tfrecord.py)
berdasarkan nama lebel

ganti file (labelmap.pbtxt)
berdasarkan nama lebel

ganti file (ssd_mobilenet_v2_quantized_300x300_coco.config)
ganti jumlah label (10an)
ganti num_sampel (180an)
sesuai jumlah img test

(folder images sama (train dan test))
python xml_to_csv.py
Successfully converted xml to csv.
Successfully converted xml to csv.

setelah pengecekan pada file csv test
sesuai dengan label
python generate_tfrecord.py --csv_input=images\train_labels.csv --image_dir=images\train --output_path=train.record

python generate_tfrecord.py --csv_input=images\test_labels.csv --image_dir=images\test --output_path=test.record

repository-- https://github.com/tensorflow

all files--
https://drive.google.com/

(Please download specific version)
For Raspberry pi 3/4 :-(Faster but less accurate)(4-10 FPS depending on raspberry pi)
tensorflow.ipynb -- https://drive.google.com/
ssd_mobilenet_v2_quantized_300x300_coco.config -- https://drive.google.com/

ONLY For Raspberry pi 4 :-(slower but more accurate)(1-2 FPS)
tensorflow.ipynb-- https://drive.google.com/
ssd_mobilenet_v2_quantized_300x300_coco.config -- https://drive.google.com/