tarun-ssharma’s gists

tarun-ssharma / tf2.8_infer_with_signature_saved_model.py

Created June 4, 2022 07:38

	import tensorflow as tf
	from object_detection.utils import label_map_util
	from object_detection.utils import visualization_utils as viz_utils
	import random
	from PIL import Image
	import numpy as np
	import glob
	import os

	# TO LOAD

tarun-ssharma / convert_tf2.8_ckpt_to_saved_model.py

Created June 4, 2022 07:33

	from object_detection.utils import config_util
	from object_detection.builders import model_builder

	#Where to save the SavedModel
	output_directory = './try2_trained_model'
	#Recover our saved model with the latest checkpoint:
	pipeline_config = 'pipeline.config'
	#Put the last ckpt from training in here, don't use long pathnames:
	model_dir = './trained_checkpoints/ckpt-31'

tarun-ssharma / jupyter_imports_tf_od_api.md

Created June 4, 2022 07:14

import matplotlib
import matplotlib.pyplot as plt

import os
import random
import io
import imageio
import glob
import scipy.misc

tarun-ssharma / ubuntu1804cuda10.sh

Created May 9, 2022 06:19 — forked from fo40225/ubuntu1804cuda10.sh

install cuda 10 on ubuntu 18.04

	# echo "blacklist nouveau" \| sudo tee -a /etc/modprobe.d/blacklist-nouveau.conf > /dev/null
	# echo "options nouveau modeset=0" \| sudo tee -a /etc/modprobe.d/blacklist-nouveau.conf > /dev/null
	# sudo update-initramfs -u

	sudo apt update
	sudo apt -y install build-essential

	sudo apt update
	sudo apt -y install linux-headers-$(uname -r)

tarun-ssharma / convert_tflite_rep_data_int8.md

Last active April 22, 2022 02:44 — forked from dansitu/convert.py

Convert to tflite with randomly picked representative dataset

model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(input_shape=(28, 28)),
  tf.keras.layers.Dense(128, activation='relu'),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Dense(10, activation='softmax')
])

model.compile(optimizer='adam',

tarun-ssharma / detection_PC.py

Created April 22, 2022 02:40 — forked from iwatake2222/detection_PC.py

Object detection using MobileNet SSD with tensorflow lite (with and without Edge TPU)

	# -- coding: utf-8 --
	import cv2
	import tensorflow as tf
	import numpy as np

	# https://www.tensorflow.org/lite/guide/hosted_models
	# http://storage.googleapis.com/download.tensorflow.org/models/tflite/coco_ssd_mobilenet_v1_1.0_quant_2018_06_29.zip


	def detect_from_camera():

tarun-ssharma / yolov4.py

Created March 8, 2022 12:21 — forked from YashasSamaga/yolov4.py

YOLOv4 on OpenCV DNN

	import cv2
	import time

	CONFIDENCE_THRESHOLD = 0.2
	NMS_THRESHOLD = 0.4
	COLORS = [(0, 255, 255), (255, 255, 0), (0, 255, 0), (255, 0, 0)]

	class_names = []
	with open("classes.txt", "r") as f:
	class_names = [cname.strip() for cname in f.readlines()]

tarun-ssharma / quantized-inference-example.py

Created March 8, 2022 12:21 — forked from ShawnHymel/quantized-inference-example.py

TensorFlow Lite (TFLite) Python Inference Example with Quantization

	# TFLite quantized inference example
	#
	# Based on:
	# https://www.tensorflow.org/lite/performance/post_training_integer_quant
	# https://www.tensorflow.org/lite/api_docs/java/org/tensorflow/lite/Tensor.QuantizationParams

	import numpy as np
	import tensorflow as tf

	# Location of tflite model file (float32 or int8 quantized)

tarun-ssharma / Enabling case-insensitive auto-complete for terminal in MacOS BigSur.md

Last active July 21, 2025 14:42

Enable case-insensitive auto-complete whilst giving preference to actual matches

Add these two lines to your ~/.zshrc file:

zstyle ':completion:*' matcher-list '' 'm:{a-zA-Z}={A-Za-z}' 'r:|=*' 'l:|=* r:|=*'
autoload -Uz compinit && compinit

Tested on:

MacOS BigSur Version 11.5.1

tarun-ssharma / calculate_mean_ap.py

Created February 21, 2022 15:42 — forked from sadjadasghari/calculate_mean_ap.py

Calculate mean Average Precision (mAP) for a set of ground truth and predicted bounding boxes for a set of images.

	"""
	author: Timothy C. Arlen
	date: 28 Feb 2018

	revised by: Sadjad Asghari Esfeden
	date: 10 Sep 2018

	Calculate Mean Average Precision (mAP) for a set of bounding boxes corresponding to specific
	image Ids. Usage:

tarun tarun-ssharma