import os
import sys
sys.path.append("Monk_Object_Detection/12_tf_obj_1/lib/")
from infer_detector import Infer
gtf = Infer()
# Model loading takes time on nano boards
gtf.set_model_params('trt_fp16_dir/trt_graph.pb', "ship/classes.txt")
# Running for the first time builds the tensorRT engine for the model based on the plan saved in trt_fp16_dir folder
# Oputput will be saved as output.png
scores, bboxes, labels = gtf.infer_on_image('ship/test/img5.jpg',
thresh=0.5,
img_size=300)
# Run speed benchmark
gtf.benchmark_for_speed('ship/test/img1.jpg', img_size=300)
sys.stdout = Unbuffered(sys.stdout)
print("Predicting....")
with open('obj_7_yolov3_infer.json') as json_file:
system = json.load(json_file)
system["conf_thresh"] = float(system["conf_thresh"])
system["iou_thresh"] = float(system["iou_thresh"])
class_file = system["class_file"]
f = open(class_file, 'r')
class_list = f.readlines()
f.close()
for i in range(len(class_list)):
class_list[i] = class_list[i][:len(class_list[i]) - 1]
gtf = Infer()
gtf.Model(system["model"],
class_list,
system["weights"],
use_gpu=True,
input_size=416)
output = gtf.Predict(system["img_file"],
conf_thres=system["conf_thresh"],
iou_thres=system["iou_thresh"])
print("Completed")
import sys
import cv2 as cv
sys.path.append("../lib")
from utils.datasets import LoadWebcam
from infer_detector import Infer
infer=Infer(0)
classPath="classes.txt"
# weightsDir="../lib/weights/last_5epoch_2numgen.pt"
weightsDir="../lib/weights/last_5epoch_2numgen.pt"
infer.Model("yolov3", classPath, weightsDir, use_gpu=True, half_precision=True)
web=LoadWebcam()
iter(web)
while True:
imgPath,img,img0,_=next(web)
cv.imwrite("webcamFrames/frame.jpg", img0)
infer.Predict("D:\\Repo\\Monk_Object_Detection\\7_yolov3\HandsTrain\\webcamFrames\\frame.jpg")
out=cv.imread("output/frame.jpg")
cv.imshow("frame", out)
with open('obj_6_cornernet_lite_infer.json') as json_file:
system = json.load(json_file)
system["conf_thresh"] = float(system["conf_thresh"])
class_file = system["class_file"];
f = open(class_file, 'r');
class_list = f.readlines();
f.close();
for i in range(len(class_list)):
class_list[i] = class_list[i][:len(class_list[i])-1]
gtf = Infer();
gtf.Model(class_list,
base=system["model"],
model_path=system["weights"])
output = gtf.Predict(system["img_file"],
vis_thresh=system["conf_thresh"],
output_img="output.jpg");
print("Completed")
sys.stdout = Unbuffered(sys.stdout)
print("Predicting....")
with open('obj_4_efficientdet_infer.json') as json_file:
system = json.load(json_file)
if (system["use_gpu"] == "yes"):
system["use_gpu"] = True
else:
system["use_gpu"] = False
system["conf_thresh"] = float(system["conf_thresh"])
class_file = system["class_file"]
f = open(class_file, 'r')
class_list = f.readlines()
f.close()
for i in range(len(class_list)):
class_list[i] = class_list[i][:len(class_list[i]) - 1]
gtf = Infer()
gtf.Model(model_dir=system["weights_dir"])
scores, labels, boxes = gtf.Predict(system["img_file"],
class_list,
vis_threshold=system["conf_thresh"])
print("Completed")
timg_dir = "images";
tset_dir = "Train";
vroot_dir = "Root_Dir";
vcoco_dir = "Coco_style";
vimg_dir = "images";
vset_dir = "Val";
model.Train_Dataset(troot_dir, tcoco_dir, timg_dir, tset_dir, batch_size=8, image_size=352, use_gpu=True)
model.Val_Dataset(vroot_dir, vcoco_dir, vimg_dir, vset_dir)
model.Model(model_name="resnet34"); # resnet 50 brought cuda memory error.
model.Set_Hyperparams(lr=0.0001, val_interval=1, print_interval=20)
model.Train(num_epochs=300,output_model_name="karen_model.pt");
from infer_detector import Infer
gtf = Infer();
gtf.Model(model_path="/content/karen_model.pt");
#predictions are quite bad at the moment.
class_list=[]
with open("/content/Root_Dir/Coco_style/annotations/classes.txt") as file:
for line in file:
class_list.append(line.rstrip("\n"))
class_list=class_list[:-1]
img_p="/content/Images_and_Labels/Images/0000002_00005_d_0000014_jpg.rf.555bf2106d899e56d45da0a48295f04c.jpg"
scores, labels, boxes = gtf.Predict(img_p, class_list, vis_threshold=0.4);
from IPython.display import Image
Image(filename='output.jpg')
import cv2 as cv
import os
import sys
import shutil
import time
sys.path.append("../lib")
from infer_detector import Infer
from utils.datasets import LoadWebcam
infer = Infer(0)
classPath = "classes.txt"
weightsDir = "../lib/weights/last_7e_2n.pt"
#weightsDir="../lib/weights/yolov3-tiny-8e-2n.pt"
infer.Model("yolov3", classPath, weightsDir, use_gpu=True)
# start=time.time()
# disp=2
# fps=0
cap = cv.VideoCapture(0)
prev = 0
new = 0
while True:
ret, frame = cap.read()
# frame=cv.flip(frame, 1)
#frame=cv.resize(frame, (416, 416))
print(frame.shape)
cv.imwrite("webcamFrames/frame.jpg", frame)
font = cv.FONT_HERSHEY_SIMPLEX
new = time.time()
fps = 1 / (new - prev)
prev = new
fps = int(fps)