-
Notifications
You must be signed in to change notification settings - Fork 0
/
index_old.html
388 lines (338 loc) · 19.3 KB
/
index_old.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
<!DOCTYPE html>
<html class="" lang="en"><head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="keywords" content="Deep Learning and Computer Vision Specialist, Abhishek, Agrawal, Deep Learning programmer">
<title>Abhishek Agrawal </title>
<link rel="icon" href="https://abhiagwl4262.github.io/favicon-S.ico" type="image/x-icon">
<link href="abhishek_agrawal_files/css.css" rel="stylesheet">
<link rel="stylesheet" href="abhishek_agrawal_files/font-awesome.css">
<link href="abhishek_agrawal_files/ .css" rel="stylesheet">
<link href="abhishek_agrawal_files/styles.css" rel="stylesheet">
</head>
<body>
<div id="mobile-menu-open" class="shadow-large">
<i class="fa fa-bars" aria-hidden="true"></i>
</div>
<!-- End #mobile-menu-toggle -->
<header>
<div id="mobile-menu-close">
<span>Close</span> <i class="fa fa-times" aria-hidden="true"></i>
</div>
<ul id="menu" class="shadow">
<li>
<a href="#about">About</a>
</li>
<li>
<a href="#skills">Skills</a>
</li>
<li>
<a href="#experience">Experience</a>
</li>
<li>
<a href="#projects">Projects</a>
</li>
<li>
<a href="#education">Education</a>
</li>
<li>
<a href="#contact">Contact</a>
</li>
</ul>
</header>
<!-- End header -->
<div id="lead">
<div id="lead-content">
<h1 style="color: #42b3f4">Abhishek Agrawal</h1>
<h2>Deep Learning Engineer</h2>
<a href="https://abhiagwl4262.github.io/docs/abhishek_resume.pdf" class="btn-rounded-white" target="_blank">Download Resume</a>
<div style="margin: 10px; position:relative">
<a style="padding:10px; font-size:50px; opacity:1.0; color:white" href="https://github.com/abhiagwl4262/" target="_blank"><i class="fa fa-github"></i></a>
<a style="padding:10px; font-size:50px; opacity:1.0; color:white" href="https://www.linkedin.com/in/abhiagwl/" target="_blank"><i class="fa fa-linkedin-in"></i></a>
<!--a style="padding:10px; font-size:50px; color:white" href="http://aimlrobots.blogspot.in/" target="_blank"><i class="fa fa-blogger" aria-hidden="true"></i></a-->
</div>
</div>
<!-- End #lead-content -->
<!--div id="lead-overlay"></div-->
<div id="lead-down">
<span>
<i class="fa fa-chevron-down" aria-hidden="true"></i>
</span>
</div>
<!-- End #lead-down -->
</div>
<!-- End #lead -->
<div id="about">
<div class="container">
<div class="row">
<div class="col-md-4">
<h2 class="heading">
About Me
</h2>
<!-- <div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px; display:inline-block"> -->
<!-- <img src="abhishek_agrawal_files/tracking.png"> -->
<img class="profile_image" src="abhishek_agrawal_files/IMG_2328.JPG" alt="profile_pic">
<!-- <img class="profile_image" src="abhishek_agrawal_files/IMG_2328.JPG" alt="profile_pic"> -->
</div>
<div class="col-md-8">
<p style="text-align:left">
<br>
<br>
<br>
Exploring Artificial Intelligenace world. I have always been fascinated with the interdisciplinary scientific field of computer vision. Deep Learning entered into my carrer in form of an internship opportunity. Then there was no looking back. Medical Imaging, Autonomous Vehicles and huge possibilities in AgriTech excite me the most about.
<br>
<br>
I am currently exploring one of the of biggest challenge of Deep Learning, challenge of making things quick. As part of ThinCI, I am working on reducing the FLOPs and memory uses in order to optimize inferencing. Unfortunately even after so much exploitation of this field, it still takes days to get results in deep learning.
<br>
<br>
Previously, I worked on ADAS (Advance Driver Assistive System) which features FCW(Forward Collision Warning), LDW(Lane Departure warning), BSD(Blind Spot Detection). I spent most of my time exploring the most talked problem of Object Detection and Object Tracking followed by Depth Estimation. It is fascinating to see the way transportation industry is transforming. A lot of fleets, automobile and insurance companies are eagerly looking at automation industry.
</p>
</div>
</div>
</div>
</div>
<!-- End #about -->
<div id="skills">
<h2 class="heading">Skills</h2>
<ul>
<li>C</li>
<li>C++</li>
<li>Python</li>
<li>ONNX</li>
<li>Pytorch</li>
<li>Tensorflow</li>
<li>CUDA</li>
<li>OpenCV</li>
<li>Eigen</li>
<li>Linux</li>
<li>ROS</li>
<li>NVIDIA JETSON TX1</li>
</ul>
</div>
<!-- End #skills -->
<div id="experience" class="background-alt">
<h2 class="heading">Experience</h2>
<div id="experience-timeline">
<div class="vtimeline-point"><div class="vtimeline-icon"><i class="fa fa-map-marker"></i></div><div class="vtimeline-block"><span class="vtimeline-date">Feb 2019 to Present Day</span><div data-date="Feb 2019 to Present Day" class="vtimeline-content">
<h3>Thinci Semiconductor Technologies India Pvt. Ltd, Hyderabad, India </h3>
<h4>Research Engineer I</h4>
<p>
Distiller for Convolution Neural Network
</p>
</div></div></div>
<div class="vtimeline-point"><div class="vtimeline-icon"><i class="fa fa-map-marker"></i></div><div class="vtimeline-block"><span class="vtimeline-date">June 2017 – Jan 2019</span><div data-date="June 2017 – Jan 2019" class="vtimeline-content">
<h3>TheHiTechRoboticSystemz Ltd, Gurugram, India</h3>
<h4>Research Engineer</h4>
<p>
Advanced Driver Assistance Systems
</p>
</div></div></div>
<div class="vtimeline-point"><div class="vtimeline-icon"><i class="fa fa-map-marker"></i></div><div class="vtimeline-block"><span class="vtimeline-date">May 2016 – July 2016</span><div data-date="May 2016 – July 2016" class="vtimeline-content">
<h3>TCS Innovation Lab Bangalore, India</h3>
<h4>Summer Internship</h4>
<p>
Cattle Recognition for Advanced Driver Assistance Systems
</p>
</div></div></div>
</div>
</div>
<!-- End #experience -->
<div id="projects" class="background-alt">
<h2 class="heading">Projects</h2>
<div class="container">
<div class="row">
<div class="project shadow-large">
<!--div class="project-image"-->
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px; display:inline-block">
<img src="abhishek_agrawal_files/tracking.png">
</div>
<!-- End .project-image -->
<div class="project-info">
<h3>MultiObjectTracking</h3>
<p>
<!--For a successful pick of object using a suction based gripper requires a robust estimation of position(touch point for suction) and direction of approach for the robot.-->
This project implements the realtime multiple object tracking for the driving scene. In order to predict the spatio temporal behaviour of any object in moving environment, motion cues are generated with tracking algorithm. This algorithm is of great importance for survillance too.
</p>
<a href="https://abhiagwl4262.github.io/pages/Multi Object Tracking.html" target="_blank">View Project</a>
</div>
<!-- End .project-info -->
</div>
<div class="project shadow-large">
<!--div class="project-image"-->
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px; display:inline-block">
<img src="abhishek_agrawal_files/Forward-Collision-Mitigation-System.png">
</div>
<!-- End .project-image -->
<div class="project-info">
<h3>Forward Collision Warning System</h3>
<p>
<!--For a successful pick of object using a suction based gripper requires a robust estimation of position(touch point for suction) and direction of approach for the robot.-->
We built a system that senses the possibility of any hazardous collision between subject vehicle with any other vehicle in front and warns the driver about it or algorithm might take control from the driver as well. As reaction time of an algorithm can be way too less than that of a human and supremely, <b>Algorithms Don't Panic</b>.
</p>
<!-- <a href="https://abhiagwl4262.github.io/pages/Forward Collision Warning System.html" target="_blank">View Project</a> -->
<a href="https://drive.google.com/file/d/17WMlOddwIb0mRZyWM8CIN-qV3bt2t56h/view" target="_blank">View Project</a>
</div>
<!-- End .project-info -->
</div>
<!-- End .project -->
<div class="project shadow-large">
<!--div class="project-image"-->
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px;">
<img src="abhishek_agrawal_files/disparity.png">
</div>
<!-- End .project-image -->
<div class="project-info">
<h3>Unsupervised Monocular Depth Estimation</h3>
<p>
This project acts in the direction of <b>learning disparity map using stero pair in unsupervised manner</b>. As the process of generating groundtruth for 3D depth is very costly and time consuming, we want to train the network without any needs of groundtruth depth data.
</p>
<a href="https://abhiagwl4262.github.io/pages/Stereo Disparity.html" target="_blank">View Project</a>
</div>
<!-- End .project-info -->
</div>
<!-- End .project -->
<div class="project shadow-large">
<!--div class="project-image"-->
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px;">
<img src="abhishek_agrawal_files/tensor_Decomposition.png">
</div>
<!-- End .project-image -->
<div class="project-info">
<h3>Low rank approximation of CNNs</h3>
<p>
This project deals with the distilling of Convolution Neural Networks. We change the topology of the network in such fashion that we get almost same performace with reduction of FLOPs upto 0.5times. <b>The distilled model would be faster and with less memory footprint<b></b>.
</p>
<a href="https://abhiagwl4262.github.io/pages/LRA_model.html" target="_blank">View Project</a>
<!-- a href="#">View Project</a -->
</div>
<!-- End .project-info -->
</div>
<!-- End .project -->
<div class="project shadow-large">
<!--div class="project-image"-->
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px;">
<img src="abhishek_agrawal_files/Onnx.jpeg">
</div>
<!-- End .project-image -->
<div class="project-info">
<h3> ONNX to Pytorch Convertor for Low rank approximation of CNNs </h3>
<p>
This project is about transforming the given model into low rank space model. This tool automates the process of Low rank model creation using the defined guidelines followed by well conditioning and quantization of the given model. We feed trained CNN model in ONNX format along with the data and training parameters as input and then this tool makes the necessary changes and generates the converted low rank optimized PyTorch model.
</p>
<!--a href="#">View Project</a-->
</div>
<!-- End .project-info -->
</div>
<!-- End .project -->
<div class="project shadow-large">
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px;">
<img src="abhishek_agrawal_files/fast_convergance.jpeg">
</div>
<!-- End .project-image -->
<div class="project-info">
<h3>Faster Convergance of CNNs</h3>
<p>
This project aims towards reducing the training needs for any deep neural network. We explore the method so that neural networks can be trained faster than with standard training methods. A primary insight that allows fast-convergence training is that large learning rates regularize the training.
</p>
<!-- <a href="https://youtu.be/g2B7S7s61vo" target="_blank">View Video</a> -->
</div>
<!-- End .project-info -->
</div>
<!-- End .project -->
<!-- <div class="project shadow-large">
<div style="float:left;margin:22px 20px 22px 10px; border: 1px solid gray;border-radius: 4px; padding: 5px;">
<img src="images/coolarm2.png" />
</div>
<!End .project-image -->
<!-- <div class="project-info">
<h3>Simulation of motion planning for Cool1000 robotic manipulator in RViz</h3>
<p>
This project presents the motion planning simulated in RViz environment for a custom built robotic manipulator Cool1000. The video demonstrates the capabilities of MoveIt for placing collision objects, picking and attaching an object to the gripper and transfer the object to another place. The blog presents a brief overview of steps to be followed in order to create a URDF file for a robot.
</p>
<a href="https://youtu.be/1IkDwN0fcu8" target="_blank">View Video</a>
<a href="http://aimlrobots.blogspot.in/2017/11/creating-urdf-robot-model-from-scratch.html" target="_blank">View Blog</a>
</div-->
<!-- End .project-info -->
<!-- End .project -->
</div>
</div>
</div>
<!-- End #projects -->
<div id="education">
<h2 class="heading">Education</h2>
<div class="education-block">
<h3>Indian Institute of Technology Jodhpur (IIT Jodhpur), India</h3>
<span class="education-date">2013 - 2017</span>
<h4>Bachelor of Technology in Electrical Engineering </h4>
<!-- <br><br> -->
<!-- <h2>Projects</h2> -->
<!-- <h4>Automatic License Plate Recognition</h4> -->
<!-- <br><br> -->
<!-- <h4>Eye Blink Dtection with Optical Flow</h4> -->
<!-- <br><br> -->
<!-- <h4>Generating Synthetic Handwriting Data</h4> -->
<!-- <br><br> -->
<!-- <h4>Graph Based Image Segmentation </h4> -->
</div>
<!-- End .education-block -->
<!--div class="education-block">
<h3>BMS College of Engineering, Bangalore, India</h3>
<span class="education-date">2006 - 2010</span>
<h4>Bachelor of Engineering in Electronics & Communication Engineering</h4>
</div-->
<!-- End .education-block -->
</div>
<!-- End #education -->
<div id="contact" class="contact">
<h2>Get in Touch</h2>
<div id="contact-form">
<form method="POST" action="mailto:abhiagwl4262@gmail.com">
<input type="hidden" name="_subject" value="Contact request from personal website">
<input type="email" name="_replyto" placeholder="Your email" required="">
<textarea name="message" placeholder="Your message" required=""></textarea>
<button type="submit">Send</button>
</form>
</div>
<!-- End #contact-form -->
</div>
<!-- End #contact -->
<footer>
<div class="container">
<div class="row">
<div class="col-sm-5 copyright">
<p>
Copyright © 2019 Abhishek Agrawal
</p></div>
<div class="col-sm-2 top">
<span id="to-top">
<i class="fa fa-chevron-up" aria-hidden="true"></i>
</span>
</div>
</div>
</div>
<!--div class="col-sm-5 social">
<ul>
<li>
<a href="https://github.com/sharathrjtr/" target="_blank"><i class="fa fa-github" aria-hidden="true"></i></a>
</li>
<li>
<a href="https://stackoverflow.com/" target="_blank"><i class="fa fa-stack-overflow" aria-hidden="true"></i></a>
</li>
<li>
<a href="https://www.facebook.com/" target="_blank"><i class="fa fa-facebook" aria-hidden="true"></i></a>
</li>
<li>
<a href="https://twitter.com/" target="_blank"><i class="fa fa-twitter" aria-hidden="true"></i></a>
</li>
<li>
<a href="https://www.linkedin.com/in/sharath-jotawar" target="_blank"><i class="fa fa-linkedin" aria-hidden="true"></i></a>
</li>
</ul>
</div-->
</footer>
<!-- End footer -->
<script src="abhishek_agrawal_files/jquery.js"></script>
<script src="abhishek_agrawal_files/scripts.js"></script>
</body></html>