|
41 | 41 | "cell_type": "code",
|
42 | 42 | "execution_count": 1,
|
43 | 43 | "metadata": {
|
44 |
| - "collapsed": false, |
45 | 44 | "scrolled": true
|
46 | 45 | },
|
47 | 46 | "outputs": [
|
|
80 | 79 | {
|
81 | 80 | "cell_type": "code",
|
82 | 81 | "execution_count": 2,
|
83 |
| - "metadata": { |
84 |
| - "collapsed": false |
85 |
| - }, |
| 82 | + "metadata": {}, |
86 | 83 | "outputs": [
|
87 | 84 | {
|
88 | 85 | "data": {
|
|
140 | 137 | {
|
141 | 138 | "cell_type": "code",
|
142 | 139 | "execution_count": 4,
|
143 |
| - "metadata": { |
144 |
| - "collapsed": false |
145 |
| - }, |
| 140 | + "metadata": {}, |
146 | 141 | "outputs": [
|
147 | 142 | {
|
148 | 143 | "data": {
|
|
181 | 176 | {
|
182 | 177 | "cell_type": "code",
|
183 | 178 | "execution_count": 5,
|
184 |
| - "metadata": { |
185 |
| - "collapsed": false |
186 |
| - }, |
| 179 | + "metadata": {}, |
187 | 180 | "outputs": [],
|
188 | 181 | "source": [
|
189 | 182 | "import inception5h"
|
|
217 | 210 | {
|
218 | 211 | "cell_type": "code",
|
219 | 212 | "execution_count": 7,
|
220 |
| - "metadata": { |
221 |
| - "collapsed": false |
222 |
| - }, |
| 213 | + "metadata": {}, |
223 | 214 | "outputs": [
|
224 | 215 | {
|
225 | 216 | "name": "stdout",
|
|
263 | 254 | "cell_type": "code",
|
264 | 255 | "execution_count": 9,
|
265 | 256 | "metadata": {
|
266 |
| - "collapsed": false, |
267 | 257 | "scrolled": true
|
268 | 258 | },
|
269 | 259 | "outputs": [
|
|
671 | 661 | " # Calculate the value of the gradient.\n",
|
672 | 662 | " # This tells us how to change the image so as to\n",
|
673 | 663 | " # maximize the mean of the given layer-tensor.\n",
|
674 |
| - " grad = tiled_gradient(gradient=gradient, image=img)\n", |
| 664 | + " grad = tiled_gradient(gradient=gradient, image=img,\n", |
| 665 | + " tile_size=tile_size)\n", |
675 | 666 | " \n",
|
676 | 667 | " # Blur the gradient with different amounts and add\n",
|
677 | 668 | " # them together. The blur amount is also increased\n",
|
|
839 | 830 | {
|
840 | 831 | "cell_type": "code",
|
841 | 832 | "execution_count": 21,
|
842 |
| - "metadata": { |
843 |
| - "collapsed": false |
844 |
| - }, |
| 833 | + "metadata": {}, |
845 | 834 | "outputs": [
|
846 | 835 | {
|
847 | 836 | "data": {
|
|
870 | 859 | "cell_type": "code",
|
871 | 860 | "execution_count": 22,
|
872 | 861 | "metadata": {
|
873 |
| - "collapsed": false, |
874 | 862 | "scrolled": true
|
875 | 863 | },
|
876 | 864 | "outputs": [
|
|
900 | 888 | {
|
901 | 889 | "cell_type": "code",
|
902 | 890 | "execution_count": 23,
|
903 |
| - "metadata": { |
904 |
| - "collapsed": false |
905 |
| - }, |
| 891 | + "metadata": {}, |
906 | 892 | "outputs": [
|
907 | 893 | {
|
908 | 894 | "name": "stdout",
|
|
1147 | 1133 | "cell_type": "code",
|
1148 | 1134 | "execution_count": 25,
|
1149 | 1135 | "metadata": {
|
1150 |
| - "collapsed": false, |
1151 | 1136 | "scrolled": true
|
1152 | 1137 | },
|
1153 | 1138 | "outputs": [
|
|
1350 | 1335 | {
|
1351 | 1336 | "cell_type": "code",
|
1352 | 1337 | "execution_count": null,
|
1353 |
| - "metadata": { |
1354 |
| - "collapsed": false |
1355 |
| - }, |
| 1338 | + "metadata": {}, |
1356 | 1339 | "outputs": [],
|
1357 | 1340 | "source": [
|
1358 | 1341 | "layer_tensor = model.layer_tensors[6]\n",
|
|
1371 | 1354 | {
|
1372 | 1355 | "cell_type": "code",
|
1373 | 1356 | "execution_count": null,
|
1374 |
| - "metadata": { |
1375 |
| - "collapsed": false |
1376 |
| - }, |
| 1357 | + "metadata": {}, |
1377 | 1358 | "outputs": [],
|
1378 | 1359 | "source": [
|
1379 | 1360 | "layer_tensor = model.layer_tensors[7][:,:,:,0:3]\n",
|
|
1392 | 1373 | {
|
1393 | 1374 | "cell_type": "code",
|
1394 | 1375 | "execution_count": null,
|
1395 |
| - "metadata": { |
1396 |
| - "collapsed": false |
1397 |
| - }, |
| 1376 | + "metadata": {}, |
1398 | 1377 | "outputs": [],
|
1399 | 1378 | "source": [
|
1400 | 1379 | "layer_tensor = model.layer_tensors[11][:,:,:,0]\n",
|
|
1413 | 1392 | {
|
1414 | 1393 | "cell_type": "code",
|
1415 | 1394 | "execution_count": null,
|
1416 |
| - "metadata": { |
1417 |
| - "collapsed": false |
1418 |
| - }, |
| 1395 | + "metadata": {}, |
1419 | 1396 | "outputs": [],
|
1420 | 1397 | "source": [
|
1421 | 1398 | "image = load_image(filename='images/giger.jpg')\n",
|
|
1425 | 1402 | {
|
1426 | 1403 | "cell_type": "code",
|
1427 | 1404 | "execution_count": null,
|
1428 |
| - "metadata": { |
1429 |
| - "collapsed": false |
1430 |
| - }, |
| 1405 | + "metadata": {}, |
1431 | 1406 | "outputs": [],
|
1432 | 1407 | "source": [
|
1433 | 1408 | "layer_tensor = model.layer_tensors[3]\n",
|
|
1439 | 1414 | {
|
1440 | 1415 | "cell_type": "code",
|
1441 | 1416 | "execution_count": null,
|
1442 |
| - "metadata": { |
1443 |
| - "collapsed": false |
1444 |
| - }, |
| 1417 | + "metadata": {}, |
1445 | 1418 | "outputs": [],
|
1446 | 1419 | "source": [
|
1447 | 1420 | "layer_tensor = model.layer_tensors[5]\n",
|
|
1461 | 1434 | "cell_type": "code",
|
1462 | 1435 | "execution_count": null,
|
1463 | 1436 | "metadata": {
|
1464 |
| - "collapsed": false, |
1465 | 1437 | "scrolled": true
|
1466 | 1438 | },
|
1467 | 1439 | "outputs": [],
|
|
1473 | 1445 | {
|
1474 | 1446 | "cell_type": "code",
|
1475 | 1447 | "execution_count": null,
|
1476 |
| - "metadata": { |
1477 |
| - "collapsed": false |
1478 |
| - }, |
| 1448 | + "metadata": {}, |
1479 | 1449 | "outputs": [],
|
1480 | 1450 | "source": [
|
1481 | 1451 | "layer_tensor = model.layer_tensors[6]\n",
|
|
1561 | 1531 | "metadata": {
|
1562 | 1532 | "anaconda-cloud": {},
|
1563 | 1533 | "kernelspec": {
|
1564 |
| - "display_name": "Python [default]", |
| 1534 | + "display_name": "Python 3", |
1565 | 1535 | "language": "python",
|
1566 | 1536 | "name": "python3"
|
1567 | 1537 | },
|
|
1575 | 1545 | "name": "python",
|
1576 | 1546 | "nbconvert_exporter": "python",
|
1577 | 1547 | "pygments_lexer": "ipython3",
|
1578 |
| - "version": "3.5.2" |
| 1548 | + "version": "3.6.6" |
1579 | 1549 | }
|
1580 | 1550 | },
|
1581 | 1551 | "nbformat": 4,
|
1582 |
| - "nbformat_minor": 0 |
| 1552 | + "nbformat_minor": 1 |
1583 | 1553 | }
|
0 commit comments