File size: 52,428 Bytes
4a1f918 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 |
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import numpy as np\n",
"import torch\n",
"import os\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"from PIL import Image\n"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"import matplotlib.pyplot as plt"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"im_path = \"/media/ubuntu/New Volume/jay/fundus_images/archive/REFUGE/val/Masks/V0001.png\"\n",
"img = torch.as_tensor(np.array(Image.open(im_path).convert(\"RGB\")))\n"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([0, 1, 2], dtype=uint8)"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"np.unique(img[:,:,0])"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"model_dict_path = \"./eval/cholec8k/svd_shiftscale_cholec_tal_focal075_1e-4.pth\"\n",
"model_dict = torch.load(model_dict_path, map_location='cpu')"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"keys = list(model_dict.keys())"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['sam_encoder.pos_embed',\n",
" 'sam_encoder.patch_embed.proj.weight',\n",
" 'sam_encoder.patch_embed.proj.bias',\n",
" 'sam_encoder.blocks.0.norm1.weight',\n",
" 'sam_encoder.blocks.0.norm1.bias',\n",
" 'sam_encoder.blocks.0.attn.qkv.weight',\n",
" 'sam_encoder.blocks.0.attn.qkv.bias',\n",
" 'sam_encoder.blocks.0.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.0.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.0.attn.proj.weight',\n",
" 'sam_encoder.blocks.0.attn.proj.bias',\n",
" 'sam_encoder.blocks.0.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.0.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.0.norm2.weight',\n",
" 'sam_encoder.blocks.0.norm2.bias',\n",
" 'sam_encoder.blocks.0.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.0.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.0.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.0.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.0.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.0.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.0.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.0.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.1.norm1.weight',\n",
" 'sam_encoder.blocks.1.norm1.bias',\n",
" 'sam_encoder.blocks.1.attn.qkv.weight',\n",
" 'sam_encoder.blocks.1.attn.qkv.bias',\n",
" 'sam_encoder.blocks.1.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.1.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.1.attn.proj.weight',\n",
" 'sam_encoder.blocks.1.attn.proj.bias',\n",
" 'sam_encoder.blocks.1.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.1.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.1.norm2.weight',\n",
" 'sam_encoder.blocks.1.norm2.bias',\n",
" 'sam_encoder.blocks.1.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.1.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.1.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.1.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.1.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.1.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.1.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.1.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.2.norm1.weight',\n",
" 'sam_encoder.blocks.2.norm1.bias',\n",
" 'sam_encoder.blocks.2.attn.qkv.weight',\n",
" 'sam_encoder.blocks.2.attn.qkv.bias',\n",
" 'sam_encoder.blocks.2.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.2.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.2.attn.proj.weight',\n",
" 'sam_encoder.blocks.2.attn.proj.bias',\n",
" 'sam_encoder.blocks.2.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.2.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.2.norm2.weight',\n",
" 'sam_encoder.blocks.2.norm2.bias',\n",
" 'sam_encoder.blocks.2.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.2.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.2.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.2.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.2.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.2.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.2.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.2.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.3.norm1.weight',\n",
" 'sam_encoder.blocks.3.norm1.bias',\n",
" 'sam_encoder.blocks.3.attn.qkv.weight',\n",
" 'sam_encoder.blocks.3.attn.qkv.bias',\n",
" 'sam_encoder.blocks.3.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.3.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.3.attn.proj.weight',\n",
" 'sam_encoder.blocks.3.attn.proj.bias',\n",
" 'sam_encoder.blocks.3.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.3.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.3.norm2.weight',\n",
" 'sam_encoder.blocks.3.norm2.bias',\n",
" 'sam_encoder.blocks.3.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.3.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.3.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.3.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.3.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.3.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.3.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.3.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.4.norm1.weight',\n",
" 'sam_encoder.blocks.4.norm1.bias',\n",
" 'sam_encoder.blocks.4.attn.qkv.weight',\n",
" 'sam_encoder.blocks.4.attn.qkv.bias',\n",
" 'sam_encoder.blocks.4.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.4.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.4.attn.proj.weight',\n",
" 'sam_encoder.blocks.4.attn.proj.bias',\n",
" 'sam_encoder.blocks.4.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.4.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.4.norm2.weight',\n",
" 'sam_encoder.blocks.4.norm2.bias',\n",
" 'sam_encoder.blocks.4.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.4.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.4.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.4.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.4.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.4.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.4.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.4.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.5.norm1.weight',\n",
" 'sam_encoder.blocks.5.norm1.bias',\n",
" 'sam_encoder.blocks.5.attn.qkv.weight',\n",
" 'sam_encoder.blocks.5.attn.qkv.bias',\n",
" 'sam_encoder.blocks.5.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.5.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.5.attn.proj.weight',\n",
" 'sam_encoder.blocks.5.attn.proj.bias',\n",
" 'sam_encoder.blocks.5.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.5.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.5.norm2.weight',\n",
" 'sam_encoder.blocks.5.norm2.bias',\n",
" 'sam_encoder.blocks.5.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.5.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.5.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.5.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.5.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.5.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.5.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.5.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.6.norm1.weight',\n",
" 'sam_encoder.blocks.6.norm1.bias',\n",
" 'sam_encoder.blocks.6.attn.qkv.weight',\n",
" 'sam_encoder.blocks.6.attn.qkv.bias',\n",
" 'sam_encoder.blocks.6.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.6.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.6.attn.proj.weight',\n",
" 'sam_encoder.blocks.6.attn.proj.bias',\n",
" 'sam_encoder.blocks.6.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.6.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.6.norm2.weight',\n",
" 'sam_encoder.blocks.6.norm2.bias',\n",
" 'sam_encoder.blocks.6.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.6.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.6.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.6.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.6.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.6.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.6.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.6.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.7.norm1.weight',\n",
" 'sam_encoder.blocks.7.norm1.bias',\n",
" 'sam_encoder.blocks.7.attn.qkv.weight',\n",
" 'sam_encoder.blocks.7.attn.qkv.bias',\n",
" 'sam_encoder.blocks.7.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.7.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.7.attn.proj.weight',\n",
" 'sam_encoder.blocks.7.attn.proj.bias',\n",
" 'sam_encoder.blocks.7.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.7.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.7.norm2.weight',\n",
" 'sam_encoder.blocks.7.norm2.bias',\n",
" 'sam_encoder.blocks.7.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.7.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.7.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.7.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.7.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.7.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.7.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.7.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.8.norm1.weight',\n",
" 'sam_encoder.blocks.8.norm1.bias',\n",
" 'sam_encoder.blocks.8.attn.qkv.weight',\n",
" 'sam_encoder.blocks.8.attn.qkv.bias',\n",
" 'sam_encoder.blocks.8.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.8.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.8.attn.proj.weight',\n",
" 'sam_encoder.blocks.8.attn.proj.bias',\n",
" 'sam_encoder.blocks.8.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.8.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.8.norm2.weight',\n",
" 'sam_encoder.blocks.8.norm2.bias',\n",
" 'sam_encoder.blocks.8.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.8.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.8.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.8.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.8.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.8.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.8.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.8.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.9.norm1.weight',\n",
" 'sam_encoder.blocks.9.norm1.bias',\n",
" 'sam_encoder.blocks.9.attn.qkv.weight',\n",
" 'sam_encoder.blocks.9.attn.qkv.bias',\n",
" 'sam_encoder.blocks.9.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.9.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.9.attn.proj.weight',\n",
" 'sam_encoder.blocks.9.attn.proj.bias',\n",
" 'sam_encoder.blocks.9.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.9.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.9.norm2.weight',\n",
" 'sam_encoder.blocks.9.norm2.bias',\n",
" 'sam_encoder.blocks.9.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.9.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.9.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.9.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.9.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.9.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.9.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.9.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.10.norm1.weight',\n",
" 'sam_encoder.blocks.10.norm1.bias',\n",
" 'sam_encoder.blocks.10.attn.qkv.weight',\n",
" 'sam_encoder.blocks.10.attn.qkv.bias',\n",
" 'sam_encoder.blocks.10.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.10.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.10.attn.proj.weight',\n",
" 'sam_encoder.blocks.10.attn.proj.bias',\n",
" 'sam_encoder.blocks.10.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.10.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.10.norm2.weight',\n",
" 'sam_encoder.blocks.10.norm2.bias',\n",
" 'sam_encoder.blocks.10.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.10.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.10.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.10.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.10.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.10.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.10.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.10.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.blocks.11.norm1.weight',\n",
" 'sam_encoder.blocks.11.norm1.bias',\n",
" 'sam_encoder.blocks.11.attn.qkv.weight',\n",
" 'sam_encoder.blocks.11.attn.qkv.bias',\n",
" 'sam_encoder.blocks.11.attn.qkv.trainable_scale',\n",
" 'sam_encoder.blocks.11.attn.qkv.trainable_shift',\n",
" 'sam_encoder.blocks.11.attn.proj.weight',\n",
" 'sam_encoder.blocks.11.attn.proj.bias',\n",
" 'sam_encoder.blocks.11.attn.proj.trainable_scale',\n",
" 'sam_encoder.blocks.11.attn.proj.trainable_shift',\n",
" 'sam_encoder.blocks.11.norm2.weight',\n",
" 'sam_encoder.blocks.11.norm2.bias',\n",
" 'sam_encoder.blocks.11.mlp.lin1.weight',\n",
" 'sam_encoder.blocks.11.mlp.lin1.bias',\n",
" 'sam_encoder.blocks.11.mlp.lin1.trainable_scale',\n",
" 'sam_encoder.blocks.11.mlp.lin1.trainable_shift',\n",
" 'sam_encoder.blocks.11.mlp.lin2.weight',\n",
" 'sam_encoder.blocks.11.mlp.lin2.bias',\n",
" 'sam_encoder.blocks.11.mlp.lin2.trainable_scale',\n",
" 'sam_encoder.blocks.11.mlp.lin2.trainable_shift',\n",
" 'sam_encoder.neck.0.weight',\n",
" 'sam_encoder.neck.0.trainable_scale',\n",
" 'sam_encoder.neck.0.trainable_shift',\n",
" 'sam_encoder.neck.1.weight',\n",
" 'sam_encoder.neck.1.bias',\n",
" 'sam_encoder.neck.2.weight',\n",
" 'sam_encoder.neck.2.trainable_scale',\n",
" 'sam_encoder.neck.2.trainable_shift',\n",
" 'sam_encoder.neck.3.weight',\n",
" 'sam_encoder.neck.3.bias',\n",
" 'clip_model.positional_embedding',\n",
" 'clip_model.text_projection',\n",
" 'clip_model.logit_scale',\n",
" 'clip_model.visual.class_embedding',\n",
" 'clip_model.visual.positional_embedding',\n",
" 'clip_model.visual.proj',\n",
" 'clip_model.visual.conv1.weight',\n",
" 'clip_model.visual.ln_pre.weight',\n",
" 'clip_model.visual.ln_pre.bias',\n",
" 'clip_model.visual.transformer.resblocks.0.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.0.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.0.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.0.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.0.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.0.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.0.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.0.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.0.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.0.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.0.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.0.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.1.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.1.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.1.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.1.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.1.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.1.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.1.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.1.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.1.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.1.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.1.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.1.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.2.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.2.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.2.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.2.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.2.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.2.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.2.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.2.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.2.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.2.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.2.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.2.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.3.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.3.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.3.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.3.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.3.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.3.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.3.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.3.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.3.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.3.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.3.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.3.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.4.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.4.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.4.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.4.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.4.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.4.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.4.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.4.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.4.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.4.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.4.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.4.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.5.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.5.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.5.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.5.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.5.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.5.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.5.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.5.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.5.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.5.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.5.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.5.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.6.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.6.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.6.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.6.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.6.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.6.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.6.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.6.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.6.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.6.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.6.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.6.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.7.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.7.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.7.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.7.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.7.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.7.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.7.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.7.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.7.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.7.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.7.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.7.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.8.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.8.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.8.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.8.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.8.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.8.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.8.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.8.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.8.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.8.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.8.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.8.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.9.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.9.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.9.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.9.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.9.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.9.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.9.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.9.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.9.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.9.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.9.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.9.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.10.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.10.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.10.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.10.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.10.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.10.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.10.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.10.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.10.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.10.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.10.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.10.ln_2.bias',\n",
" 'clip_model.visual.transformer.resblocks.11.attn.in_proj_weight',\n",
" 'clip_model.visual.transformer.resblocks.11.attn.in_proj_bias',\n",
" 'clip_model.visual.transformer.resblocks.11.attn.out_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.11.attn.out_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.11.ln_1.weight',\n",
" 'clip_model.visual.transformer.resblocks.11.ln_1.bias',\n",
" 'clip_model.visual.transformer.resblocks.11.mlp.c_fc.weight',\n",
" 'clip_model.visual.transformer.resblocks.11.mlp.c_fc.bias',\n",
" 'clip_model.visual.transformer.resblocks.11.mlp.c_proj.weight',\n",
" 'clip_model.visual.transformer.resblocks.11.mlp.c_proj.bias',\n",
" 'clip_model.visual.transformer.resblocks.11.ln_2.weight',\n",
" 'clip_model.visual.transformer.resblocks.11.ln_2.bias',\n",
" 'clip_model.visual.ln_post.weight',\n",
" 'clip_model.visual.ln_post.bias',\n",
" 'clip_model.transformer.resblocks.0.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.0.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.0.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.0.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.0.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.0.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.0.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.0.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.0.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.0.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.0.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.0.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.1.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.1.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.1.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.1.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.1.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.1.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.1.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.1.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.1.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.1.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.1.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.1.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.2.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.2.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.2.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.2.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.2.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.2.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.2.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.2.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.2.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.2.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.2.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.2.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.3.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.3.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.3.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.3.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.3.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.3.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.3.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.3.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.3.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.3.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.3.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.3.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.4.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.4.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.4.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.4.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.4.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.4.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.4.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.4.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.4.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.4.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.4.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.4.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.5.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.5.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.5.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.5.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.5.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.5.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.5.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.5.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.5.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.5.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.5.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.5.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.6.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.6.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.6.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.6.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.6.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.6.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.6.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.6.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.6.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.6.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.6.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.6.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.7.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.7.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.7.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.7.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.7.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.7.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.7.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.7.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.7.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.7.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.7.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.7.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.8.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.8.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.8.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.8.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.8.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.8.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.8.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.8.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.8.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.8.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.8.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.8.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.9.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.9.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.9.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.9.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.9.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.9.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.9.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.9.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.9.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.9.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.9.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.9.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.10.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.10.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.10.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.10.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.10.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.10.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.10.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.10.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.10.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.10.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.10.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.10.ln_2.bias',\n",
" 'clip_model.transformer.resblocks.11.attn.in_proj_weight',\n",
" 'clip_model.transformer.resblocks.11.attn.in_proj_bias',\n",
" 'clip_model.transformer.resblocks.11.attn.out_proj.weight',\n",
" 'clip_model.transformer.resblocks.11.attn.out_proj.bias',\n",
" 'clip_model.transformer.resblocks.11.ln_1.weight',\n",
" 'clip_model.transformer.resblocks.11.ln_1.bias',\n",
" 'clip_model.transformer.resblocks.11.mlp.c_fc.weight',\n",
" 'clip_model.transformer.resblocks.11.mlp.c_fc.bias',\n",
" 'clip_model.transformer.resblocks.11.mlp.c_proj.weight',\n",
" 'clip_model.transformer.resblocks.11.mlp.c_proj.bias',\n",
" 'clip_model.transformer.resblocks.11.ln_2.weight',\n",
" 'clip_model.transformer.resblocks.11.ln_2.bias',\n",
" 'clip_model.token_embedding.weight',\n",
" 'clip_model.ln_final.weight',\n",
" 'clip_model.ln_final.bias',\n",
" 'prompt_encoder.pe_layer.positional_encoding_gaussian_matrix',\n",
" 'prompt_encoder.point_embeddings.0.weight',\n",
" 'prompt_encoder.point_embeddings.1.weight',\n",
" 'prompt_encoder.point_embeddings.2.weight',\n",
" 'prompt_encoder.point_embeddings.3.weight',\n",
" 'prompt_encoder.not_a_point_embed.weight',\n",
" 'prompt_encoder.mask_downscaling.0.weight',\n",
" 'prompt_encoder.mask_downscaling.0.bias',\n",
" 'prompt_encoder.mask_downscaling.1.weight',\n",
" 'prompt_encoder.mask_downscaling.1.bias',\n",
" 'prompt_encoder.mask_downscaling.3.weight',\n",
" 'prompt_encoder.mask_downscaling.3.bias',\n",
" 'prompt_encoder.mask_downscaling.4.weight',\n",
" 'prompt_encoder.mask_downscaling.4.bias',\n",
" 'prompt_encoder.mask_downscaling.6.weight',\n",
" 'prompt_encoder.mask_downscaling.6.bias',\n",
" 'prompt_encoder.no_mask_embed.weight',\n",
" 'mask_decoder.transformer.layers.0.self_attn.q_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.self_attn.q_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.self_attn.k_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.self_attn.k_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.self_attn.v_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.self_attn.v_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.self_attn.out_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.self_attn.out_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.norm1.weight',\n",
" 'mask_decoder.transformer.layers.0.norm1.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.q_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.q_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.k_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.k_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.v_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.v_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.out_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_token_to_image.out_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.norm2.weight',\n",
" 'mask_decoder.transformer.layers.0.norm2.bias',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin1.weight',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin1.bias',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin1.trainable_scale',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin1.trainable_shift',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin2.weight',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin2.bias',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin2.trainable_scale',\n",
" 'mask_decoder.transformer.layers.0.mlp.lin2.trainable_shift',\n",
" 'mask_decoder.transformer.layers.0.norm3.weight',\n",
" 'mask_decoder.transformer.layers.0.norm3.bias',\n",
" 'mask_decoder.transformer.layers.0.norm4.weight',\n",
" 'mask_decoder.transformer.layers.0.norm4.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.q_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.q_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.k_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.k_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.v_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.v_proj.bias',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.out_proj.weight',\n",
" 'mask_decoder.transformer.layers.0.cross_attn_image_to_token.out_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.self_attn.q_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.self_attn.q_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.self_attn.k_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.self_attn.k_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.self_attn.v_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.self_attn.v_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.self_attn.out_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.self_attn.out_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.norm1.weight',\n",
" 'mask_decoder.transformer.layers.1.norm1.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.q_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.q_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.k_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.k_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.v_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.v_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.out_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_token_to_image.out_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.norm2.weight',\n",
" 'mask_decoder.transformer.layers.1.norm2.bias',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin1.weight',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin1.bias',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin1.trainable_scale',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin1.trainable_shift',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin2.weight',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin2.bias',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin2.trainable_scale',\n",
" 'mask_decoder.transformer.layers.1.mlp.lin2.trainable_shift',\n",
" 'mask_decoder.transformer.layers.1.norm3.weight',\n",
" 'mask_decoder.transformer.layers.1.norm3.bias',\n",
" 'mask_decoder.transformer.layers.1.norm4.weight',\n",
" 'mask_decoder.transformer.layers.1.norm4.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.q_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.q_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.k_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.k_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.v_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.v_proj.bias',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.out_proj.weight',\n",
" 'mask_decoder.transformer.layers.1.cross_attn_image_to_token.out_proj.bias',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.q_proj.weight',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.q_proj.bias',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.k_proj.weight',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.k_proj.bias',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.v_proj.weight',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.v_proj.bias',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.out_proj.weight',\n",
" 'mask_decoder.transformer.final_attn_token_to_image.out_proj.bias',\n",
" 'mask_decoder.transformer.norm_final_attn.weight',\n",
" 'mask_decoder.transformer.norm_final_attn.bias',\n",
" 'mask_decoder.iou_token.weight',\n",
" 'mask_decoder.mask_tokens.weight',\n",
" 'mask_decoder.output_upscaling.0.weight',\n",
" 'mask_decoder.output_upscaling.0.bias',\n",
" 'mask_decoder.output_upscaling.1.weight',\n",
" 'mask_decoder.output_upscaling.1.bias',\n",
" 'mask_decoder.output_upscaling.3.weight',\n",
" 'mask_decoder.output_upscaling.3.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.0.layers.0.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.0.layers.0.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.0.layers.1.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.0.layers.1.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.0.layers.2.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.0.layers.2.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.1.layers.0.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.1.layers.0.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.1.layers.1.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.1.layers.1.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.1.layers.2.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.1.layers.2.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.2.layers.0.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.2.layers.0.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.2.layers.1.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.2.layers.1.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.2.layers.2.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.2.layers.2.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.3.layers.0.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.3.layers.0.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.3.layers.1.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.3.layers.1.bias',\n",
" 'mask_decoder.output_hypernetworks_mlps.3.layers.2.weight',\n",
" 'mask_decoder.output_hypernetworks_mlps.3.layers.2.bias',\n",
" 'mask_decoder.iou_prediction_head.layers.0.weight',\n",
" 'mask_decoder.iou_prediction_head.layers.0.bias',\n",
" 'mask_decoder.iou_prediction_head.layers.1.weight',\n",
" 'mask_decoder.iou_prediction_head.layers.1.bias',\n",
" 'mask_decoder.iou_prediction_head.layers.2.weight',\n",
" 'mask_decoder.iou_prediction_head.layers.2.bias',\n",
" 'Text_Embedding_Affine.0.weight',\n",
" 'Text_Embedding_Affine.0.bias',\n",
" 'Text_Embedding_Affine.2.weight',\n",
" 'Text_Embedding_Affine.2.bias',\n",
" 'Text_Embedding_Affine.2.running_mean',\n",
" 'Text_Embedding_Affine.2.running_var',\n",
" 'Text_Embedding_Affine.2.num_batches_tracked']"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"keys"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"def basic_stats(key):\n",
" print('min: ', model_dict[key].min())\n",
" print('max: ', model_dict[key].max())\n",
" print('mean: ', model_dict[key].mean())\n",
" print('norm: ', np.linalg.norm(model_dict[key].numpy()))\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"min: tensor(0.)\n",
"max: tensor(0.)\n",
"mean: tensor(0.)\n",
"norm: 0.0\n"
]
}
],
"source": [
"basic_stats('mask_decoder.transformer.layers.0.mlp.lin1.trainable_shift')"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'a.b.c.d'"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"a = \"a.b.c.d\"\n",
"a.replace(\"c\",\"e\")\n",
"a"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"import nibabel as nib\n",
"import numpy as np\n",
"import os"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['case0005_slice018.npz',\n",
" 'case0027_slice043.npz',\n",
" 'case0027_slice044.npz',\n",
" 'case0027_slice045.npz',\n",
" 'case0027_slice046.npz',\n",
" 'case0027_slice047.npz',\n",
" 'case0027_slice048.npz',\n",
" 'case0027_slice049.npz',\n",
" 'case0027_slice050.npz',\n",
" 'case0027_slice051.npz',\n",
" 'case0027_slice052.npz',\n",
" 'case0027_slice053.npz',\n",
" 'case0027_slice054.npz',\n",
" 'case0027_slice055.npz',\n",
" 'case0027_slice056.npz',\n",
" 'case0027_slice057.npz',\n",
" 'case0027_slice058.npz',\n",
" 'case0027_slice059.npz',\n",
" 'case0027_slice060.npz',\n",
" 'case0027_slice061.npz',\n",
" 'case0027_slice062.npz',\n",
" 'case0027_slice063.npz',\n",
" 'case0027_slice064.npz',\n",
" 'case0027_slice065.npz',\n",
" 'case0027_slice066.npz',\n",
" 'case0027_slice067.npz',\n",
" 'case0027_slice068.npz',\n",
" 'case0010_slice080.npz',\n",
" 'case0010_slice081.npz',\n",
" 'case0010_slice082.npz',\n",
" 'case0010_slice083.npz',\n",
" 'case0010_slice084.npz']"
]
},
"execution_count": 18,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"os.listdir('/media/ubuntu/New Volume/jay/BTCV/train_npz')[18:50]"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"# im = nib.load('/media/ubuntu/New Volume/jay/LiTS/images/1.nii')\n",
"im = np.load('/media/ubuntu/New Volume/jay/BTCV/train_npz/case0005_slice000.npz')"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(512, 512)\n"
]
}
],
"source": [
"list(im.keys())\n",
"print(im['image'].shape)"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import torch"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(tensor([0, 0, 1, 1]), tensor([0, 4, 3, 4]))"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"a = torch.Tensor([\n",
" [1,0,0,0,1],\n",
" [0,0,0,1,1]\n",
"])\n",
"torch.where(a==1)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"tensor([[315., 581.],\n",
" [339., 340.]])\n"
]
}
],
"source": [
"a = [[torch.tensor(315), torch.tensor(581)], [torch.tensor(339), torch.tensor(340)]]\n",
"print(torch.Tensor(a))"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3.8.16 ('dassl')",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.16"
},
"orig_nbformat": 4,
"vscode": {
"interpreter": {
"hash": "5b0d24c0401191df5ff06ef3cb04a21077c1fd7ca08d243336ea8a8a1206ff02"
}
}
},
"nbformat": 4,
"nbformat_minor": 2
}
|