python-doctr 0.12.0__py3-none-any.whl → 1.0.1__py3-none-any.whl

python_doctr-1.0.1.dist-info/RECORD

@@ -0,0 +1,149 @@
+doctr/__init__.py,sha256=sdqGeYFfPLRsRH54PsedllScz5FD8yWwyekcsOq3JNc,110
+doctr/file_utils.py,sha256=DsQvazaicyDSM1bNuI70p1DNewO0IOxQ5TRBgXSju3w,999
+doctr/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+doctr/version.py,sha256=6p2_N2LdNEx0s2rC1vd5LYuVCJ2ya740g-riPQhSMfQ,23
+doctr/contrib/__init__.py,sha256=EKeAGa3TOuJVWR4H_HJvuKO5VPEnJWXW305JMw3ufic,40
+doctr/contrib/artefacts.py,sha256=fP-UMWBYVibCv0aMHXFd-D3HF4yQIbaWfV0EDl4_iAU,5291
+doctr/contrib/base.py,sha256=uKV-B_9Y02jR_X6of55N3aJCcn8e1asMTHybsTQzRbw,3273
+doctr/datasets/__init__.py,sha256=QzWe91FU1d3Vi5h4ayPtJUE-Y_SSs_4Xp6MAkZ8XKTc,504
+doctr/datasets/coco_text.py,sha256=3-D4EjqOdWF-ioMXgdNj_plYWGrIa-gZt4FasphXAGA,5496
+doctr/datasets/cord.py,sha256=uRGYIDtRsq01UTRShLKEUyOeWphydHyMQA9Rows-1z0,5315
+doctr/datasets/detection.py,sha256=jeTNggksKEYsVF7_Km1LSJNc7VHTcES4BLuYt_UhXpg,3531
+doctr/datasets/doc_artefacts.py,sha256=tYTdwfC5d2f_wkxhJ-2-XMf7Bk8pS4LQ_zLrk-JTVvQ,3230
+doctr/datasets/funsd.py,sha256=siX7xUP38dHT3ECL3ReQ78v4RI-q0mwcuaX1TADkwzI,4733
+doctr/datasets/ic03.py,sha256=l_f0ACbR9DD2_mCe57eLO8dnwFgb2JNP0jjpPmvXCP4,5564
+doctr/datasets/ic13.py,sha256=EWv67V5TwkKcgOzE9qCz38Rnff6b5Bqo7FwXNECDy-I,4528
+doctr/datasets/iiit5k.py,sha256=Yz3brhWUVCqXhAAg75xRU3qWJeX87j2lEG70zkkMrlo,4583
+doctr/datasets/iiithws.py,sha256=8iMytN32T23BpIUMoAEzHrsUgzII_thjoMxHzAEu-nk,2768
+doctr/datasets/imgur5k.py,sha256=MJwrMlBFsxlfJs65rgkKzRjhepNf04kez7kP90Tlt74,7525
+doctr/datasets/mjsynth.py,sha256=SBlhKk7s1onMXF8AJj2WMnIF2n5XXBzTJTvOZxcW0TI,4075
+doctr/datasets/ocr.py,sha256=61SxWHtDOty7ZtkAkuiPAw68n5wZYXI2YZJxP4LhrHc,2523
+doctr/datasets/orientation.py,sha256=b5XsuSWChDxmU_zJ33LtBz76_Rkz5N6r6NogW9uAMZc,1091
+doctr/datasets/recognition.py,sha256=FdaTGAoRq0VD5g7oo8NKphsI2YU8fSZzSEj9IppiT-Y,1858
+doctr/datasets/sroie.py,sha256=9moG-O_PiFU0uXuYzsbzb-TxXKwCOK0NSyRSde0FlT4,4430
+doctr/datasets/svhn.py,sha256=duB_HMWRLY3FKPEgMt2x9k3x4qEjCvbZftcsfuayRWE,5784
+doctr/datasets/svt.py,sha256=B_39RFw3qWYWqGbpVcd51SyhUwi2vjBSCFQKhE1PpiA,5046
+doctr/datasets/synthtext.py,sha256=F_5BS1vJ5b194kRaYpofsSWWIQqBoon9URF742jME5w,6301
+doctr/datasets/utils.py,sha256=rt5-1B_ZbqPpZmfTRAMCqaUPpi3d2O5vfGfiVWkzp04,8106
+doctr/datasets/vocabs.py,sha256=vKGhlVxA8KuJz1X15QcX2i8AwLT1u0eHxvQSUp0zFVA,99032
+doctr/datasets/wildreceipt.py,sha256=vqGYVaaSS97GxSHuvYvpeCIShACAEYhgScJhpA943R0,5213
+doctr/datasets/datasets/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/datasets/datasets/base.py,sha256=hyNTt3HCvMpeRA1rYMneLhAbeSTvAbC5JE8MKQBFk5g,4826
+doctr/datasets/datasets/pytorch.py,sha256=U9xtyOSgXLnlGu9vG9ECvp0cf1SvXZ2SABVk2v0cp6g,2030
+doctr/datasets/generator/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/datasets/generator/base.py,sha256=xiVAvAzP_r4fTpZy75O_daN93fzsjjpaN-vxn-GtyLU,5706
+doctr/datasets/generator/pytorch.py,sha256=vvpbdu6PgDLbUkKmXYqZYPsoGaRQIV2NqAS7t99jP7E,2116
+doctr/io/__init__.py,sha256=kS7tKGFvzxOCWBOun-Y8n9CsziwRKNynjwpZEUUI03M,106
+doctr/io/elements.py,sha256=74pRYfxbe4yCB8S1It-PymGWbDG0S4qdbjf_P57_M0s,26474
+doctr/io/html.py,sha256=gKu4hxMYaEgRDPcd45HBlctNXw0Fnk3CY5Jgr9JZRMw,698
+doctr/io/pdf.py,sha256=LYxPh7auAIlPh3Xdh462peVnBbCMMjB7ouDfiQq7eok,1308
+doctr/io/reader.py,sha256=gwQmziIkF-cozyE8K4KisZYmFmtyoZUSl4LWyI_CPCI,2732
+doctr/io/image/__init__.py,sha256=bJLj2I8OOTYLuTDjdinao0nkOIWQOLbzIuww23EX3gw,43
+doctr/io/image/base.py,sha256=EgAl_IaklfK67h4nY9nhedTMu1_MUqIqzNcJSH1IGzs,1681
+doctr/io/image/pytorch.py,sha256=UYlll8CPBaOZ7tDyWN-xQ2DMVqGRnBI173Q4NF8ij7A,3201
+doctr/models/__init__.py,sha256=yn_mXUL8B5L27Uaat1rLGRQHgLR8VLVxzBuPfNuN1YE,124
+doctr/models/_utils.py,sha256=U66VEsTHHmJEgauqX_t6tl6fRUTgWiDPKmDL4i-Ptu0,8455
+doctr/models/builder.py,sha256=LkleWWqwhkFdBMMAFWD9TWGDKXSIEx-cn3hi1CToIxc,20370
+doctr/models/core.py,sha256=KPTn_zbiwTV7OoPDv6o-xx7Tzb-QnAfLne1HuNQlWu4,482
+doctr/models/zoo.py,sha256=uPgHXZP_wcRaZXmwZyIzAcLaDHhzo-B_8a0cx64mOVk,9276
+doctr/models/classification/__init__.py,sha256=piTKyOGbTiKBzZSAHQLXf622mqjxaAJPicNtH3b-A0k,173
+doctr/models/classification/zoo.py,sha256=dXgrlE9V7-zNM6yd0hfgpcpTQpSED9X7KuR2jGbjcd0,4265
+doctr/models/classification/magc_resnet/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/magc_resnet/pytorch.py,sha256=yVnkta1vNSXB5tF8C4TLMeAptWjrbNGbtDmQ4z_lJ6w,5480
+doctr/models/classification/mobilenet/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/mobilenet/pytorch.py,sha256=33ou3bbirjFhSUe-ws9C6foPosiO_cse-RgDBfxYL0Q,9826
+doctr/models/classification/predictor/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/predictor/pytorch.py,sha256=bu2dWcwrOBD_0XloiNXzyJ7JrjKiczRv0YWjh46iHbQ,2523
+doctr/models/classification/resnet/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/resnet/pytorch.py,sha256=nmD1sg9-Vd5u_NRaVaES5lENi4d06jcy00TSl82RT50,13250
+doctr/models/classification/textnet/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/textnet/pytorch.py,sha256=8a4i33cNgb5zdGjJa8nDgZfHyEMDa4mvtz1rJ153PPM,10424
+doctr/models/classification/vgg/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/vgg/pytorch.py,sha256=1fO451jkdTLMnc4rDEkvsvkYVapyfK19TKEE0mHoPHY,3679
+doctr/models/classification/vip/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/vip/pytorch.py,sha256=4IXmfXOil4tEp-g5VmoQIbXWlud2aFXsc0wuIDrGG58,16134
+doctr/models/classification/vip/layers/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/vip/layers/pytorch.py,sha256=lK01uxaC6QAbTf3DkgxEaYwNFhcPk7rT9TPs6FEkUI4,21019
+doctr/models/classification/vit/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/classification/vit/pytorch.py,sha256=rrBLNjHvpmxqsNwSZenhrX81u4SBeUiGenFYVQKuaWc,6371
+doctr/models/detection/__init__.py,sha256=RqSz5beehLiqhW0PwFLFmCfTyMjofO-0umcQJLDMHjY,105
+doctr/models/detection/core.py,sha256=3qHxwPRXD_Pcpk3uvUtoCXtW07xagLOjbvsVFzRli7g,3444
+doctr/models/detection/zoo.py,sha256=fmTcKXfOioUk9faOJSnr_1_c_Xviv1JAoPs0FMhZKng,3671
+doctr/models/detection/_utils/__init__.py,sha256=bJLj2I8OOTYLuTDjdinao0nkOIWQOLbzIuww23EX3gw,43
+doctr/models/detection/_utils/base.py,sha256=1Ozf6nQoXt_OhKEt-QPhP1iJ0HJoMugbdDFNYK2wHG8,2475
+doctr/models/detection/_utils/pytorch.py,sha256=NyQxrS4e_EVQ5FriZSQYbvTvtenmAsvSC5QGFbZUhV4,1021
+doctr/models/detection/differentiable_binarization/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/detection/differentiable_binarization/base.py,sha256=gMfQBrpuxflQUF46CyBFA1WNztU11fdkQxL2J3M3CIQ,15788
+doctr/models/detection/differentiable_binarization/pytorch.py,sha256=sioxicRYT_kr5BsqzfdH6Pa2GTPtmycA2LzQrurKvC0,16453
+doctr/models/detection/fast/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/detection/fast/base.py,sha256=BQ1Z9ARWakLbTH0VtHjUvmKZKhITFG-RTZkGKucjCyA,10225
+doctr/models/detection/fast/pytorch.py,sha256=fzUoDDPrrKS9H_WyO8jcPdUZrrqVMZHL4d04DIJq_yY,16702
+doctr/models/detection/linknet/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/detection/linknet/base.py,sha256=EKU5rPreZ6HO3brR93R8x5ssBmEu1ynbtc8iffHgQ0g,10007
+doctr/models/detection/linknet/pytorch.py,sha256=KrKuZj16Xiwgk89oSaaVWSgz9jItvkCVv20rjoaqfDw,14333
+doctr/models/detection/predictor/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/detection/predictor/pytorch.py,sha256=jdU5A10FDBXMLXyu1zJsVxlqSmt860u-oT30H7cIoKM,2627
+doctr/models/factory/__init__.py,sha256=cKPoH2V2157lLMTR2zsljG3_IQHziodqR-XK_LG0D_I,19
+doctr/models/factory/hub.py,sha256=F-CG-9dG3Iqbsj2mKkxf0JqKuZbg4hOfDoFGTO6fAIs,6889
+doctr/models/kie_predictor/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/kie_predictor/base.py,sha256=oLVDjIZ5sKq4ma5bopWVYUzq6q4ZAxF6mtEEQNxwBxk,2291
+doctr/models/kie_predictor/pytorch.py,sha256=pjFtZcjDDO9mpC7Zscu9CYcmkfF87J1ioJ8N_dzUSq4,7704
+doctr/models/modules/__init__.py,sha256=pouP7obVTu4p6aHkyaqa1yHKbynpvT0Hgo-LO_1U2R4,83
+doctr/models/modules/layers/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/modules/layers/pytorch.py,sha256=tjTZYcQt30TSpFwpzBfwEc7QN7bQdxnYMGOcBI4MqwE,8667
+doctr/models/modules/transformer/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/modules/transformer/pytorch.py,sha256=AVFeQCElFgg7ondhTYc_XFPocTKMt3uFJBHUFSEUevk,7657
+doctr/models/modules/vision_transformer/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/modules/vision_transformer/pytorch.py,sha256=ltnBAevcfJl612yblI6WhcJAPEgUO0w-5jFL7vBhXoA,3943
+doctr/models/predictor/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/predictor/base.py,sha256=YYvEepdYY7bphhOmBdFsVGFI6EaBZDip9hut8qz2uA8,8541
+doctr/models/predictor/pytorch.py,sha256=XplCn6sCdeREsINz6foxqLVTTJ0jiO2vMJgcVF0_Dxw,6241
+doctr/models/preprocessor/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/preprocessor/pytorch.py,sha256=vcMwCb04UlJdCJ5vh-ZXPzoy_-U0sPOn-OWgREvL42g,4395
+doctr/models/recognition/__init__.py,sha256=bgAvbwjO14Z2RQFD2XKZcSYJNsyxPa96SGHmX_nqbAQ,145
+doctr/models/recognition/core.py,sha256=InV-HLYkbQG65B6BYIud_JCWdbeoacC2MURjPZqceoo,1524
+doctr/models/recognition/utils.py,sha256=WLUr6Qa3Kq803xyuLWmlFBS2Vf3P5GntyZVXW-y1LqU,3739
+doctr/models/recognition/zoo.py,sha256=SeJr1IqZsFXzRGYrExgFb2aDyWp0fZWhR6Rr1RJTQHQ,2991
+doctr/models/recognition/crnn/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/crnn/pytorch.py,sha256=jy5LVkG0NdLObFIaW2lWsdv6LWYYLIkZDZS-aTjFa6I,12234
+doctr/models/recognition/master/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/master/base.py,sha256=KU94CO3YZYyYh83gW6F65IjeJfoAkwuandY3A_DxfM8,1471
+doctr/models/recognition/master/pytorch.py,sha256=mL9o3O_L-iVw-Wq71tF-Y3a_G6qoGlUv2_3SvcWBu2I,12756
+doctr/models/recognition/parseq/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/parseq/base.py,sha256=SMGEETj94ZIJUdSLe2d2PhLTpeptrW5MB7cW3WtVOhQ,1465
+doctr/models/recognition/parseq/pytorch.py,sha256=hI_6qBDrKqe4358TKKz0YzXm06K9FKyGn2_6s8caHf8,20801
+doctr/models/recognition/predictor/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/predictor/_utils.py,sha256=8q5ate9L7xeqOY-45foX1itPlHUBWw3RpLfCgYJ2ZJo,4695
+doctr/models/recognition/predictor/pytorch.py,sha256=amUzwltBWhQNaCKB2hjqsQuRHtxfeknZO1uwSX3ASzY,2753
+doctr/models/recognition/sar/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/sar/pytorch.py,sha256=_YMeCgu7p8-v0N75AiT6dR2O3ww7UFT8E82IWdF9lQM,15639
+doctr/models/recognition/viptr/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/viptr/pytorch.py,sha256=IQQeuVyBbknN0BbEYy2Kak8WeBU4dDcpP0wUvjPO330,9355
+doctr/models/recognition/vitstr/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/recognition/vitstr/base.py,sha256=tRMhC5i_l0S207_YFEs7wFAtkznyAJoH8sUZVMcZUBk,1419
+doctr/models/recognition/vitstr/pytorch.py,sha256=UMXLvGHi065j3XQxbVwrqN2AL3YShfJ1pl-sKkRU0Cs,10080
+doctr/models/utils/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/models/utils/pytorch.py,sha256=BHNEDG89XSW9TYNRwtGuf2RcpM2OfjnBb_92-OpYRr8,5804
+doctr/transforms/__init__.py,sha256=0VHbvUZ7llFd1e_7_JdWTaxYMCvaR6KbUJaYJequmQI,23
+doctr/transforms/functional/__init__.py,sha256=hxLu7FmRYn01ysoOQalrMa5K_qg4rLNf8OAuoemzd6I,23
+doctr/transforms/functional/base.py,sha256=LRIM3L6nZQaZ3qyQKEFQ0ZLJRRg7rMYtshADK_ZCOi8,6851
+doctr/transforms/functional/pytorch.py,sha256=mSToyPGn8O3WgK9rQJoGZcOHyIm06MQtWua1YA7p1kU,5102
+doctr/transforms/modules/__init__.py,sha256=bJLj2I8OOTYLuTDjdinao0nkOIWQOLbzIuww23EX3gw,43
+doctr/transforms/modules/base.py,sha256=fAJzcvq6rZv9oj8NRZ3Veo9z2-wFR40w3JjNmIdsP7w,7679
+doctr/transforms/modules/pytorch.py,sha256=XTUbZPC3aahoB0cqwIPlLtsG8dbc7eReG7SAQsjYWJg,11802
+doctr/utils/__init__.py,sha256=uQY9ibZ24V896fmihIsK23QOIZdKtk0HyKoCVJ_lLuM,95
+doctr/utils/common_types.py,sha256=qK-gSvbAGJAhtqTUCO1DJ3Vr04KVFUW5I3Pn0T4Btp8,534
+doctr/utils/data.py,sha256=eXLIuO5TJk6hjlVtodADQFKkt-dlEmLjEhIQcFVF2Zs,4174
+doctr/utils/fonts.py,sha256=fxEZ8o67iyceZpiw87SskP6v8lcC5iHVcfWLPo_m6rQ,1265
+doctr/utils/geometry.py,sha256=y4Xz-heV_sA7OfqoP40vHzJe4bKAgmzIBEA5FvV7Pgc,18678
+doctr/utils/metrics.py,sha256=7u2Fz7KnY32PlV-Lo9M9iYKfHHMXEE6UdSTq-rgJAvU,20249
+doctr/utils/multithreading.py,sha256=Z7cz0M-Zsqq5jRMrAW4Q5L7n2pNvC5qPKOfkXX4w05A,1972
+doctr/utils/reconstitution.py,sha256=i6wqDQmCmu5WPRcIGFxGvGegy0mUv42lLK90DKUFtjc,7256
+doctr/utils/repr.py,sha256=ShDOvCiPjdj1_v1AfLQ7AN_tunu07p0S_scFZYip7GY,2087
+doctr/utils/visualization.py,sha256=sAB283iegS86gq0btK9H6PFLxP7SPXpuX9JWT_v53Iw,13118
+python_doctr-1.0.1.dist-info/licenses/LICENSE,sha256=75RTSsXOsAYhGpxsHc9U41ep6GS7vrUPufeekgoeOXM,11336
+python_doctr-1.0.1.dist-info/METADATA,sha256=gOwRkDU1ufra-cglBOjbW0OpxTwHBvVgFFM9YY_6iMM,32370
+python_doctr-1.0.1.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+python_doctr-1.0.1.dist-info/top_level.txt,sha256=lCgp4pmjPI3HYph62XhfzA3jRwM715kGtJPmqIUJ9t8,6
+python_doctr-1.0.1.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+python_doctr-1.0.1.dist-info/RECORD,,

{python_doctr-0.12.0.dist-info → python_doctr-1.0.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.9.0)
+Generator: setuptools (80.10.2)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

doctr/datasets/datasets/tensorflow.py

@@ -1,59 +0,0 @@
-# Copyright (C) 2021-2025, Mindee.
-
-# This program is licensed under the Apache License 2.0.
-# See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
-
-import os
-from copy import deepcopy
-from typing import Any
-
-import numpy as np
-import tensorflow as tf
-
-from doctr.io import read_img_as_tensor, tensor_from_numpy
-
-from .base import _AbstractDataset, _VisionDataset
-
-__all__ = ["AbstractDataset", "VisionDataset"]
-
-
-class AbstractDataset(_AbstractDataset):
-    """Abstract class for all datasets"""
-
-    def _read_sample(self, index: int) -> tuple[tf.Tensor, Any]:
-        img_name, target = self.data[index]
-
-        # Check target
-        if isinstance(target, dict):
-            assert "boxes" in target, "Target should contain 'boxes' key"
-            assert "labels" in target, "Target should contain 'labels' key"
-        elif isinstance(target, tuple):
-            assert len(target) == 2
-            assert isinstance(target[0], str) or isinstance(target[0], np.ndarray), (
-                "first element of the tuple should be a string or a numpy array"
-            )
-            assert isinstance(target[1], list), "second element of the tuple should be a list"
-        else:
-            assert isinstance(target, str) or isinstance(target, np.ndarray), (
-                "Target should be a string or a numpy array"
-            )
-
-        # Read image
-        img = (
-            tensor_from_numpy(img_name, dtype=tf.float32)
-            if isinstance(img_name, np.ndarray)
-            else read_img_as_tensor(os.path.join(self.root, img_name), dtype=tf.float32)
-        )
-
-        return img, deepcopy(target)
-
-    @staticmethod
-    def collate_fn(samples: list[tuple[tf.Tensor, Any]]) -> tuple[tf.Tensor, list[Any]]:
-        images, targets = zip(*samples)
-        images = tf.stack(images, axis=0)
-
-        return images, list(targets)
-
-
-class VisionDataset(AbstractDataset, _VisionDataset):  # noqa: D101
-    pass

doctr/datasets/generator/tensorflow.py

@@ -1,58 +0,0 @@
-# Copyright (C) 2021-2025, Mindee.
-
-# This program is licensed under the Apache License 2.0.
-# See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
-
-import tensorflow as tf
-
-from .base import _CharacterGenerator, _WordGenerator
-
-__all__ = ["CharacterGenerator", "WordGenerator"]
-
-
-class CharacterGenerator(_CharacterGenerator):
-    """Implements a character image generation dataset
-
-    >>> from doctr.datasets import CharacterGenerator
-    >>> ds = CharacterGenerator(vocab='abdef', num_samples=100)
-    >>> img, target = ds[0]
-
-    Args:
-        vocab: vocabulary to take the character from
-        num_samples: number of samples that will be generated iterating over the dataset
-        cache_samples: whether generated images should be cached firsthand
-        font_family: font to use to generate the text images
-        img_transforms: composable transformations that will be applied to each image
-        sample_transforms: composable transformations that will be applied to both the image and the target
-    """
-
-    def __init__(self, *args, **kwargs) -> None:
-        super().__init__(*args, **kwargs)
-
-    @staticmethod
-    def collate_fn(samples):
-        images, targets = zip(*samples)
-        images = tf.stack(images, axis=0)
-
-        return images, tf.convert_to_tensor(targets)
-
-
-class WordGenerator(_WordGenerator):
-    """Implements a character image generation dataset
-
-    >>> from doctr.datasets import WordGenerator
-    >>> ds = WordGenerator(vocab='abdef', min_chars=1, max_chars=32, num_samples=100)
-    >>> img, target = ds[0]
-
-    Args:
-        vocab: vocabulary to take the character from
-        min_chars: minimum number of characters in a word
-        max_chars: maximum number of characters in a word
-        num_samples: number of samples that will be generated iterating over the dataset
-        cache_samples: whether generated images should be cached firsthand
-        font_family: font to use to generate the text images
-        img_transforms: composable transformations that will be applied to each image
-        sample_transforms: composable transformations that will be applied to both the image and the target
-    """
-
-    pass

doctr/datasets/loader.py

@@ -1,94 +0,0 @@
-# Copyright (C) 2021-2025, Mindee.
-
-# This program is licensed under the Apache License 2.0.
-# See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
-
-import math
-from collections.abc import Callable
-
-import numpy as np
-import tensorflow as tf
-
-__all__ = ["DataLoader"]
-
-
-def default_collate(samples):
-    """Collate multiple elements into batches
-
-    Args:
-        samples: list of N tuples containing M elements
-
-    Returns:
-        tuple of M sequences containing N elements each
-    """
-    batch_data = zip(*samples)
-
-    tf_data = tuple(tf.stack(elt, axis=0) for elt in batch_data)
-
-    return tf_data
-
-
-class DataLoader:
-    """Implements a dataset wrapper for fast data loading
-
-    >>> from doctr.datasets import CORD, DataLoader
-    >>> train_set = CORD(train=True, download=True)
-    >>> train_loader = DataLoader(train_set, batch_size=32)
-    >>> train_iter = iter(train_loader)
-    >>> images, targets = next(train_iter)
-
-    Args:
-        dataset: the dataset
-        shuffle: whether the samples should be shuffled before passing it to the iterator
-        batch_size: number of elements in each batch
-        drop_last: if `True`, drops the last batch if it isn't full
-        collate_fn: function to merge samples into a batch
-    """
-
-    def __init__(
-        self,
-        dataset,
-        shuffle: bool = True,
-        batch_size: int = 1,
-        drop_last: bool = False,
-        collate_fn: Callable | None = None,
-    ) -> None:
-        self.dataset = dataset
-        self.shuffle = shuffle
-        self.batch_size = batch_size
-        nb = len(self.dataset) / batch_size
-        self.num_batches = math.floor(nb) if drop_last else math.ceil(nb)
-        if collate_fn is None:
-            self.collate_fn = self.dataset.collate_fn if hasattr(self.dataset, "collate_fn") else default_collate
-        else:
-            self.collate_fn = collate_fn
-        self.reset()
-
-    def __len__(self) -> int:
-        return self.num_batches
-
-    def reset(self) -> None:
-        # Updates indices after each epoch
-        self._num_yielded = 0
-        self.indices = np.arange(len(self.dataset))
-        if self.shuffle is True:
-            np.random.shuffle(self.indices)
-
-    def __iter__(self):
-        self.reset()
-        return self
-
-    def __next__(self):
-        if self._num_yielded < self.num_batches:
-            # Get next indices
-            idx = self._num_yielded * self.batch_size
-            indices = self.indices[idx : min(len(self.dataset), idx + self.batch_size)]
-
-            samples = list(map(self.dataset.__getitem__, indices))
-
-            batch_data = self.collate_fn(samples)
-
-            self._num_yielded += 1
-            return batch_data
-        else:
-            raise StopIteration

doctr/io/image/tensorflow.py

@@ -1,101 +0,0 @@
-# Copyright (C) 2021-2025, Mindee.
-
-# This program is licensed under the Apache License 2.0.
-# See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
-
-
-import numpy as np
-import tensorflow as tf
-from PIL import Image
-from tensorflow.keras.utils import img_to_array
-
-from doctr.utils.common_types import AbstractPath
-
-__all__ = ["tensor_from_pil", "read_img_as_tensor", "decode_img_as_tensor", "tensor_from_numpy", "get_img_shape"]
-
-
-def tensor_from_pil(pil_img: Image.Image, dtype: tf.dtypes.DType = tf.float32) -> tf.Tensor:
-    """Convert a PIL Image to a TensorFlow tensor
-
-    Args:
-        pil_img: a PIL image
-        dtype: the output tensor data type
-
-    Returns:
-        decoded image as tensor
-    """
-    npy_img = img_to_array(pil_img)
-
-    return tensor_from_numpy(npy_img, dtype)
-
-
-def read_img_as_tensor(img_path: AbstractPath, dtype: tf.dtypes.DType = tf.float32) -> tf.Tensor:
-    """Read an image file as a TensorFlow tensor
-
-    Args:
-        img_path: location of the image file
-        dtype: the desired data type of the output tensor. If it is float-related, values will be divided by 255.
-
-    Returns:
-        decoded image as a tensor
-    """
-    if dtype not in (tf.uint8, tf.float16, tf.float32):
-        raise ValueError("insupported value for dtype")
-
-    img = tf.io.read_file(img_path)
-    img = tf.image.decode_jpeg(img, channels=3)
-
-    if dtype != tf.uint8:
-        img = tf.image.convert_image_dtype(img, dtype=dtype)
-        img = tf.clip_by_value(img, 0, 1)
-
-    return img
-
-
-def decode_img_as_tensor(img_content: bytes, dtype: tf.dtypes.DType = tf.float32) -> tf.Tensor:
-    """Read a byte stream as a TensorFlow tensor
-
-    Args:
-        img_content: bytes of a decoded image
-        dtype: the desired data type of the output tensor. If it is float-related, values will be divided by 255.
-
-    Returns:
-        decoded image as a tensor
-    """
-    if dtype not in (tf.uint8, tf.float16, tf.float32):
-        raise ValueError("insupported value for dtype")
-
-    img = tf.io.decode_image(img_content, channels=3)
-
-    if dtype != tf.uint8:
-        img = tf.image.convert_image_dtype(img, dtype=dtype)
-        img = tf.clip_by_value(img, 0, 1)
-
-    return img
-
-
-def tensor_from_numpy(npy_img: np.ndarray, dtype: tf.dtypes.DType = tf.float32) -> tf.Tensor:
-    """Read an image file as a TensorFlow tensor
-
-    Args:
-        npy_img: image encoded as a numpy array of shape (H, W, C) in np.uint8
-        dtype: the desired data type of the output tensor. If it is float-related, values will be divided by 255.
-
-    Returns:
-        same image as a tensor of shape (H, W, C)
-    """
-    if dtype not in (tf.uint8, tf.float16, tf.float32):
-        raise ValueError("insupported value for dtype")
-
-    if dtype == tf.uint8:
-        img = tf.convert_to_tensor(npy_img, dtype=dtype)
-    else:
-        img = tf.image.convert_image_dtype(npy_img, dtype=dtype)
-        img = tf.clip_by_value(img, 0, 1)
-
-    return img
-
-
-def get_img_shape(img: tf.Tensor) -> tuple[int, int]:
-    """Get the shape of an image"""
-    return img.shape[:2]

doctr/models/classification/magc_resnet/tensorflow.py

@@ -1,196 +0,0 @@
-# Copyright (C) 2021-2025, Mindee.
-
-# This program is licensed under the Apache License 2.0.
-# See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
-
-import math
-from copy import deepcopy
-from functools import partial
-from typing import Any
-
-import tensorflow as tf
-from tensorflow.keras import activations, layers
-from tensorflow.keras.models import Sequential
-
-from doctr.datasets import VOCABS
-
-from ...utils import _build_model
-from ..resnet.tensorflow import ResNet
-
-__all__ = ["magc_resnet31"]
-
-
-default_cfgs: dict[str, dict[str, Any]] = {
-    "magc_resnet31": {
-        "mean": (0.694, 0.695, 0.693),
-        "std": (0.299, 0.296, 0.301),
-        "input_shape": (32, 32, 3),
-        "classes": list(VOCABS["french"]),
-        "url": "https://doctr-static.mindee.com/models?id=v0.9.0/magc_resnet31-16aa7d71.weights.h5&src=0",
-    },
-}
-
-
-class MAGC(layers.Layer):
-    """Implements the Multi-Aspect Global Context Attention, as described in
-    <https://arxiv.org/pdf/1910.02562.pdf>`_.
-
-    Args:
-        inplanes: input channels
-        headers: number of headers to split channels
-        attn_scale: if True, re-scale attention to counteract the variance distibutions
-        ratio: bottleneck ratio
-        **kwargs
-    """
-
-    def __init__(
-        self,
-        inplanes: int,
-        headers: int = 8,
-        attn_scale: bool = False,
-        ratio: float = 0.0625,  # bottleneck ratio of 1/16 as described in paper
-        **kwargs,
-    ) -> None:
-        super().__init__(**kwargs)
-
-        self.headers = headers  # h
-        self.inplanes = inplanes  # C
-        self.attn_scale = attn_scale
-        self.ratio = ratio
-        self.planes = int(inplanes * ratio)
-
-        self.single_header_inplanes = int(inplanes / headers)  # C / h
-
-        self.conv_mask = layers.Conv2D(filters=1, kernel_size=1, kernel_initializer=tf.initializers.he_normal())
-
-        self.transform = Sequential(
-            [
-                layers.Conv2D(filters=self.planes, kernel_size=1, kernel_initializer=tf.initializers.he_normal()),
-                layers.LayerNormalization([1, 2, 3]),
-                layers.ReLU(),
-                layers.Conv2D(filters=self.inplanes, kernel_size=1, kernel_initializer=tf.initializers.he_normal()),
-            ],
-            name="transform",
-        )
-
-    def context_modeling(self, inputs: tf.Tensor) -> tf.Tensor:
-        b, h, w, c = (tf.shape(inputs)[i] for i in range(4))
-
-        # B, H, W, C -->> B*h, H, W, C/h
-        x = tf.reshape(inputs, shape=(b, h, w, self.headers, self.single_header_inplanes))
-        x = tf.transpose(x, perm=(0, 3, 1, 2, 4))
-        x = tf.reshape(x, shape=(b * self.headers, h, w, self.single_header_inplanes))
-
-        # Compute shorcut
-        shortcut = x
-        # B*h, 1, H*W, C/h
-        shortcut = tf.reshape(shortcut, shape=(b * self.headers, 1, h * w, self.single_header_inplanes))
-        # B*h, 1, C/h, H*W
-        shortcut = tf.transpose(shortcut, perm=[0, 1, 3, 2])
-
-        # Compute context mask
-        # B*h, H, W, 1
-        context_mask = self.conv_mask(x)
-        # B*h, 1, H*W, 1
-        context_mask = tf.reshape(context_mask, shape=(b * self.headers, 1, h * w, 1))
-        # scale variance
-        if self.attn_scale and self.headers > 1:
-            context_mask = context_mask / math.sqrt(self.single_header_inplanes)
-        # B*h, 1, H*W, 1
-        context_mask = activations.softmax(context_mask, axis=2)
-
-        # Compute context
-        # B*h, 1, C/h, 1
-        context = tf.matmul(shortcut, context_mask)
-        context = tf.reshape(context, shape=(b, 1, c, 1))
-        # B, 1, 1, C
-        context = tf.transpose(context, perm=(0, 1, 3, 2))
-        # Set shape to resolve shape when calling this module in the Sequential MAGCResnet
-        batch, chan = inputs.get_shape().as_list()[0], inputs.get_shape().as_list()[-1]
-        context.set_shape([batch, 1, 1, chan])
-        return context
-
-    def call(self, inputs: tf.Tensor, **kwargs) -> tf.Tensor:
-        # Context modeling: B, H, W, C  ->  B, 1, 1, C
-        context = self.context_modeling(inputs)
-        # Transform: B, 1, 1, C  ->  B, 1, 1, C
-        transformed = self.transform(context, **kwargs)
-        return inputs + transformed
-
-
-def _magc_resnet(
-    arch: str,
-    pretrained: bool,
-    num_blocks: list[int],
-    output_channels: list[int],
-    stage_downsample: list[bool],
-    stage_conv: list[bool],
-    stage_pooling: list[tuple[int, int] | None],
-    origin_stem: bool = True,
-    **kwargs: Any,
-) -> ResNet:
-    kwargs["num_classes"] = kwargs.get("num_classes", len(default_cfgs[arch]["classes"]))
-    kwargs["input_shape"] = kwargs.get("input_shape", default_cfgs[arch]["input_shape"])
-    kwargs["classes"] = kwargs.get("classes", default_cfgs[arch]["classes"])
-
-    _cfg = deepcopy(default_cfgs[arch])
-    _cfg["num_classes"] = kwargs["num_classes"]
-    _cfg["classes"] = kwargs["classes"]
-    _cfg["input_shape"] = kwargs["input_shape"]
-    kwargs.pop("classes")
-
-    # Build the model
-    model = ResNet(
-        num_blocks,
-        output_channels,
-        stage_downsample,
-        stage_conv,
-        stage_pooling,
-        origin_stem,
-        attn_module=partial(MAGC, headers=8, attn_scale=True),
-        cfg=_cfg,
-        **kwargs,
-    )
-    _build_model(model)
-
-    # Load pretrained parameters
-    if pretrained:
-        # The number of classes is not the same as the number of classes in the pretrained model =>
-        # skip the mismatching layers for fine tuning
-        model.from_pretrained(
-            default_cfgs[arch]["url"], skip_mismatch=kwargs["num_classes"] != len(default_cfgs[arch]["classes"])
-        )
-
-    return model
-
-
-def magc_resnet31(pretrained: bool = False, **kwargs: Any) -> ResNet:
-    """Resnet31 architecture with Multi-Aspect Global Context Attention as described in
-    `"MASTER: Multi-Aspect Non-local Network for Scene Text Recognition",
-    <https://arxiv.org/pdf/1910.02562.pdf>`_.
-
-    >>> import tensorflow as tf
-    >>> from doctr.models import magc_resnet31
-    >>> model = magc_resnet31(pretrained=False)
-    >>> input_tensor = tf.random.uniform(shape=[1, 224, 224, 3], maxval=1, dtype=tf.float32)
-    >>> out = model(input_tensor)
-
-    Args:
-        pretrained: boolean, True if model is pretrained
-        **kwargs: keyword arguments of the ResNet architecture
-
-    Returns:
-        A feature extractor model
-    """
-    return _magc_resnet(
-        "magc_resnet31",
-        pretrained,
-        [1, 2, 5, 3],
-        [256, 256, 512, 512],
-        [False] * 4,
-        [True] * 4,
-        [(2, 2), (2, 1), None, None],
-        False,
-        stem_channels=128,
-        **kwargs,
-    )