TF 2.x Linux GPU (Compute 5.2-9.0), Python 3.x, CUDA 12.6, cuDNN 9.4, AVX2, MKL, TensorRT 10.4

These wheels were built with the manylinux_2_18_x86_64 image.

Google Drive folder
Files and SHA256 checksums:

96fb7e870548202dfb1ae7ae39013f21e8b548c582a2d73f2d7867167c366dd9  tensorflow_io-0.37.1-cp310-cp310-linux_x86_64.whl
7bbfbe5c28053bcb90faee6bb95f7ef0e5fe9351675c2b389e4013060e569c40  tensorflow_io-0.37.1-cp311-cp311-linux_x86_64.whl
4fd848f9cca34599db78a19ac0a1be8573a08f1b8b1240d2484e115972c39bba  tensorflow_io-0.37.1-cp312-cp312-linux_x86_64.whl
08996cab97e82a982ee0f0f25a44cfe9494d56d584c3c9317936009f742fef3b  tensorflow_io-0.37.1-cp39-cp39-linux_x86_64.whl
cd8ef7e196d929aa336a38f51706a45580a5e622a270542d4c350977efc81bde  tensorflow_io_gcs_filesystem-0.37.1-cp310-cp310-linux_x86_64.whl
044b5276b687fc1973f66b25176afe2709ff32894a13e4428589ef9046494237  tensorflow_io_gcs_filesystem-0.37.1-cp311-cp311-linux_x86_64.whl
e26d8c14bb91ea326fdb5d062cbb48c68855969cd4c1a9ad9944651a852d5826  tensorflow_io_gcs_filesystem-0.37.1-cp312-cp312-linux_x86_64.whl
db3c12a9db2d9c239af9fd838c670fd268de503d5a16ff731416e28ea7165ee8  tensorflow_io_gcs_filesystem-0.37.1-cp39-cp39-linux_x86_64.whl
d5d1fdd1171ea083067e488c9ef278dd93c833541dee9ccdeb09bbd331e58b5d  tensorflow-2.16.2-cp310-cp310-linux_x86_64.whl
8b98ee1ab2200d6cb1d6f02a814ec27eb1fadc632d306902f566963c6416ea7a  tensorflow-2.16.2-cp311-cp311-linux_x86_64.whl
b04d09dfd740387c2c463501c27e8e2eb0c36af1561837a456dafc95e3097947  tensorflow-2.16.2-cp312-cp312-linux_x86_64.whl
d1de2ca98199f5797e725674e00e1913df5501baaebe5ca9d7477dc2d77b3356  tensorflow-2.16.2-cp39-cp39-linux_x86_64.whl
cedf9676fbf52adf782b92dd827ad05c3a9701884364a83f0ce4113cd24d3d4d  tensorflow_text-2.16.1-cp310-cp310-linux_x86_64.whl
d15f91b299c63ded66afae3f9f2017a9b5a1490d7b6913db4351c42f691420a8  tensorflow_text-2.16.1-cp311-cp311-linux_x86_64.whl
66430c68dcf6eedd54e640611a73f7ba654173ab0a511e741c9b3a72270cb908  tensorflow_text-2.16.1-cp312-cp312-linux_x86_64.whl
0324b12ef68932384c3062a22126d020f69c9e4c856a1665ca863d6e34e5b233  tensorflow_text-2.16.1-cp39-cp39-linux_x86_64.whl

TF 2.x Linux GPU (Compute 5.2-9.0), Python 3.x, CUDA 12.3, cuDNN 8.9, AVX2, MKL, TensorRT 8.6

These wheels were built with the manylinux2014_x86_64 image.

Google Drive folder
Files and SHA256 checksums:

bdf46e003d126fb15d4f7a9aec09c7ef993cca4178514d0482542acf2c510064  keras-2.14.0-py2.py3-none-any.whl
8a016b58272ab1b22518c5f5f6c32b68a1c9878a61f84aba8d9fae32259e71e9  keras-2.15.0-py2.py3-none-any.whl
e3840cbafa4cb41c7e4a3012f16df0dabc3f30c0d41465b82d5cf6957742d1a4  tensorflow-2.14.1-cp310-cp310-linux_x86_64.whl
cf22aa8b678030e39a7b7e1992412dc3b928963e3e7396408477a1378044d293  tensorflow-2.14.1-cp311-cp311-linux_x86_64.whl
81666e99bac0b69b788db6c3373224527293c4f726340207cdf5d96ffa488042  tensorflow-2.14.1-cp312-cp312-linux_x86_64.whl
499fdea88ca8799dc3673c7415da14dbb326646e408b87261f4b1327e3f03057  tensorflow-2.14.1-cp38-cp38-linux_x86_64.whl
035f79c7d05956ce260169fe1cde7a01528a4aa4a56646a4ae713b0b7ea32d4f  tensorflow-2.14.1-cp39-cp39-linux_x86_64.whl
6af908a1e6c5f02f59a0bdf6d3a582c32dac30bcfe34e5bd6f579e8333fd86ba  tensorflow-2.15.0-cp310-cp310-linux_x86_64.whl
8328d9ca7de6d7fc6eed15f783f27f0dffd369c9f31ee417f20ef93eef9be95f  tensorflow-2.15.0-cp311-cp311-linux_x86_64.whl
810deaa49b1abf1723ea8ced40eb1ca1e5b4d9f59359237ed496d0ae0ce62bdb  tensorflow-2.15.0-cp312-cp312-linux_x86_64.whl
58c59eaeeb40a04ba9fd00963368f9d649945363e05ca7a252c0d2eae11fbb68  tensorflow-2.15.0-cp39-cp39-linux_x86_64.whl
e30d02d21a91b4d1977dbe52c13247142ba8fe742b5a4adad1b69cdf898285df  tensorflow_estimator-2.14.0-py2.py3-none-any.whl
9ea040e5a96e269dcf9174fee59dc3ef00a2c48792bd83b2592ae21130a04aa3  tensorflow_estimator-2.15.0-py2.py3-none-any.whl
91a11857ce509e7cdbf05e6a6a295ef77e784de24268d1b12065babae8816a71  tensorflow_text-2.14.0-cp310-cp310-linux_x86_64.whl
d75e455cd354574cefafb634bbe559e9b15c3a8f9a6d954ac22d51c3ab4fab58  tensorflow_text-2.14.0-cp311-cp311-linux_x86_64.whl
1f1993bb578b36df5e7e5646d84ff75f5858ef5a78a8b2d408c6b6998d0274fa  tensorflow_text-2.14.0-cp312-cp312-linux_x86_64.whl
c218b3b7b25901111006cc9585e8f25c92716ebdd3e0373aa16ce53a0dfcb835  tensorflow_text-2.14.0-cp38-cp38-linux_x86_64.whl
437a735332d438980b240cd44d24234f179448230e2ba7d13b1cddca39624be2  tensorflow_text-2.14.0-cp39-cp39-linux_x86_64.whl
bc885aac06d9e18994061bfa88c0784bc1c9586ce326ec00d47c87fd209d1c32  tensorflow_text-2.15.0-cp310-cp310-linux_x86_64.whl
98b2d6ad8a7a6b87e8bd2d67816dd38d47a89eb0cdd4f05e37991aff1a6e34b1  tensorflow_text-2.15.0-cp311-cp311-linux_x86_64.whl
e21463b2c3ebc5ef955c9a119c6c3b97e72a4336792dd7575daf5f56911b1e67  tensorflow_text-2.15.0-cp312-cp312-linux_x86_64.whl
85ce4b946a483e939773c2f00ce0b88c5893b9e13f8d3c10e5683ddaaddcd049  tensorflow_text-2.15.0-cp39-cp39-linux_x86_64.whl

TF 2.13.0 Linux GPU (Compute 5.2-9.0), Python 3.x, CUDA 12.2, cuDNN 8.9, AVX2, MKL, TensorRT 8.6

These wheels were built with the manylinux2014_x86_64 image.

Google Drive folder
Files and SHA256 checksums:

82aafd257d2046ffd6a2bfc8c6cc14f8692ea7b78fe6407d1c40255d4b8979b3  keras-2.13.1-py2.py3-none-any.whl
031d068aad9fc29924e51308b5272dca19bb3b08f173d754b4190f9cbe9922de  tensorflow-2.13.0-cp310-cp310-linux_x86_64.whl
9cbf2609af55fca4a484bab0c938324918059395014b72ca8e944702e8463a8d  tensorflow-2.13.0-cp311-cp311-linux_x86_64.whl
c15d4541a9dd1b7b4b7470385c697ab6f9efb2859810d2b5be7f5f4c572ff23e  tensorflow-2.13.0-cp38-cp38-linux_x86_64.whl
a7e741b9c2afc6bd5d12dad3f2c59543de03dba4f1e9ae5513bb9adbcee99b80  tensorflow-2.13.0-cp39-cp39-linux_x86_64.whl
bbd860ed7e4b0612eb0a8ef3d80ad700d778bb72e13bb84f0ca3f982eb00cde3  tensorflow_estimator-2.13.0-py2.py3-none-any.whl
4e1490eb70d574a1f4a98bca0fb3f90b3f20bd440f0db38973154aefe9325ab3  tensorflow_io-0.33.0-cp310-cp310-linux_x86_64.whl
de61e1b5eb707a32ce2b20be560cd3a9e867a4ce0dade56fd62f91300bccd759  tensorflow_io-0.33.0-cp311-cp311-linux_x86_64.whl
40e36752fd6f0b0f9690f86968d9bd3eab5d332070de8bcb91a8638c5081c784  tensorflow_io-0.33.0-cp38-cp38-linux_x86_64.whl
99204ace7e68e700a77d0c6c6b0ff6abd14c57594649d3b77829121cfeb86a3c  tensorflow_io-0.33.0-cp39-cp39-linux_x86_64.whl
67921c65c34e7aad3535b73f75f480aab4dd0e19defcfb4540dacfa7b0d4ce50  tensorflow_io_gcs_filesystem-0.33.0-cp310-cp310-linux_x86_64.whl
d02722b9dc92339c807bbcc180097aeeee87820430019a4a686694ab47575b56  tensorflow_io_gcs_filesystem-0.33.0-cp311-cp311-linux_x86_64.whl
abaca5bde0a843b1c5ca826003d78318058622405b6012f1e045311cafb2661b  tensorflow_io_gcs_filesystem-0.33.0-cp38-cp38-linux_x86_64.whl
d09ea080df29575484171cdc43078b453f2161dcdf08dda8b4a2f5af29b615dd  tensorflow_io_gcs_filesystem-0.33.0-cp39-cp39-linux_x86_64.whl
0468e302324f9667795bc2ea85df2f25c054e68716a194a8607fcddcfc1a62b4  tensorflow_text-2.13.0-cp310-cp310-linux_x86_64.whl
e4e346564581a9531cdac2bf6d122bea6965cf861f4194d384f16a4168988360  tensorflow_text-2.13.0-cp311-cp311-linux_x86_64.whl
4ab42924285fbfc497973520fecf0458d3bb3d3eff4320289df6fdfd240ad012  tensorflow_text-2.13.0-cp38-cp38-linux_x86_64.whl
7301224328db169cb8bfa408c9f0674a5f9f32ff3ce2d56874e4b06987c459f6  tensorflow_text-2.13.0-cp39-cp39-linux_x86_64.whl

TF 2.x Linux GPU (Compute 5.2-9.0), Python 3.x, CUDA 12.1, cuDNN 8.9, AVX2, MKL, TensorRT 8.6

These wheels were built with the manylinux2014_x86_64 image.

Google Drive folder
Files and SHA256 checksums:

a630f9af373b5f37a401aa1406c7cfe3a49f2facb3a6b7f77e50dffb96370c00  tensorflow-2.12.0-cp310-cp310-linux_x86_64.whl
a7deede05cdf7328ed05822650b8a084c850e6704b74efa5ca3c246fa37dc98c  tensorflow-2.12.0-cp311-cp311-linux_x86_64.whl
d15186c8eb25fa25061b0b32be4942731ca3baab7254a1817b545fd9057932c3  tensorflow-2.12.0-cp38-cp38-linux_x86_64.whl
715264651837fb07ed94a08ddd823bc24225b7fde71d6397f5c991f97a8498ff  tensorflow-2.12.0-cp39-cp39-linux_x86_64.whl

TF 2.7-2.11 Linux GPU (Compute 5.2-9.0), Python 3.7-3.11, CUDA 11.8, cuDNN 8.8, AVX2, MKL, TensorRT 8.5

These wheels were built with the manylinux2014_x86_64 image. Corresponding versions of Keras were built alongside, in case you have issues with Keras incompatibility.

Google Drive folder
Files and SHA256 checksums:

03e25be1a9a0f203914d17e4caca79215e343d10f839af2e035f23c4be966094  tensorflow-2.10.1-cp310-cp310-linux_x86_64.whl
7163464599e21b7bb0ff2985b2429dab4aabae69b143eea4ee69f2a57fed89cb  tensorflow-2.10.1-cp311-cp311-linux_x86_64.whl
8b25658a1f127f2d1907442b737daf0b024ce6dd5e62bc30e5c2d7ac378ecc7c  tensorflow-2.10.1-cp37-cp37m-linux_x86_64.whl
4215de2c5e128a66cf8566246d659cbede1d0606a6ee0cbb229ddede565eab3b  tensorflow-2.10.1-cp38-cp38-linux_x86_64.whl
2c4fe6f436019a89bd9a24356bf8fe223e3b5ed784a385e8bd2c08f4f9301f9a  tensorflow-2.10.1-cp39-cp39-linux_x86_64.whl
836e904801552cec541744e6452f0542336ae603dabaa64c021f771151b0ac45  tensorflow-2.11.0-cp310-cp310-linux_x86_64.whl
ea9ea683adb66f1157b6decb3d72fb88103378a6ea90081d46b24e0abfc6c699  tensorflow-2.11.0-cp311-cp311-linux_x86_64.whl
78955c445c0c1f557f093be6d2ae5cff26911980e61928c8c8df441cb3626c93  tensorflow-2.11.0-cp37-cp37m-linux_x86_64.whl
5db07f9f34141dc16212406498501825d95cc79db4cf16e23eb01737d18789a1  tensorflow-2.11.0-cp38-cp38-linux_x86_64.whl
2efef4659a6e623cbf85a8542b826ef1692f66ce93fbe2eaafc78c6d2764823a  tensorflow-2.11.0-cp39-cp39-linux_x86_64.whl
6460ee254086051d4ce1de52a419e23534009f989b58fcf4b5866e47f25f8e29  tensorflow-2.7.4-cp310-cp310-linux_x86_64.whl
62476202b862ea07a0e6c9d707a4f9b7490b6a3e68bb81d888455e2427cb18cb  tensorflow-2.7.4-cp311-cp311-linux_x86_64.whl
453c8f6a010f91c3ca2cc458593acac6b0f08884249ad7c1a3fd34207f13dfa1  tensorflow-2.7.4-cp36-cp36m-linux_x86_64.whl
d22e193958a283d56e35c16c674d15e1eb403074219ff170c32657effd1de636  tensorflow-2.7.4-cp37-cp37m-linux_x86_64.whl
5a60179de2422978866ad32d52fe31694abe55a07b7432ebc70d4f31c9f8bdc5  tensorflow-2.7.4-cp38-cp38-linux_x86_64.whl
bd5a61cef7bb6222a4205631a84eeb725f283faa67c1e68b1a6f510892c3df90  tensorflow-2.7.4-cp39-cp39-linux_x86_64.whl
51cb208301a860e5ea88f82e54c331472e6983fcdeb92826305194a8d3986646  tensorflow-2.8.4-cp310-cp310-linux_x86_64.whl
ed76643b94403e313d517c8ff56bcd174d717eec48a945042e20236057580313  tensorflow-2.8.4-cp311-cp311-linux_x86_64.whl
8832e94ff4ee7b20fa1d73407cc2f9a31c690434f6e255f5dc268a4e965a8a48  tensorflow-2.8.4-cp37-cp37m-linux_x86_64.whl
ef7cd8d16b6fc248718038ce7fc8854f622b9953fd6129b06e3fc40eb8bcb9d9  tensorflow-2.8.4-cp38-cp38-linux_x86_64.whl
04aa111905be4239056b0252a35a95113fdc82cc2ea3d6627b48bad14cef956b  tensorflow-2.8.4-cp39-cp39-linux_x86_64.whl
4f13bbd3e36f046b5ba27dc36ec199e3772ec0829b0737ae550b7dbe60b44cc8  tensorflow-2.9.3-cp310-cp310-linux_x86_64.whl
fe2c8fac6166b38dadc7e00167e27a39868e6f5c5e37a1687dd90f30433890c6  tensorflow-2.9.3-cp311-cp311-linux_x86_64.whl
85397ff99140d5e3ed4e6b416172e9379ea5b831c746cb7e9927da3c8ad93ea1  tensorflow-2.9.3-cp37-cp37m-linux_x86_64.whl
f65b5b5aedb6be8865465b4493b36f3ac172e2acaba820be3f266130a532363d  tensorflow-2.9.3-cp38-cp38-linux_x86_64.whl
778363aaeec324c48a9461d3e4a156aa7be393c83fa62505c562966a4dd58ccc  tensorflow-2.9.3-cp39-cp39-linux_x86_64.whl
b3432b0e26683cb6ad052f3612d140322885722589706dcdc3edf092afe5befd  keras-2.10.0-py2.py3-none-any.whl
d6ad46896b0a482ed2154f69cd25334bd7f15454ab15260d19adc2a22a3c5a05  keras-2.11.0-py2.py3-none-any.whl
84305794f6cd70a20cace19a7b8d853978542620c8bead4785bb5d5188b9399b  keras-2.7.0-py2.py3-none-any.whl
0f52eea0255f61a12296589abe63dba7cfbc0f082c783924a48de63274f8e70b  keras-2.8.0-py2.py3-none-any.whl
092a2fd248a2d4c2c24780bfd2cf9ebd80aad4e52aa7df744a1bbe81e45c2d5a  keras-2.9.0-py2.py3-none-any.whl

TF 2.11.0 (nightly) Linux GPU (Compute 5.2-9.0), Python 3.11, CUDA 11.8, cuDNN 8, AVX2, MKL, TensorRT 8

This wheel is built as tf-nightly so that it pulls the latest versions of its dependencies, which are compatible with Python 3.11. However, tensorflow-io-gcs-filesystem is a requirement but has broken wheels which won't install under Python 3.11, so you need to install as pip install --ignore-requires-python tensorflow-io-gcs-filesystem.

Checksums

MD5: 2735a771bebcacb149a7ab2bcbe4488f
SHA1: 8e32dfcb2306f3df643d6bc94e9d65b4d3e2b341
SHA256: 3f2e6a5c8f030d804da8deea9f18e6fe16a5c161be1f0795931f69e2fb66c7da
SHA512: 0060751618f028761f7259a10bf903596f217d91e2dc72fce7953410afef755e24f39da131924d2f5cb7ef19c78d93f701681802de77f1aa2a3e80ac1a9e81cf
B2: d627f70b6120058cc93994c78ee4bf033e0552bc6f04c079182633d937b8ed0656255cc90678ad34f2f83c87938d9d7b315aa9dbe3bdf9cdf77eeba2029d17f7

TF 2.11.0 Linux GPU (Compute 5.2-9.0), Python 3.7-3.10, CUDA 11.8, cuDNN 8, AVX2, MKL, TensorRT 8

These were compiled on a Ubuntu 20.04 Docker image with Glibc 2.31.

TF 2.10.0 Linux GPU (Compute 5.2-8.7), Python 3.9, CUDA 11.7, cuDNN 8, AVX2, MKL, TensorRT 8

This was built with commit tensorflow/tensorflow@cd8cb49 in order to fix the Unable to register cuBLAS factory error tensorflow/tensorflow#57663

Checksums

SHA1: 1e3bea09b4e7c54f6cedbddd8879e9b2264ed6bd
SHA256: 7360f1d24db56d961a69791f696ffa9a7a6afcb2964450d44e087e1007370f5d
SHA512: b5f29afb11eb182014aed39b46dbab200e3efa3e1aaf069d9cc7ea7748b47849b346dc07571ef5e112b16b38eb05911f86419a1b2ba59f7bce87483728f6b48c
B2: 19aca612799a6357a9365ad45776b97b69e0b83b9d23577a0bee5352f5d5e9070d61adff57a4d51c9935bc91cabbafa022fca5cb5d98e0362b6db3aea72f47eb
MD5: c1b3d1d56040002f7d1d6187394d9244

TF 2.10.0 Linux GPU (Compute 5.2-8.7), Python 3.10, CUDA 11.7, cuDNN 8, AVX2, MKL, TensorRT 8

This was built with commit tensorflow/tensorflow@cd8cb49 in order to fix the Unable to register cuBLAS factory error tensorflow/tensorflow#57663

Checksums

SHA1: 8386111c5853296571baf6f57a25181e17b75ea0
SHA256: fa29ce59147a96509283fa92bbab96399482bb74abea5c3812626ba063f2471d
SHA512: 3cc768418348103201887b7a5017daaf3e2b46ed005de0d235cf8c53b992cdd8ff1a3b18afd9eeafca3944e687fe6aefbe55d318487a58e4e0c88a93fbb6f4aa
B2: 0f8054b3e86af991f0733a2261d9b6fb3fb49501cb9297e29798004d91b49c9e4c3ebc82d417a7336a1eafbba46ab0a2d3c05755508b31d406069e5916f6c9c5
MD5: badf87c57783be0da77b11fa3a42c926

TF 2.10.0 Linux GPU (Compute 5.2-8.7), Python 3.8, CUDA 11.7, cuDNN 8, AVX2, MKL, TensorRT 8

This was built with commit tensorflow/tensorflow@cd8cb49 in order to fix the Unable to register cuBLAS factory error tensorflow/tensorflow#57663

Checksums

SHA1: 42eb7853f6d0cc64c332e8fb693e960fdc7b9596
SHA256: f353aeddd4a627f1214b1e92f1c0535b1aa0a3089d55c6193f7eeb181d827c8c
SHA512: 5b7f02cd806c9e8eabeebccb4d6aca85b413e8c676a9193edc348cf2a2d28a3d8ac8974f77cf335c9da6f1840b8422565d0dd8ae3e6a9ee1888a73f092179dcb
B2: 71c092474ef7dbac68c5e57c7f6e6598b329910f9db2febcc2de40dba80e9408e2cd0caa197dd488034c54258c3b6053767fdfb0ecb14b67ba7c8209921fec27
MD5: 02cfc5098eba73e44843620583b06a15