microsoft
/
caffe
зеркало из https://github.com/microsoft/caffe.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
1 981 коммит 3 Ветки 11 Теги 51 MiB
Граф коммитов

1981 Коммитов

Автор SHA1 Сообщение Дата
Jeff Donahue ac9275bd6f .Doxyfile: modify to generate C++ docs, excluding tests 2014-09-03 10:59:23 -07:00
Jeff Donahue 3223ca4c98 .gitignore doxygen-generated documentation 2014-09-03 10:59:23 -07:00
Jeff Donahue c34ed497ff add "make {docs,doxygen}" targets to build doxygen-generated docs 2014-09-03 10:59:23 -07:00
Jeff Donahue eebf2e29ae add .Doxyfile: the default Doxygen config file from `doxygen -g` 2014-09-03 10:59:23 -07:00
Jeff Donahue 59eaba19ac [wip] vision layers, start convolution 2014-09-03 10:59:23 -07:00
Jeff Donahue d15405a4b0 use kramdown for markdown syntax; add mathjax 2014-09-03 10:59:23 -07:00
Evan Shelhamer 17252db7a0 [docs] add note on Caffe convolution 2014-09-03 10:59:04 -07:00
Evan Shelhamer b256a7601a [docs] draft tutorial subjects 2014-09-03 10:59:04 -07:00
Evan Shelhamer 5d8c93c201 [docs] skeleton documentation subjects 2014-09-03 10:30:33 -07:00
Evan Shelhamer cdcf888a45 Merge pull request #1025 from mohomran/minor_fix_to_mnist_solver_prototxt 
Snapshot naming issue in mnist_autoencoder_solver_adagrad.prototxt
 2014-09-02 08:38:07 -07:00
Mohamed Omran d3c6eb6207 fixed relative path and prefix for adagrad-optimised autoencoder snapshots 2014-09-02 13:01:13 +02:00
Evan Shelhamer bf19aaf731 Merge pull request #1023 from longjon/unbreak-pycaffe 
Fix pycaffe break due to dropping {blob,layer}.name
 2014-09-01 20:05:57 -07:00
Jonathan L Long 3e12d49324 [pycaffe] use _blob_names, _layer_names instead of removed .name 2014-09-01 19:52:57 -07:00
Jonathan L Long afd8f370b4 [pycaffe] expose Net.blob_names and Net.layer_names 2014-09-01 19:52:38 -07:00
Jonathan L Long ed13a61990 [pycaffe] add converter for vector<string> used by _*_names 2014-09-01 19:52:17 -07:00
Jeff Donahue 396e4afd63 add CUDA 6.5 error CUBLAS_STATUS_LICENSE_ERROR to cublasGetErrorString 
enum
 2014-09-01 16:46:21 -07:00
Jeff Donahue 77d66618af revert tools/train_net.cpp to previous, depecated version 2014-09-01 14:51:59 -07:00
Jeff Donahue 88e17976f6 Merge branch 'qipeng-solvers' into dev 2014-09-01 14:13:56 -07:00
Jeff Donahue b0f97fda76 make MNIST autoencoder solvers start from base_lr 0.01 and step (much 
better performance) and terminate at iter 65K
 2014-09-01 12:35:03 -07:00
Jeff Donahue eaf28fe575 make adagrad/nesterov train scripts follow new "run-from-root" 
convention
 2014-09-01 12:32:54 -07:00
Jeff Donahue b49b2d3e8e Add "test-on-train" stage to test accuracy on the training data; correct 
test_iter (should be 100 instead of 50)
 2014-09-01 12:04:31 -07:00
Jeff Donahue 5894f039a4 mnist_autoencoder: always compute both cross-entropy loss and L2 
(euclidean) error
 2014-09-01 11:40:14 -07:00
qipeng 972264946c hot fix for warning 2014-09-01 11:33:41 -07:00
qipeng 36f9de4303 lint 2014-09-01 11:33:41 -07:00
qipeng a464df45a7 Re-added solver switch into the new caffe main excutable; fixed AdaGrad MNIST example 2014-09-01 11:33:41 -07:00
qipeng 06f335f04b lint 2014-09-01 11:33:41 -07:00
qipeng c1ff97c947 Added sanity check for AdaGradSolver; added MNIST examples for solvers 2014-09-01 11:33:41 -07:00
Jeff Donahue f206c6479f Merge Test{SGD,AdaGrad,Nesterov}Solver; they become subclasses of 
TestGradientBasedSolver
 2014-09-01 11:33:41 -07:00
Jeff Donahue dbb9296050 cleanup caffe.proto 2014-09-01 11:33:41 -07:00
qipeng 7f2e66e6cc added unit test for solvers and fixed solver bugs 2014-09-01 11:33:41 -07:00
qipeng 29b3b2442d proto conflit, lint, and math_functions (compiler complaint) 2014-09-01 11:33:41 -07:00
qipeng 23d4430869 fixes after rebase 2014-09-01 11:33:41 -07:00
qipeng 3f7a910e05 Addressed Yangqing's comments 2014-09-01 11:33:41 -07:00
qipeng b0ec5314f4 fixed caffe.proto after a mistaken rebase 2014-09-01 11:33:41 -07:00
qipeng a683c40d63 Added L1 regularization support for the weights 2014-09-01 11:33:41 -07:00
qipeng 76ef2ca1f6 bugfixes for AdaGrad 2014-09-01 11:33:41 -07:00
qipeng 0144de68e6 improved numerical stability for AdaGrad 2014-09-01 11:33:41 -07:00
qipeng 8b3dde08fc fixed solver constructor in train_net.cpp 2014-09-01 11:33:41 -07:00
qipeng ed8b1da57f converted pointers to shared_ptr 2014-09-01 11:33:41 -07:00
qipeng 8a9c268bd5 restored vituals in solver.hpp 2014-09-01 11:33:41 -07:00
qipeng 9d10569a79 Solver switching support & implementation of Nesterov's accelerated gradient and AdaGrad 2014-09-01 11:33:41 -07:00
Jeff Donahue 5f350aa246 use LMDB in mnist autoencoder examples 2014-09-01 11:33:19 -07:00
Evan Shelhamer f7458b5cd2 make no GPU error in CPU-only mode a little clearer 2014-08-31 23:28:28 -07:00
Evan Shelhamer fe95024678 include comment on CPU mode fine-tuning for Flickr example 2014-08-31 23:25:40 -07:00
Evan Shelhamer c80c6f355d Merge pull request #1014 from longjon/cleaner-pycaffe 
Clean up pycaffe core
 2014-08-31 23:22:21 -07:00
Evan Shelhamer 8bf1e60cdc Merge pull request #1008 from mohomran/mnist_with_lmdb 
convert MNIST demo to lmdb, fixes
 2014-08-31 11:02:45 -07:00
Evan Shelhamer 92f6c26829 Merge pull request #1015 from mohomran/fixing_parse_log_script 
fix for tools/extra/parse_log.sh
 2014-08-31 11:00:48 -07:00
Mohamed Omran aae8545427 parse_log.sh adapted to new training log format + fixed typos and updated description 2014-08-31 16:15:20 +02:00
Mohamed Omran 5615bf6c17 updated lenet_train_test.prototxt + minor correction to create_mnist.sh 2014-08-31 14:27:33 +02:00
Mohamed Omran c98ed3b8b2 minor changes to variable names and error messages + set default backed in convert_mnist_data.cpp to lmdb 2014-08-31 14:05:51 +02:00