Age | Commit message (Collapse) | Author | Files | Lines | |
---|---|---|---|---|---|
2015-08-30 | Merge pull request #2998 from longjon/data-race | Jeff Donahue | 1 | -1/+3 | |
Fix a recently introduced race condition in DataLayer | |||||
2015-08-29 | Merge pull request #2894 from mfigurnov/fix-truncation-warning | Ronghang Hu | 1 | -1/+2 | |
Fix truncation of value warning | |||||
2015-08-29 | fix GPU data race | Jonathan L Long | 1 | -1/+3 | |
Previously, the prefetch GPU -> top GPU and prefetch CPU -> prefetch GPU copies were launched concurrently in separate streams, allowing the next batch to be copied in before the current one is read. This patch explicitly synchronizes the prefetch -> top copy wrt the host, preventing the CPU -> GPU from being launched until its completion. | |||||
2015-08-27 | Merge pull request #2989 from jyegerlehner/embed-layer-compile-warning | Jeff Donahue | 1 | -1/+0 | |
Fix EmbedLayer compiler warning for unused variable. | |||||
2015-08-27 | Fix EmbedLayer compiler warning for unused variable. | J Yegerlehner | 1 | -1/+0 | |
2015-08-27 | Add extra openblas search path | Matt Dawkins | 1 | -0/+2 | |
2015-08-26 | Merge pull request #2944 from philkr/python_layer_param | Jon Long | 2 | -1/+71 | |
Give the python layer parameter/weight blobs. | |||||
2015-08-26 | Merge pull request #2970 from ronghanghu/spp-fix | Ronghang Hu | 2 | -7/+39 | |
Fix SPPLayer top blob num and address `pyramid_height_ == 1` | |||||
2015-08-26 | Fix previous mistake on unimplemented top and address pyramid_height_==1 in ↵ | Ronghang Hu | 2 | -7/+39 | |
SPPLayer also, do nothing in SPPLayer Reshape if already reshaped once and bottom size unchanged | |||||
2015-08-26 | Merge pull request #2964 from jyegerlehner/mvn-layer-fixes | Jeff Donahue | 3 | -8/+27 | |
Fix MVNLayer | |||||
2015-08-26 | Merge pull request #2981 from maaskola/draw-deconvolution | Jeff Donahue | 1 | -3/+3 | |
Draw Deconvolution layers like Convolution layers | |||||
2015-08-26 | Draw Deconvolution layers like Convolution layers | Jonas Maaskola | 1 | -3/+3 | |
2015-08-26 | MVNLayer fixes. | J Yegerlehner | 3 | -8/+27 | |
Fix the MVNLayer tests so they actually test what they claim. MVNLayer fixes: sum_multiplier_ sized correctly; backward gradient calculation. Gradient calculation per analysis of seanbell, found here: https://github.com/BVLC/caffe/issues/1938 Fixes according to review comments. | |||||
2015-08-25 | Merge pull request #2083 from jeffdonahue/tile-layer | Jeff Donahue | 5 | -1/+332 | |
TileLayer | |||||
2015-08-25 | TileLayer: add CUDA kernels | Jeff Donahue | 1 | -14/+39 | |
2015-08-25 | Add TileLayer | Jeff Donahue | 5 | -1/+307 | |
2015-08-25 | Merge pull request #2032 from jeffdonahue/embed-layer | Evan Shelhamer | 7 | -4/+488 | |
Embed layer for lookup table of one hot encodings | |||||
2015-08-25 | Merge pull request #2927 from jeffdonahue/improve-net-init-error-msgs | Evan Shelhamer | 2 | -7/+33 | |
improve net config and shape mismatch error messages | |||||
2015-08-25 | Merge pull request #2972 from jeffdonahue/concat-backward-fix | Evan Shelhamer | 3 | -14/+25 | |
[fix] properly backprop through ConcatLayer with propagate_down set | |||||
2015-08-25 | Python parameter test added | philkr | 2 | -2/+60 | |
2015-08-25 | Allow the python layer have weight/parameter blobs. | philkr | 1 | -2/+14 | |
2015-08-24 | Merge pull request #2963 from longjon/superfluous-toproto | Jeff Donahue | 1 | -6/+0 | |
Remove superfluous code in Net::ToProto | |||||
2015-08-24 | bugfix for ConcatLayer with propagate_down set | Jeff Donahue | 2 | -14/+16 | |
if propagate_down[i] was set, offset_concat_axis was not correctly updated for subsequent bottoms i+1, i+2, ... | |||||
2015-08-24 | TestConcatLayer: add gradient check for bottom[1] only (to verify | Jeff Donahue | 1 | -0/+9 | |
propagate_down[0] == false correctness) | |||||
2015-08-22 | Merge pull request #2935 from rmanor/accuracies | Ronghang Hu | 3 | -1/+134 | |
Output accuracies per class. | |||||
2015-08-22 | Merge pull request #2253 from jyegerlehner/snapshot_on_signal | Ronghang Hu | 5 | -10/+268 | |
Snapshot on signal | |||||
2015-08-23 | Output accuracies per class. | Ran | 3 | -1/+134 | |
Fixed case where number of samples in class can be zero. - Fixed ignore_label case, also added a test. - Two other fixes. Fixed lint errors. Small fix. | |||||
2015-08-22 | Add signal handler and early exit/snapshot to Solver. | J Yegerlehner | 5 | -10/+268 | |
Add signal handler and early exit/snapshot to Solver. Add signal handler and early exit/snapshot to Solver. Also check for exit and snapshot when testing. Skip running test after early exit. Fix more lint. Rebase on master. Finish rebase on master. Fixups per review comments. Redress review comments. Lint. Correct error message wording. | |||||
2015-08-21 | remove superfluous code in Net::ToProto | Jonathan L Long | 1 | -6/+0 | |
2015-08-20 | Merge pull request #2949 from jeffdonahue/deconv-gpu-backward-no-repeat-im2col | Jon Long | 1 | -1/+2 | |
DeconvolutionLayer Backward_gpu fix: don't redo im2col | |||||
2015-08-20 | Merge pull request #2950 from lukeyeager/use-input_shape | Jon Long | 11 | -44/+66 | |
Use input_shape instead of input_dim in examples | |||||
2015-08-20 | Use input_shape instead of input_dim in examples | Luke Yeager | 11 | -44/+66 | |
2015-08-20 | Merge pull request #2947 from BlGene/bdl_fixup | Jeff Donahue | 1 | -2/+1 | |
BasePrefetchingDataLayer fixup. | |||||
2015-08-20 | DeconvolutionLayer Backward_gpu fix: don't redo im2col | Jeff Donahue | 1 | -1/+2 | |
2015-08-20 | Merge pull request #2930 from lukeyeager/pycaffe-layer_type_list | Jeff Donahue | 4 | -11/+34 | |
Expose LayerFactory::LayerTypeList in pycaffe | |||||
2015-08-20 | Expose LayerFactory::LayerTypeList in pycaffe | Luke Yeager | 4 | -11/+34 | |
Useful for validating NetParameters without crashing on SIGABRT | |||||
2015-08-20 | In BasePrefetchingDataLayer::Forward_cpu hanged top[0]->Reshape to | max argus | 1 | -2/+1 | |
top[0]->ReshapeLike, in line with other calls. | |||||
2015-08-18 | Add information about how to get GPU topology from nvidia-smi | mhouston | 1 | -0/+2 | |
2015-08-18 | Add some documentation on Multi-GPU support | mhouston | 1 | -0/+24 | |
2015-08-16 | Merge pull request #2812 from philkr/python_loss_weight | Jon Long | 2 | -0/+14 | |
Expose blob loss weight to python | |||||
2015-08-15 | Malloc at least 1 byte for MultiGPU P2PSync buffers | Ronghang Hu | 1 | -1/+2 | |
2015-08-15 | Merge pull request #2931 from ronghanghu/fix-gpu-test | Ronghang Hu | 1 | -2/+13 | |
Fix MultiGPU solver test crash | |||||
2015-08-14 | Fix MultiGPU solver test with TEST_GPUID != 1 | Ronghang Hu | 1 | -2/+13 | |
This is a patch for multi-gpu testing issue (#2926). The problem fixed in this commit is that when calling make runtest with TEST_GPUID != 0 on a MultiGPU machine, solver tests will crash because gpu ids in multi-gpu tests doesn't match that of single GPU test. | |||||
2015-08-14 | Merge pull request #2928 from cypof/lib_make_target | Jeff Donahue | 1 | -2/+4 | |
New make target to only build the library. | |||||
2015-08-14 | New make target to only build the library. | Cyprien Noel | 1 | -2/+4 | |
2015-08-14 | [net] improve net config and shape mismatch error messages | Jeff Donahue | 2 | -7/+33 | |
2015-08-14 | Exposing blob loss weight to python | philkr | 2 | -0/+14 | |
2015-08-14 | Merge pull request #2925 from flx42/fix_classification_example | Jeff Donahue | 1 | -0/+2 | |
Make classification.bin support models with less than 5 classes | |||||
2015-08-14 | Make classification.bin support models with less than 5 classes | Felix Abecassis | 1 | -0/+2 | |
The example program would crash if the number of classes was less than 5, since it was still attempting to get the top 5 predictions. Close #2585 | |||||
2015-08-14 | Merge pull request #2921 from buaaliyi/multi_gpu | Ronghang Hu | 1 | -2/+7 | |
Destroy CUDA stream when finished |