summaryrefslogtreecommitdiff
path: root/aten/src
AgeCommit message (Expand)AuthorFilesLines
2018-04-18Add mutex to THC random number generator (#6527)Will Feng21-88/+148
2018-04-18Implement torch.einsum (fixes #1889) (#6307)Thomas Viehmann2-10/+167
2018-04-18Better dispatch (#6687)Zeming Lin1-2/+2
2018-04-17__STDC_FORMAT_MACROS was conflicting with some thirdparty include from google...Dmytro Dzhulgakov1-0/+2
2018-04-17Adding dispatch to Tensors (#6664)Zeming Lin2-0/+11
2018-04-17randperm supports n=0 (#6656)Francisco Massa1-2/+2
2018-04-17Support gpu triangle solve (#6648)Du Phan3-0/+40
2018-04-16Add dtypes (with reasonable defaults) to sum, prod, cumsum, cumprod. (#6573)gchanan5-13/+237
2018-04-16Create safe and unsafe versions of sparse_coo_tensor (#6058)li-roy7-145/+185
2018-04-16Fix bilinear performance regression (#6110)Thomas Viehmann3-6/+194
2018-04-13 More precise digamma (#6517)Richard Zou4-34/+179
2018-04-12Use THC allocation for CUFFT workspace (#6568)Tongzhou Wang1-7/+15
2018-04-12Support arbitrary number of batch dimensions in *FFT (#6528)Tongzhou Wang1-16/+31
2018-04-12Separate cuda-ness from dtype. (#6470)gchanan1-0/+17
2018-04-11Fix #6398, Add MKL threading support for Windows (#6416)peterjc1232-1/+9
2018-04-11Fix THTensor_(take) negative index check (#6482)Tongzhou Wang1-7/+7
2018-04-11Add SmallVector from llvm (#6485)Sam Gross4-0/+1172
2018-04-10[fft] [3 of 3] Implements backward of fft ifft rfft irfft (#5537)Tongzhou Wang5-18/+18
2018-04-10Fix fft when any of the input dimensions is not aligned (#6118)Tongzhou Wang4-99/+222
2018-04-10[pytorch] Fix signed random_ (#6463)albanD2-7/+11
2018-04-10Add torch.randint and torch.randint_like functions (#6136)Naman Jain2-0/+60
2018-04-10Fix reflection padding boundary checks (#6438)Tongzhou Wang4-12/+12
2018-04-09Another CUDA 8 fix for Windows (#6383)peterjc1232-0/+10
2018-04-08[pytorch] add static linkage support for CuDNN and NCCL (#6410)Soumith Chintala1-15/+4
2018-04-08fix incorrect error message in convolution_expand_param_if_needed (#6409)Tongzhou Wang1-1/+1
2018-04-07Fix typos in docs (#6389)Kento NOZAWA1-1/+1
2018-04-07Several minor fixes for Windows build (#6332)peterjc1231-1/+1
2018-04-05[ready] Implement log2 and log10 in PyTorch (#6272)Vishwak Srinivasan12-0/+119
2018-04-05Use reshape({-1}) (#6281)Peter Goldsborough1-2/+2
2018-04-05Move instruction set specific code to anonymous namespace (#6314)Sam Gross7-12/+17
2018-04-05default build with MKL for desktop (#6266)Sergei Belousov1-2/+1
2018-04-05Add default args to loss functions in native_functions.yaml (#6289)li-roy1-3/+3
2018-04-04[pytorch] minor fixes around binary builds (#6291)Soumith Chintala1-4/+0
2018-04-04Add max_values and argmax convenience functions to ATen (#6201)Peter Goldsborough2-0/+56
2018-04-04Fix potential UB when input is empty (#6242)Sam Gross2-4/+8
2018-04-04Do not use cpuinfo on PowerPC (#6255)Simeon Monov1-0/+3
2018-04-03fix sparse embedding backward when input contains only padding_idx (#6211)Tongzhou Wang1-3/+9
2018-04-03Fix memory leak in maxpool3d backwards (#6230)Richard Zou1-1/+0
2018-04-03Reduce flakiness of math tests in test_torch.py (#6200)Sam Gross2-6/+21
2018-04-03avx_mathfun.h is imprecise (#6192)cpuhrsch2-20/+1
2018-04-02Add a missed parenthesis to the LogSigmoid documentation (#6209)Joo-Kyung Kim1-1/+1
2018-03-31Delete NNPACK (#6151)Edward Z. Yang5-630/+5
2018-03-31Fix SparseMM compiler warning (#6156)Richard Zou1-1/+0
2018-03-31Give ATen errors backtraces (#6112)Peter Goldsborough1-9/+197
2018-03-30Use THC cached CUDA device property when get_device_name and get_device_capab...Tongzhou Wang1-0/+10
2018-03-30Move LayerNorm to ATen; remove tracking_running_stats functionality (#5983)Tongzhou Wang2-9/+81
2018-03-29Enable MKLDNN convolution forward and backward (#6062)Ma Mingfei7-2/+545
2018-03-29Speed up sum over a dimension (#6026)Sam Gross16-868/+956
2018-03-29Add source location information to error messages (#6059)Peter Goldsborough30-181/+264
2018-03-29Update FFT comments from #5856 (#6089)Tongzhou Wang2-3/+3