Commit Graph

173 Commits (ec9bc1bd9f2381e95a7933e2157974c2a38d89f7)

Author SHA1 Message Date
guomingz 6d1d7c8acf Fix the KL algorithm bug when calculated the size of tensor. (#17198)
7 years ago
Zhen Wang 65541d83b0
add scale pass for calculating the output scales.test=develop (#17259)
7 years ago
Zhen Wang a40121e4c8
fix the initialization process error. test=develop (#17213)
7 years ago
wopeizl 06325c4de1
disable the test_distillation_strategy temporarily (#17227)
7 years ago
guomingz e4a52e082f resolve #17159 (#17172)
7 years ago
Yibing Liu beda78258f
Init mixed precision training interface (#16856)
7 years ago
Zhen Wang b8c166f632
Merge pull request #17029 from wzzju/add_graph_checkpoint
7 years ago
Zhen Wang bf6b9d6d37 add checkpoint functions for graph. test=develop
7 years ago
Wang, Chuanqi 45ebafc33e Split test_calibration test to two tests (#16977)
7 years ago
whs 486f7d8ed6
Restore quantization and distillation stategy before loading persistables. (#16958)
7 years ago
Tao Luo bc037c13c7 use multi-thread to speedup CI tests
7 years ago
chuanqiw 1818d1d9f0 Reduce test_calibration test time by using large batchsize
7 years ago
AIFollowers fb1ae722d1 Fix model_stat importing in paddle/fluid/contrib
7 years ago
chengduo bb80dae7d0
Add DecoupledWeightDecay (#16427)
7 years ago
whs 73c4f2b7b6
Fix distillation for soft label. (#16538)
7 years ago
Zhen Wang 4c1ec41de3
Merge pull request #16531 from wanghaoshuang/quan_ck
7 years ago
AIFollowers e18ab78f67 add model_stat.py (#16512)
7 years ago
wanghaoshuang d41b623a72 Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into quan_ck
7 years ago
wanghaoshuang 6db7c2a500 Fix checkpoint of quantization.
7 years ago
Zhen Wang f86429dbd9 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_slim_quant_bugs
7 years ago
whs 679a4c28fc
Fix lost of learning rate variable in distillatoin when using lr decay. (#16471)
7 years ago
Zhen Wang 6b854f3e1f fix the save_in_nodes bug.
7 years ago
Zhen Wang 183bacebe3 clean codes and fix some bugs. test=develop
7 years ago
Tao Luo 1b4e4e7ef7
Merge pull request #16453 from chuanqi129/calibration_readme_refine
7 years ago
liuwei1031 8d22bc17a4
Memory optimize (#16410)
7 years ago
chuanqiw c512516ff4 Update INT8 calibration README
7 years ago
Zhen Wang 27d05203e7 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_quan_hang
7 years ago
Zhen Wang 1c11f817e9 Use the resolve hazard method.
7 years ago
Wang, Chuanqi 85e1cc1e02 Update Readme with new accuracy and performance data measured on 6271 (#16437)
7 years ago
whs e9bec9369b
[slim] Add quantization strategy and distillation strategy. (#16408)
7 years ago
Zhen Wang 2ccbfd5e10 Fix some bugs for quantization passes.
7 years ago
liuwei1031 de3b70a101
fix cdn issue, test=develop (#16423)
7 years ago
chuanqiw 431068c9ca Enhance test calibration script on accuracy assert
7 years ago
whs 2e5831f0dc
[slim] Refine framework of slim and add filter pruning strategy (#16226)
7 years ago
Zhen Wang ec11135d54
Merge pull request #16341 from wzzju/add_channel_wise_in_quant_pass
7 years ago
Zhen Wang 8965819fbb rewrite the cuda kernels of channel_wise_quant_op and channe_wise_dequant_op. test=develop
7 years ago
Zhen Wang ec88b6cc5a add channel wise quantization in ir pass.
7 years ago
tangwei12 8ea4218ce1
update load persistables for increment, test=develop (#15576)
7 years ago
achao2013 81b4fad8b9 add moving average absmax op and fix bug (#15155)
7 years ago
Zhen Wang 41b8cf0bae
Merge pull request #16162 from wzzju/fix_nan_static_quant
7 years ago
Zhen Wang b8d1f5038e Add the executor test for the graph clone API. test=develop
7 years ago
Zhen Wang 4ab0607634 fix NaN bugs.
7 years ago
Zhen Wang ac6ef06ffa Add the Clone method in Graph. test=develop
7 years ago
Zhen Wang 1b9c8d5f06 add clone function for IrGraph. test=develop
7 years ago
Zhen Wang 3cf0ee414d update some details. test=develop
7 years ago
Zhen Wang 0cb50bb983 avoid ce fails on windows.
7 years ago
Zhen Wang 7c8f7df2fe add some op_des funs to IrOpNode and add some var_des funs to IrVarNode. test=develop
7 years ago
Zhen Wang 33f99d6197 add IrNode&IrVarNode&IrOpNode. test=develop
7 years ago
Zhang, Guoming e5d3d7c63d resolve #15724
7 years ago
Zhen Wang bc95a4ccfe
Merge branch 'develop' into quantization_inference_passes
7 years ago
sneaxiy 7e399b0628 rename
7 years ago
sneaxiy 42f6d0f899 modify API.spec
7 years ago
sneaxiy f85245b409 test=develop
7 years ago
WangZhen 28dfad5e27 fix some bugs about python3. test=develop
7 years ago
WangZhen 2175292634 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into quantization_inference_passes
7 years ago
WangZhen a7efab7ec1 add comments for public API. test=develop
7 years ago
Haihao Shen b5ebca47a3 Add INT8 calibration README (#15548)
7 years ago
Haihao Shen 1b8047b712 Add INT8 calibration support in Paddle package (#15569)
7 years ago
Haihao Shen 294d594450 Enable performance measurement in INT8 calibration unit test (#15560)
7 years ago
WangZhen 0db41a9c44 add op_role attr when creating op node.
7 years ago
WangZhen c67b29c178 fix some bugs of graph.to_program and get_pass.
7 years ago
Haihao Shen f82515800c Enable INT8 Calibration Unit Test for MobileNet-V1 (#15539)
7 years ago
WangZhen c64f22048a add convert_to_int8 pass and transform_for_mobile pass and their UTs.
7 years ago
WangZhen c8095eeb82 add freeze pass, and UT is passed.
7 years ago
Haihao Shen 2a82c5651e Refine INT8 calibration API; shorten the iteration number to reduce test time; test=develop
7 years ago
乔龙飞 Qiao Longfei c58555067e
Merge pull request #14731 from jacquesqiao/optimize-cpp-reader
7 years ago
WangZhen dde19a0ff8 add quantization freeze pass.
7 years ago
guomingz dbdaf15ca0 [V1.3] Add the calibration tool code for int8 inference and focus test. (#15062)
7 years ago
Qiao Longfei 119a3d4deb update comment test=develop
7 years ago
WangZhen 3b668c1574 Update some comments in the quantization transform pass. test=develop
7 years ago
WangZhen b913463e83 Update according to the reviewers' suggestion. test=develop
7 years ago
WangZhen 3ce6172052 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into graph_quantization
7 years ago
WangZhen 59e5cc51d6 Add quantization transform pass and UT.
7 years ago
Qiao Longfei cf0a057981 add document for ctr reader
7 years ago
Qiao Longfei 45578c1b48 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into optimize-cpp-reader
7 years ago
Dun 9f8f0fc2d3 Memory optimization of depthwise conv op and group norm op (#15313)
7 years ago
WangZhen e2ff300b02 add UT for quantization.
7 years ago
WangZhen 451896fce4 init quantization.
7 years ago
Qiao Longfei edad60e612 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into optimize-cpp-reader
7 years ago
whs 938705745e
Init paddle slim (#14834)
7 years ago
tangwei12 45dd3491db
lookup table utils fix (#14730)
7 years ago
heqiaozhi caa6b59677 add hdfs_utils & helper & node doc
7 years ago
heqiaozhi 7bd16e3afa fix some bug & add log
7 years ago
heqiaozhi 5d3ecbfdf5 fix hdfs bug
7 years ago
heqiaozhi 009c7cf6cc add finialize
7 years ago
heqiaozhi 57ac412b98 download data
7 years ago
Qiao Longfei daba57f752 complete ctr_reader
7 years ago
Qiao Longfei 978fd6800c update ctr_reader.py
7 years ago
Qiao Longfei 668ae9083e Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add-ctr-reader
7 years ago
tangwei12 3639d99f99
Fix save and load lookup table/optimizer vars (#14301)
7 years ago
tangwei12 dd6fd4c747
Utils for download and upload files with HDFS (#14473)
7 years ago
wopeizl d9a1f3e58e Windows/online (#14474)
7 years ago
peizhilin 1f12ba6192 gpu support, fix build issue:
7 years ago
Qiao Longfei aff54ef735 add ctr data
7 years ago
Dang Qingqing f63ab5619b Fix conflict.
7 years ago
Dang Qingqing 8f5d918a19 Disable one test in test_quantize_transpiler.
7 years ago
qingqing01 748be49e77
Fix random fail in Python3 (#13666)
7 years ago
Dang Qingqing 209f799f91 test,
7 years ago
qingqing01 479ad4bb92
Merge branch 'develop' into quantize_transpiler_update
7 years ago
chengduo 4e81e22827
add op frequence (#13328)
7 years ago