- 05 7月, 2021 11 次提交
-
-
由 jiangcheng 提交于
* reduce sum op default fp32, add into amp black list * reduce_sum default fp32 can avoid return inf when the sum value large than 65504
-
由 WangXi 提交于
-
由 WangXi 提交于
-
由 pangyoki 提交于
* change Add to AddN in sum npu op * add AddInputNames * change fp16 to fp32 because numpy has accuracy loss in fp16 adding * delete check * fix runner error
-
由 Qi Li 提交于
-
由 WangXi 提交于
-
由 ShenLiang 提交于
-
由 cc 提交于
* refine ptq according comments * reuse the module to calculate kl threshold
-
由 ShenLiang 提交于
* add amp support for pp * add amp untest
-
由 danleifeng 提交于
* fix hdfs and fleet_util for supporting save/load infer;test=develop
-
由 cc 提交于
* Save all scales to target ops * Move quant layers to paddle.nn.quant
-
- 04 7月, 2021 1 次提交
-
-
由 pangyoki 提交于
* delete useless GELU in gelu npu op * add description * fix format * add check_grad in gelu unittest
-
- 02 7月, 2021 3 次提交
- 01 7月, 2021 5 次提交
-
-
由 Yuang Liu 提交于
-
由 sunli 提交于
-
由 JZ-LIANG 提交于
* dygraph sharding * update unitest hybrid_parallel_communicate_group
-
由 taixiurong 提交于
-
由 zhangbo9674 提交于
* add get and set for Grad_scaler * refine some API name and comments * refine API name and comments * refine some comments
-
- 30 6月, 2021 3 次提交
-
-
由 jakpiase 提交于
* added matmul_v2 bf16/fp32 FWD kernel added matmul_v2 bf16/fp32 FWD kernel * added formatting * removed some tests due to timeout in CI * refactored tests * merged tests classes into one file * minor change * removed test guard for CUDA * remove skipIf * changes after review * formated one file * minor change * added skipping UT in CUDA place
-
由 Aurelius84 提交于
* refine temp_scope_vec logic * polish partial_program * fix fake var * add stop_gradient in spec * fix fake_var * fix unittest
-
由 houj04 提交于
* support set_device for NPU. * minor update doc and add more unit test.
-
- 29 6月, 2021 4 次提交
-
-
由 liu zhengxi 提交于
* add returns for beamsearchdecoder doc, test=document_fix
-
由 Thunderbrook 提交于
* remove heterbox * remove heterbox
-
由 Zhou Wei 提交于
-
由 taixiurong 提交于
-
- 28 6月, 2021 3 次提交
- 25 6月, 2021 2 次提交
- 24 6月, 2021 8 次提交
-
-
由 houj04 提交于
* in NPU environment, use CPUPlace for missing operators. * in NPU environment, use CPUPlace for missing operators. * fix TensorCopy bug and add unit test. * fix code style. * add more unit tests.
-
由 Jacek Czaja 提交于
* - fix to #33282 * - Increased threshold for elementwise_mul_bf16 grad * -disabled faulty UT * - fix to approval
-
由 Aurelius84 提交于
* support type hint * fix unittest
-
由 WeiXin 提交于
* TestSaveLoadLargeParameters use cpu place. * edit unittest
-
由 Jiangxinz 提交于
-
由 Jiangxinz 提交于
-
由 XGZhang 提交于
-
由 CtfGo 提交于
As the title
-