- 14 8月, 2023 4 次提交
-
-
由 周波涛 提交于
-
由 jiangfan06 提交于
-
由 Sonder 提交于
-
由 MarDino 提交于
* add rmsnorm residual bias add and quant * refine python interface * add rmsnorm unittest * Add layernorm * fix layernorm unittest * refine unittest * fix example code * fix review comment
-
- 11 8月, 2023 4 次提交
-
-
由 umiswing 提交于
* Add enforce for sparse_bn. * Add enforce for sp conv.
-
由 Yiqun Liu 提交于
* Fix the shape of input sin and cos for fused_rope. * Update shape in unittest.
-
由 wz1qqx 提交于
-
由 hong19860320 提交于
-
- 10 8月, 2023 5 次提交
-
-
由 LiYuRio 提交于
-
由 jiangfan06 提交于
-
由 lzy 提交于
* add variable_length_memory_efficient_attention * update variable_length_memory_efficient_attention unittest * update variable_length_mem_eff_attn's docs and unittest * update variable_length_mem_eff_attn's docs * Update test_variable_length_memory_efficient_attention.py * Update variable_length_memory_efficient_attention.cu * fix codestyle * fix variable_length_fmha's docs and unittest * fix variable_length_fmha's docs
-
由 lxd-cumt 提交于
* decompose tanh_triple_grad and add it into prim_white_list test=develop * fix TanhTripleGradKernel bugs test=develop * decompose tanh_triple_grad test=develop
-
由 RuohengMa 提交于
* [XPU kernel] fix warpctc issue * fix issue * temporal hack to circumvent depthwise_conv2d precision issue * reset test case
-
- 09 8月, 2023 6 次提交
-
-
由 Xinyu Chen 提交于
* onednn: rename macro to PADDLE_WITH_DNNL * onednn: rename macro to CINN_WITH_DNNL
-
由 cyberslack_lee 提交于
-
由 niuliling123 提交于
-
由 houj04 提交于
-
由 houj04 提交于
* [XPU] add pos_weight for sigmoid_cross_entropy_with_logits. * update xdnn version.
-
由 ronnywang 提交于
-
- 08 8月, 2023 6 次提交
-
-
由 Wang Xin 提交于
* move decayed_adagrad_op to phi * fix bug
-
由 huangjiyi 提交于
-
由 leolishaohao 提交于
-
由 freeliuzc 提交于
-
由 niuliling123 提交于
-
由 hong 提交于
-
- 07 8月, 2023 4 次提交
-
-
由 yin wei 提交于
* add mask * add backword * add enforce info * update scale * integrate code * update enforce * add enforce eq * add error type * update enforce * add test_flash_attention * Polish codes and fix compiling errors. * Set num_splits to 0 for flash-attn with tensor mask. * Fix the compiling error for non flash-attn case. --------- Co-authored-by: NLiu Yiqun <liuyiqun01@baidu.com>
-
由 gouzil 提交于
-
由 Ruibin Cheung 提交于
-
由 umiswing 提交于
* Work for fa-2 padded fwd. Code to be cleaned. * Work for fa2 unpadded fwd. * Work for padded-bwd, dk get small diff on np.random.seed(0) * Anyway I pass paddle's utest, except return softmax without dropout. * Clean code. * Modify interface. * Clean code and add some check. * Easy compile for dev. * Fix ci. * Fix ci-build. * Add std c++17 option again. * Limit max job when compiling fa2. * Remove const_cast * Add fwd params, to be cleaned. * Clean code. * Add bwd params. * Clean code. * Add enforce. * Use v2.0.4 * Pass RNG state to fa2 capi * Fix review. * Add assert * Skip compile for sm less than 80.
-
- 04 8月, 2023 4 次提交
-
-
由 kangguangli 提交于
* fix bug: feed_with_place should consider variable existence * fix * fix build scope * change method to set feed var name * remove feed_with_place to placeholder * fix * rename to data * fix * fix
-
由 hong 提交于
* refacot aot * update * fix bugs * remove some test * fix bug * fix bug * fix bug * fix bug * update
-
由 Zhenghai Zhang 提交于
-
由 jiangfan06 提交于
-
- 03 8月, 2023 7 次提交
-
-
由 Yuang Liu 提交于
-
由 yangguohao 提交于
-
由 Wang Xin 提交于
-
由 wz1qqx 提交于
-
由 hong19860320 提交于
-
由 wanghuancoder 提交于
* fix security bug
-
由 wanghuancoder 提交于
* fix security bug
-