未验证 提交 cd74b207 编写于 作者: C cc 提交者: GitHub

Add more ops to calculate output scales (#32036)

上级 ed49b418
develop Ligoml-patch-1 OliverLPH-patch-1 OliverLPH-patch-2 PaddlePM-patch-1 PaddlePM-patch-2 ZHUI-patch-1 add_default_att add_model_benchmark_ci add_some_yaml_config addfile all_new_design_exec ascendrelease cherry_undefined_var compile_windows delete_add_default_att delete_all_new_design_exec delete_compile_windows delete_delete_addfile delete_disable_iterable_dataset_unittest delete_fix_dataloader_memory_leak delete_fix_retry_ci delete_fix_undefined_var delete_improve_sccache delete_paralleltest delete_prv-disable-more-cache delete_revert-33630-bug-fix delete_revert-34159-add_npu_bce_logical_dev delete_revert-34910-spinlocks_for_allocator delete_revert-35069-revert-34910-spinlocks_for_allocator delete_revert-36057-dev/read_flags_in_ut dingjiaweiww-patch-1 disable_iterable_dataset_unittest dy2static enable_eager_model_test final_state_gen_python_c final_state_intermediate fix-numpy-issue fix_concat_slice fix_dataloader_memory_leak fix_npu_ci fix_op_flops fix_retry_ci fix_rnn_docs fix_tensor_type fix_undefined_var fixiscan fixiscan1 fixiscan2 fixiscan3 improve_sccache incubate/infrt inplace_addto make_flag_adding_easier move_embedding_to_phi move_histogram_to_pten move_sgd_to_phi move_slice_to_pten move_temporal_shift_to_phi move_yolo_box_to_phi npu_fix_alloc paralleltest preln_ernie prv-disable-more-cache prv-md-even-more prv-onednn-2.5 pten_tensor_refactor release/2.1 release/2.2 release/2.3 release/2.3-fc-ernie-fix release/2.4 revert-32290-develop-hardlabel revert-33037-forci revert-33475-fix_cifar_label_dimension revert-33630-bug-fix revert-34159-add_npu_bce_logical_dev revert-34406-add_copy_from_tensor revert-34910-spinlocks_for_allocator revert-35069-revert-34910-spinlocks_for_allocator revert-36057-dev/read_flags_in_ut revert-36201-refine_fast_threaded_ssa_graph_executor revert-36985-add_license revert-37318-refactor_dygraph_to_eager revert-37926-eager_coreops_500 revert-37956-revert-37727-pylayer_support_tuple revert-38100-mingdong revert-38301-allocation_rearrange_pr revert-38703-numpy_bf16_package_reupload revert-38732-remove_useless_header_in_elementwise_mul_grad revert-38959-Reduce_Grad revert-39143-adjust_empty revert-39227-move_trace_op_to_pten revert-39268-dev/remove_concat_fluid_kernel revert-40170-support_partial_grad revert-41056-revert-40727-move_some_activaion_to_phi revert-41065-revert-40993-mv_ele_floordiv_pow revert-41068-revert-40790-phi_new revert-41944-smaller_inference_api_test revert-42149-do-not-reset-default-stream-for-stream-safe-cuda-allocator revert-43155-fix_ut_tempfile revert-43882-revert-41944-smaller_inference_api_test revert-45808-phi/simplify_size_op revert-46827-deform_comment support_weight_transpose zhiqiu-patch-1 v2.4.0-rc0 v2.3.2 v2.3.1 v2.3.0 v2.3.0-rc0 v2.2.2 v2.2.1 v2.2.0 v2.2.0-rc0 v2.2.0-bak0 v2.1.3 v2.1.2 v2.1.1 v2.1.0 v2.1.0-rc0
无相关合并请求
......@@ -468,7 +468,7 @@ class ImperativeQuantizeOutputs(object):
"""
Whether the layer needs to calculate output scales.
"""
return isinstance(layer, tuple(utils.quant_output_layers_map.values())) \
return isinstance(layer, utils.quant_output_layers) \
or ('quantized' in layer.full_name() and \
'quantized_noweight' not in layer.full_name())
......
......@@ -43,28 +43,18 @@ fake_quantize_dequantize_types = [
"fake_quantize_dequantize_moving_average_abs_max"
]
quant_output_layers_map = {
'Conv2D': paddle.nn.Conv2D,
'Conv2DTranspose': paddle.nn.Conv2DTranspose,
'Linear': paddle.nn.Linear,
'AdaptiveAvgPool2D': paddle.nn.AdaptiveAvgPool2D,
'AdaptiveMaxPool2D': paddle.nn.AdaptiveMaxPool2D,
'AvgPool2D': paddle.nn.AvgPool2D,
'MaxPool2D': paddle.nn.MaxPool2D,
'BatchNorm': paddle.nn.BatchNorm,
'BatchNorm2D': paddle.nn.BatchNorm2D,
'SyncBatchNorm': paddle.nn.SyncBatchNorm,
'ELU': paddle.nn.ELU,
'GELU': paddle.nn.GELU,
'LeakyReLU': paddle.nn.LeakyReLU,
'PReLU': paddle.nn.PReLU,
'ReLU': paddle.nn.ReLU,
'ReLU6': paddle.nn.ReLU6,
'Sigmoid': paddle.nn.Sigmoid,
'Softmax': paddle.nn.Softmax,
'Tanh': paddle.nn.Tanh,
'Swish': paddle.nn.Swish,
}
quant_output_layers = (
paddle.nn.Conv2D, paddle.nn.Conv2DTranspose, paddle.nn.Linear,
paddle.nn.AdaptiveAvgPool2D, paddle.nn.AdaptiveMaxPool2D,
paddle.nn.AvgPool2D, paddle.nn.MaxPool2D, paddle.nn.BatchNorm,
paddle.nn.BatchNorm2D, paddle.nn.LayerNorm, paddle.nn.SyncBatchNorm,
paddle.nn.ELU, paddle.nn.GELU, paddle.nn.Hardshrink, paddle.nn.Hardsigmoid,
paddle.nn.Hardswish, paddle.nn.Hardtanh, paddle.nn.LeakyReLU,
paddle.nn.LogSigmoid, paddle.nn.LogSoftmax, paddle.nn.Maxout,
paddle.nn.PReLU, paddle.nn.ReLU, paddle.nn.ReLU6, paddle.nn.SELU,
paddle.nn.Sigmoid, paddle.nn.Softmax, paddle.nn.Softplus,
paddle.nn.Softshrink, paddle.nn.Softsign, paddle.nn.Swish, paddle.nn.Tanh,
paddle.nn.Tanhshrink, paddle.nn.ThresholdedReLU, paddle.nn.Upsample)
weight_op_types = [
"conv2d", "depthwise_conv2d", "matmul", "conv2d_transpose",
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册
反馈
建议
客服 返回
顶部