From 3a8c622f580ee2efbb0eb9a71eb5fc1df0eb2c93 Mon Sep 17 00:00:00 2001 From: Wang Guibao Date: Mon, 10 Jun 2019 10:04:27 +0800 Subject: [PATCH] Add documentation for runtime_thread_num (#8) --- doc/SERVING_CONFIGURE.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/doc/SERVING_CONFIGURE.md b/doc/SERVING_CONFIGURE.md index f5887f5c..f34436a0 100644 --- a/doc/SERVING_CONFIGURE.md +++ b/doc/SERVING_CONFIGURE.md @@ -165,7 +165,7 @@ reloadable_type: 检查reload条件：timestamp_ne/timestamp_gt/md5sum/revision/ model_data_path: 模型文件路径 -runtime_thread_num: 若大于0， 则启用bsf多线程调度框架，在每个预测bthread worker内启动多线程预测。 +runtime_thread_num: 若大于0， 则启用bsf多线程调度框架，在每个预测bthread worker内启动多线程预测。要注意的是，当启用worker内多线程预测，workflow中OP需要用Serving框架的BatchTensor类做预测的输入和输出 (predictor/framework/infer_data.h, `class BatchTensor`)。 batch_infer_size: 启用bsf多线程预测时，每个预测线程的batch size -- GitLab