apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 08:37:06 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Other] Deprecate some option api and parameters (#1243)

Browse Source 
* Optimize Poros backend

* fix error

* Add more pybind

* fix conflicts

* add some deprecate notices

* [Other] Deprecate some apis in RuntimeOption (#1240)

* Deprecate more options

* modify serving

* Update option.h

* fix tensorrt error

* Update option_pybind.cc

* Update option_pybind.cc

* Fix error in serving

* fix word spell error
This commit is contained in:
Jason
2023-02-07 17:57:46 +08:00
committed by GitHub
parent a18cc0f94c
commit 713afe7f1c
15 changed files with 380 additions and 229 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
serving/docs/EN/model_configuration-en.md
View File

@@ -162,7 +162,8 @@ optimization {
gpu_execution_accelerator : [
{
name : "tensorrt"
# Use FP16 inference in TensorRT. You can also choose: trt_fp32, trt_int8
# Use FP16 inference in TensorRT. You can also choose: trt_fp32
# If the loaded model is a quantized model, this precision will be int8 automatically
parameters { key: "precision" value: "trt_fp16" }
}
]
@@ -203,4 +204,4 @@ optimization {
}
]
}}
```
```

3
serving/docs/zh_CN/model_configuration.md
View File

@@ -162,7 +162,8 @@ optimization {
gpu_execution_accelerator : [
{
name : "tensorrt"
# 使用TensorRT的FP16推理,其他可选项为: trt_fp32、trt_int8
# 使用TensorRT的FP16推理,其他可选项为: trt_fp32
# 如果加载的是量化模型，此精度设置无效，会默认使用int8进行推理
parameters { key: "precision" value: "trt_fp16" }
}
]