Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

kunlunxin berfLarge inference configs && results #212

Merged
merged 7 commits into from
Sep 14, 2023
Merged

kunlunxin berfLarge inference configs && results #212

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
cd9127c
kunlunxin inference : add bertLarge
Aug 24, 2023
3e44ed6
Revert "kunlunxin inference : add bertLarge"
Aug 24, 2023
25011bf
kunlunxin inference : add bertLarge
Aug 24, 2023
0d7df83
kunlunxin : remove re-install transformers
Aug 25, 2023
4e59add
adjust env for bertlarge
Aug 25, 2023
c6c9b03
kunlunxin: update bertLarge performance
Aug 29, 2023
b205883
Update BertLarge performance
jinxiangshi Sep 1, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 19 additions & 1 deletion inference/benchmarks/bertLarge/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,24 @@ bert_reference_results_text_md5.txt

- TensorRT 8.5.1.7

#### 2.2 昆仑芯R200

- ##### 硬件环境
- 机器、加速卡型号: R200

- ##### 软件环境
- OS版本：Ubuntu 20.04
- OS kernel版本: 5.15.0-56-generic
- 加速卡驱动版本：4.0
- Docker 版本：20.10.21
- 依赖软件版本：
- pytorch: 1.13.0+cpu
- onnx: 1.14.0

- 推理工具包

- XTCL 2.1

### 4. 运行情况（BERT-Large）

* 指标列表
Expand All @@ -64,5 +82,5 @@ bert_reference_results_text_md5.txt
| ----------- | --------- | ---- | ---- | -------- | ----------- | ---------- | ------------- | ------------ | ----------- | ----------- |
| tensorrt | fp16 | 32 | 1283.9 | 257.3 | 260.4 | 408.3 | 418.1 | 45.3% | 0.600/0.638 | 17.4/40.0 |
| tensorrt | fp32 | 32 | 1868.8 | 150.4 | 152.2 | 190.4 | 194.1 | 42.0% | 0.638/0.638 | 16.9/40.0 |

| kunlunxin_xtcl| fp32 | 32 | 6102.2 | None | None | 49.6 | 56 | 50% | 0.638/0.638| None|

3 changes: 3 additions & 0 deletions inference/configs/bertLarge/vendor_config/kunlunxin_configurations.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,3 @@
compiler: xtcl
no_validation: true
exist_onnx_path: onnxs/bertLarge/bertLarge_bs32_pytorch_fp16False.onnx
3 changes: 2 additions & 1 deletion inference/docker_images/kunlunxin/pytorch_1.13/Dockerfile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,7 +46,8 @@ RUN /root/miniconda/envs/python38/bin/pip install \
munch \
pyyaml \
tqdm \
scipy
scipy \
transformers
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

bert的requirements中已有transformers，这属于是bert这个case而非kunlunxin硬件或编译器需要的。请删除这一改动看看是否会遇到运行错误？

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

确认了下，不需要重复 install transformers 。新的patch去掉了这处改动


RUN /root/miniconda/envs/python38/bin/pip install torch==1.13.0+cpu torchvision==0.14.0+cpu torchaudio==0.13.0 --extra-index-url https://download.pytorch.org/whl/cpu

Expand Down