-
Notifications
You must be signed in to change notification settings - Fork 1k
[Feat][Serving] Add PP-ShiTuV2 and face recognition server apps #2488
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.
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
[Feat][Serving] Add PP-ShiTuV2 and face recognition server apps #2488
Conversation
Thanks for your contribution! |
2fea5c6
to
cdd1ce0
Compare
@@ -391,9 +391,10 @@ data_root # 数据集根目录,目录名称可以改变 | |||
|
|||
<details><summary>API参考</summary> | |||
|
|||
<p>对于服务提供的所有操作:</p> | |||
<p>对于服务提供的主要操作:</p> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
考虑到与健康检查等辅助操作区分,将与产线推理API直接绑定的操作定位为“主要操作”,并作为统一规范以减小产线间差异。对各产线文档进行了修改。
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
import abc |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
原有的file_storage
主要考虑将文件存储后返回某种用户可以用于获取文件的表示(base64编码或URL)。考虑到shitu产线对“状态”的需求(需要对index进行CRUD),这里重新做了抽象,以支持可复用的外部存储/缓存功能,同时保留一定可扩展性。
return Image.open(io.BytesIO(data)) | ||
|
||
|
||
def image_to_bytes(image: Image.Image, format: str = "JPEG") -> bytes: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
换用更细粒度的API(有时需要或者原始bytes,而不进行编码)。
* modify npu predict bug (#2454) * Support serving vehicle attribute recognition pipeline (#2452) * fix bug (#2450) * add ad practise tuturial (#2465) * add ad practise tuturial * add ad practise tuturial * add ad practise tuturial * update hpip support list (#2478) * update hpip support list * optimize * update README.md hpip support list * update * fix api reference (#2475) * Doc update (#2479) * update docs * update index * modify npu models infer bug (#2474) * modify pretrained model link (#2441) * add model url for docs (#2466) * add model url for docs * update * update * update new ir blacklist for adapting paddle 3.0b2 * support serving * update: 1. change the default delimiter to space; 2. add en doc; 3. update official test dataset link; 4. fix some error in doc; * remove and append support serving * update faiss api * rm unnecessary arg in remove_index * upgrade api to save vector and id_map independently * save index_type to index_info.yaml * update doc * use index_info to cover other data except index * fix pp-yoloe-sod models train bug (#2487) * bugfix: dont pass index_type and metric_type in append_index() * fix bug where acc is always 0 during SLANet training (#2482) * fix unit to ms in benchmark csv file * update benchmark doc * [Feat][Serving] Add PP-ShiTuV2 and face recognition server apps (#2488) * Add PP-ShiTuV2 FastAPI app * Fix index_type and metric_type * Add face recognition app * Change small object detection endpoint * Fix face recognition bugs
No description provided.