updated shuguang InferDeployInstance logic

Former-commit-id: 211870e061ed7fda363887f4818472ccad8fd9d5
This commit is contained in:
tzwang 2024-09-12 17:35:16 +08:00
parent 565c1455d2
commit 5a15db9499
2 changed files with 26 additions and 2 deletions

View File

@ -84,10 +84,10 @@ var (
CardModelNameCmdMap = map[string]map[string]string{
BIV100: {"blip-image-captioning-base": "pip install -U transformers; pip install fastapi uvicorn[standard]; pip install python-multipart; cd /code; python infer_biv100.py",
"imagenet_resnet50": "pip install -U transformers; pip install fastapi uvicorn[standard]; pip install python-multipart; cd /code/infer; python infer_biv100.py",
"chatGLM_6B": "su root; pip install transformers==4.33.2; pip install fastapi uvicorn[standard]; cd /code; python infer_biv100.py"},
"ChatGLM_6B": "su root; pip install transformers==4.33.2; pip install fastapi uvicorn[standard]; cd /code; python infer_biv100.py"},
MLU: {"blip-image-captioning-base": "",
"imagenet_resnet50": "su root; . /torch/venv3/pytorch/bin/activate; pip install fastapi uvicorn[standard]; pip install python-multipart; cd /code/infer; python infer_mlu.py",
"chatGLM_6B": ""},
"ChatGLM_6B": ""},
}
)

View File

@ -24,6 +24,7 @@ import (
"gitlink.org.cn/JointCloud/pcm-coordinator/internal/scheduler/service/collector"
"gitlink.org.cn/JointCloud/pcm-coordinator/internal/scheduler/service/inference"
"gitlink.org.cn/JointCloud/pcm-coordinator/internal/types"
"gitlink.org.cn/JointCloud/pcm-coordinator/pkg/constants"
"gitlink.org.cn/JointCloud/pcm-coordinator/pkg/utils"
"mime/multipart"
"strconv"
@ -867,6 +868,25 @@ func (s *ShuguangAi) GetInferDeployInstance(ctx context.Context, id string) (*in
if err != nil || resp.Code != "0" {
return nil, err
}
var url string
if resp.Data.Status == constants.Running {
url = resp.Data.ContainerPortInfoList[0].ContentPath
}
var modelType string
var modelName string
var card string
if resp.Data.Description != "" {
str := strings.Split(resp.Data.Description, FORWARD_SLASH)
if len(str) == 3 {
modelType = str[0]
modelName = str[1]
card = str[2]
}
}
ins.InstanceName = resp.Data.InstanceServiceName
ins.InstanceId = resp.Data.Id
ins.ClusterName = s.platform
@ -874,6 +894,10 @@ func (s *ShuguangAi) GetInferDeployInstance(ctx context.Context, id string) (*in
ins.InferCard = DCU
ins.CreatedTime = resp.Data.CreateTime
ins.ClusterType = TYPE_SHUGUANGAI
ins.ModelType = modelType
ins.ModelName = modelName
ins.InferUrl = url
ins.InferCard = card
return ins, nil
}