注意事项
服务限制
请确保您账户内以下资源最低可开出数量为:
- EC2 vCPU – 16
- ml.g4dn.2xlarge for endpoint – 2
- ml.g4dn.2xlarge for training job – 1
- Amazon VPC – 1
- Amazon VPC subnet – 4
- Amazon VPC Internet Gateway – 1
- Amazon VPC NAT gateway – 2
- Amazon VPC EIP – 2
- Amazon VPC endpoint – 5
- Amazon VPC route table – 3
- Amazon VPC Security group – 10
- Amazon VPC interface endpoint – 5
- Amazon IAM roles – 60
- Amazon SSM – 40
- Amazon DynamoDB table – 10
- Amazon EFS – 1
- Amazon EFS access point – 1
- Amazon Lambda – 60
- Amazon Rest API Gateway – 1
- Amazon ECS service – 1
- Amazon ECS task – 1
- Amazon ECR private repository – 1
- Amazon S3 bucket – 1
- Amazon ACM public certificate – 1
资源冲突
如果账户中已存在以下资源,部署会出现冲突,麻烦提前删除:
- DynamoDB tables
- all-in-one-ai-industrial-model
- all-in-one-ai-training-job
- all-in-one-ai-model
- all-in-one-ai-endpoint
- all-in-one-ai-transform-job
- all-in-one-ai-pipeline
- all-in-one-ai-api
- Lambda functions
- all_in_one_ai_api
- all_in_one_ai_create_deploy_generic
- all_in_one_ai_deploy /deploy
- all_in_one_ai_create_train_generic
- all_in_one_ai_train /train
- all_in_one_ai_inference /inference
- all_in_one_ai_invoke_endpoint
- all_in_one_ai_s3 /s3
- all_in_one_ai_industrial_model /industrialmodel
- all_in_one_ai_sd_hypernetwork /sd/hypernetwork
- all_in_one_ai_sd_industrial_model /sd/industrialmodel
- all_in_one_ai_sd_model /sd/models
- all_in_one_ai_sd_user /sd/user
- all_in_one_ai_helper
- all_in_one_ai_sagemaker
- all_in_one_ai_delete_endpoint
- all_in_one_ai_describe_endpoint
- all_in_one_ai_list_endpoints
- all_in_one_ai_endpoint /endpoint
- all_in_one_ai_list_training_jobs
- all_in_one_ai_describe_training_job
- all_in_one_ai_training_job /triaining
- System Store Parameters
- parameters starting with /all_in_one_ai/*
- VPCs
- all-in-one-ai
原创文章,作者:奋斗,如若转载,请注明出处:https://blog.ytso.com/301720.html