@daocloud-proto/hydra 0.18.0-dev-15 → 0.18.0-dev-18

package/admin-api/maas/v1alpha1/maas.pb.ts

@@ -140,8 +140,18 @@ export type CreateMAASModelRequest = {
   tokenWeight?: string
   workspaceVisibility?: MAASWorkspaceVisibility
   capabilityRoutes?: MAASCapabilityRoute[]
+  modelRatelimitConfig?: ModelRatelimitConfig
 }
 
+
+type BaseModelRatelimitConfig = {
+  enabled?: boolean
+}
+
+export type ModelRatelimitConfig = BaseModelRatelimitConfig
+  & OneOf<{ rpm: number }>
+  & OneOf<{ tpm: number }>
+
 export type MAASModel = {
   model?: HydraAdmin_apiModelV1alpha1Model.Model
   upstreamModels?: UpstreamModel[]
@@ -156,6 +166,7 @@ export type MAASModel = {
   capabilityRoutes?: MAASCapabilityRoute[]
   healthStatus?: MAASModelHealthStatus
   clusterStatus?: HydraManagement_apiWorkspaceV1alpha1Workspace.ClusterStatus
+  modelRatelimitConfig?: ModelRatelimitConfig
 }
 
 export type LoadBalanceConfig = {
@@ -236,6 +247,7 @@ export type UpdateMAASModelRequest = {
   tokenWeight?: string
   workspaceVisibility?: MAASWorkspaceVisibility
   capabilityRoutes?: MAASCapabilityRoute[]
+  modelRatelimitConfig?: ModelRatelimitConfig
 }
 
 export type UpdateMAASModelStatusRequest = {

package/management-api/finetune/v1alpha1/finetuneconfig.pb.ts

@@ -8,15 +8,6 @@ import * as HydraCommonCommon from "../../../common/common.pb"
 import * as fm from "../../../fetch.pb"
 import * as GoogleProtobufTimestamp from "../../../google/protobuf/timestamp.pb"
 
-type Absent<T, K extends keyof T> = { [k in Exclude<keyof T, K>]?: undefined };
-type OneOf<T> =
-  | { [k in keyof T]?: undefined }
-  | (
-    keyof T extends infer K ?
-      (K extends string & keyof T ? { [k in K]: T[K] } & Absent<T, K>
-        : never)
-    : never);
-
 export enum TrainingStage {
   TRAINING_STAGE_UNSPECIFIED = "TRAINING_STAGE_UNSPECIFIED",
   SFT = "SFT",
@@ -138,19 +129,13 @@ export type DpoConfig = {
   refModel?: string
 }
 
-
-/* hydra modified */ export type BaseResourceConfig = {
+export type ResourceConfig = {
   gpuType?: string
-  gpuCount?: number
   cpu?: number
   memory?: number
-  perGpuMemory?: number
   gpuResourceValues?: {[key: string]: string}
 }
 
-export type ResourceConfig = BaseResourceConfig
-  & OneOf<{ gpuCores: number }>
-
 export type FineTuneConfig = {
   name?: string
   cluster?: string

package/management-api/model/v1alpha1/model.pb.ts

@@ -209,9 +209,6 @@ export type ModelServingSpec = {
   cpu?: number
   memory?: number
   gpuType?: string
-  gpuCount?: number
-  perGpuMemory?: number
-  gpuCores?: number
   gpuResourceValues?: {[key: string]: string}
 }
 

package/management-api/model_serving/v1alpha1/model_serving.pb.ts

@@ -145,13 +145,10 @@ export type ConfigMapMountConfig = {
   cpu?: number
   memory?: number
   gpuType?: string
-  gpuCount?: number
-  perGpuMemory?: number
   gpuResourceValues?: {[key: string]: string}
 }
 
 export type ResourceConfig = BaseResourceConfig
-  & OneOf<{ gpuCores: number }>
   & OneOf<{ shmSize: number }>
   & OneOf<{ enableShm: boolean }>
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@daocloud-proto/hydra",
-  "version": "v0.18.0-dev-15",
+  "version": "v0.18.0-dev-18",
   "description": "",
   "author": "",
   "license": "ISC"