Remove lmdeploy

helm/h2ogpt-chart/README.md

@@ -133,29 +133,6 @@ A Helm chart for h2oGPT
 | h2ogpt.storage.useEphemeral | bool | `true` |  |
 | h2ogpt.tolerations | string | `nil` |  |
 | h2ogpt.updateStrategy.type | string | `"RollingUpdate"` |  |
-| lmdeploy.containerArgs[0] | string | `"OpenGVLab/InternVL-Chat-V1-5"` |  |
-| lmdeploy.enabled | bool | `false` | Enable lmdeploy |
-| lmdeploy.env | object | `{}` |  |
-| lmdeploy.hfSecret | string | `nil` |  |
-| lmdeploy.image.pullPolicy | string | `"IfNotPresent"` |  |
-| lmdeploy.image.repository | string | `"gcr.io/vorvan/h2oai/h2oai-h2ogpt-lmdeploy"` |  |
-| lmdeploy.image.tag | string | `nil` |  |
-| lmdeploy.nodeSelector | string | `nil` |  |
-| lmdeploy.overrideConfig | string | `nil` |  |
-| lmdeploy.podAffinity | string | `nil` | Set hostname and zone to true for pod affinity rules based on hostname and zone. |
-| lmdeploy.podAnnotations | object | `{}` |  |
-| lmdeploy.podLabels | object | `{}` |  |
-| lmdeploy.podSecurityContext | string | `nil` |  |
-| lmdeploy.replicaCount | int | `1` |  |
-| lmdeploy.resources | string | `nil` |  |
-| lmdeploy.securityContext | string | `nil` |  |
-| lmdeploy.service.port | int | `23333` |  |
-| lmdeploy.service.type | string | `"ClusterIP"` |  |
-| lmdeploy.storage.class | string | `nil` |  |
-| lmdeploy.storage.size | string | `"512Gi"` |  |
-| lmdeploy.storage.useEphemeral | bool | `true` |  |
-| lmdeploy.tolerations | string | `nil` |  |
-| lmdeploy.updateStrategy.type | string | `"RollingUpdate"` |  |
 | nameOverride | string | `""` |  |
 | namespaceOverride | string | `""` |  |
 | tgi.containerArgs | string | `nil` |  |

helm/h2ogpt-chart/templates/h2ogpt-deployment.yaml

@@ -98,18 +98,7 @@ spec:
 
               python3 /workspace/generate.py
           {{- end }}
-          {{- if and .Values.lmdeploy.enabled (not .Values.global.externalLLM.modelLock) }}
-          args:
-            - >
-              until wget -O- http://{{ include "h2ogpt.fullname" . }}-lmdeploy-inference:{{ .Values.lmdeploy.service.port }}/ >/dev/null 2>&1;
-                do
-                  echo "Waiting for inference service to become ready...";
-                  sleep 5;
-                done
-
-              python3 /workspace/generate.py
-          {{- end }}
-          {{- if and .Values.h2ogpt.enabled (not (or .Values.vllm.enabled .Values.tgi.enabled .Values.lmdeploy.enabled)) }}
+          {{- if and .Values.h2ogpt.enabled (not (or .Values.vllm.enabled .Values.tgi.enabled )) }}
           args:
             - >
               python3 /workspace/generate.py
@@ -165,10 +154,6 @@ spec:
             - name: h2ogpt_inference_server
               value: "vllm:{{ include "h2ogpt.fullname" . }}-vllm-inference:{{ .Values.vllm.service.port }}"
           {{- end }}
-          {{- if and .Values.lmdeploy.enabled (not .Values.global.externalLLM.enabled) }}
-            - name: h2ogpt_inference_server
-              value: "http://{{ include "h2ogpt.fullname" . }}-lmdeploy-inference:{{ .Values.lmdeploy.service.port }}"
-          {{- end }}
           {{- range $key, $value := .Values.h2ogpt.env }}
             - name: "{{ $key }}"
               value: "{{ $value }}"

helm/h2ogpt-chart/templates/validations.yaml

@@ -1,12 +1,6 @@
 {{- if and .Values.vllm.enabled .Values.tgi.enabled }}
   {{- fail "Both TGI and vLLM cannot be enabled at the same time. Enable only one and try again" }}
 {{- end }}
-{{- if and .Values.vllm.enabled .Values.lmdeploy.enabled }}
-  {{- fail "Both lmdeploy and vLLM cannot be enabled at the same time. Enable only one and try again" }}
-{{- end }}
-{{- if and .Values.lmdeploy.enabled .Values.tgi.enabled }}
-  {{- fail "Both TGI and lmdeploy cannot be enabled at the same time. Enable only one and try again" }}
-{{- end }}
 {{- if and ( and .Values.h2ogpt.enabled .Values.h2ogpt.agents.enabled) .Values.agents.enabled }}
   {{- fail " Both agents in both h2ogpt.agents cannot be enabled. Enably only one and try again" }}
 {{- end }}

helm/h2ogpt-chart/values.yaml

@@ -351,48 +351,5 @@ vllm:
   podLabels: {}
 
 
-lmdeploy:
-  # -- Enable lmdeploy
-  enabled: false
-  replicaCount: 1
-
-  image:
-    repository: gcr.io/vorvan/h2oai/h2oai-h2ogpt-lmdeploy
-    tag:
-    pullPolicy: IfNotPresent
-  # -- Set hostname and zone to true for pod affinity rules based on hostname and zone.
-  podAffinity:
-    # hostname:
-    # zone:
-
-  storage:
-    size: 512Gi
-    class:
-    useEphemeral: true
-
-  overrideConfig:
-  hfSecret:
-  containerArgs:
-    - "OpenGVLab/InternVL-Chat-V1-5"
-
-  service:
-    type: ClusterIP
-    port: 23333
-
-  updateStrategy:
-    type: RollingUpdate
-
-  podSecurityContext:
-  securityContext:
-
-  resources:
-  nodeSelector:
-  tolerations:
-
-  env: {}
-
-  podAnnotations: {}
-  podLabels: {}
-
 # -- CA certs
 caCertificates: ""