Skip to content

Commit fc5cda0

Browse files
authored
Update vllm (#31)
An updated version of vllm is released to application collection registry. We no longer need the workaround.
1 parent 50a039f commit fc5cda0

1 file changed

Lines changed: 5 additions & 32 deletions

File tree

references/vllm-helm-overrides.adoc

Lines changed: 5 additions & 32 deletions
Original file line numberDiff line numberDiff line change
@@ -5,33 +5,6 @@
55

66
include::../snippets/helm-chart-overrides-intro.adoc[]
77

8-
[WARNING]
9-
.{vllm} deployment issue
10-
====
11-
There is a bug with the latest revisions of 0.3.2 `containers/lmcache-vllm-openai` and 0.9.1 `containers/vllm-openai` where the deployment fails with the error:
12-
[source]
13-
----
14-
ValueError: 'aimv2' is already used by a Transformers config, pick another name.
15-
----
16-
This will be fixed in later versions. However, you can use a specific tag revision that does not have this issue.
17-
As a workaround, update references to these image tags in your overrides YAML file.
18-
Follow the examples below with the following specific revision of the image.
19-
20-
For `containers/vllm-openai`:
21-
[source,yaml]
22-
----
23-
repository: "containers/vllm-openai"
24-
tag: "0.9.1-3.1"
25-
----
26-
27-
For `containers/lmcache-vllm-openai`:
28-
[source,yaml]
29-
----
30-
repository: "containers/lmcache-vllm-openaii"
31-
tag: "0.3.2-2.3"
32-
----
33-
====
34-
358
[#vllm-helm-overrides-minimal]
369
.Minimal configuration
3710
====
@@ -47,7 +20,7 @@ servingEngineSpec:
4720
- name: "phi3-mini-4k"
4821
registry: "dp.apps.rancher.io"
4922
repository: "containers/vllm-openai"
50-
tag: "0.9.1"
23+
tag: "0.13.0"
5124
imagePullPolicy: "IfNotPresent"
5225
modelURL: "microsoft/Phi-3-mini-4k-instruct"
5326
replicaCount: 1
@@ -153,7 +126,7 @@ servingEngineSpec:
153126
- name: "llama3" <.>
154127
registry: "dp.apps.rancher.io" <.>
155128
repository: "containers/vllm-openai" <.>
156-
tag: "0.9.1" <.>
129+
tag: "0.13.0" <.>
157130
imagePullPolicy: "IfNotPresent"
158131
modelURL: "meta-llama/Llama-3.1-8B-Instruct" <.>
159132
replicaCount: 1 <.>
@@ -295,7 +268,7 @@ servingEngineSpec:
295268
- name: "llama3"
296269
registry: "dp.apps.rancher.io"
297270
repository: "containers/vllm-openai"
298-
tag: "0.9.1"
271+
tag: "0.13.0"
299272
imagePullPolicy: "IfNotPresent"
300273
modelURL: "/models/llama-3.1-8b-it"
301274
replicaCount: 1
@@ -356,7 +329,7 @@ servingEngineSpec:
356329
- name: "llama3"
357330
registry: "dp.apps.rancher.io"
358331
repository: "containers/vllm-openai"
359-
tag: "0.9.1"
332+
tag: "0.13.0"
360333
imagePullPolicy: "IfNotPresent"
361334
modelURL: "meta-llama/Llama-3.1-8B-Instruct"
362335
replicaCount: 1
@@ -372,7 +345,7 @@ servingEngineSpec:
372345
- name: "mistral"
373346
registry: "dp.apps.rancher.io"
374347
repository: "containers/vllm-openai"
375-
tag: "0.9.1"
348+
tag: "0.13.0"
376349
imagePullPolicy: "IfNotPresent"
377350
modelURL: "mistralai/Mistral-7B-Instruct-v0.2"
378351
replicaCount: 1

0 commit comments

Comments
 (0)