jandres/GenAIExamples
1
0
Fork 0
Code Issues Pull Requests Actions 5 Packages Projects Releases Wiki Activity

[ChatQnA] Update OOB with wrapper manifests. (#823)

Browse Source
This commit is contained in:
Zhenzhong1
2024-09-19 11:03:10 +08:00
committed by GitHub
parent 88829c9381
commit 933c3d3445
6 changed files with 0 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ChatQnA/benchmark/oob/with_rerank/four_gaudi/oob_four_gaudi_with_rerank.yaml
View File

@@ -275,10 +275,6 @@ spec:
- '2048'
- --max-total-tokens
- '4096'
- --max-batch-total-tokens
- '65536'
- --max-batch-prefill-tokens
- '4096'
env:
- name: OMPI_MCA_btl_vader_single_copy_mechanism
value: none

4
ChatQnA/benchmark/oob/with_rerank/single_gaudi/oob_single_gaudi_with_rerank.yaml
View File

@@ -275,10 +275,6 @@ spec:
- '2048'
- --max-total-tokens
- '4096'
- --max-batch-total-tokens
- '65536'
- --max-batch-prefill-tokens
- '4096'
env:
- name: OMPI_MCA_btl_vader_single_copy_mechanism
value: none

4
ChatQnA/benchmark/oob/with_rerank/two_gaudi/oob_two_gaudi_with_rerank.yaml
View File

@@ -275,10 +275,6 @@ spec:
- '2048'
- --max-total-tokens
- '4096'
- --max-batch-total-tokens
- '65536'
- --max-batch-prefill-tokens
- '4096'
env:
- name: OMPI_MCA_btl_vader_single_copy_mechanism
value: none

4
ChatQnA/benchmark/oob/without_rerank/four_gaudi/oob_four_gaudi_without_rerank.yaml
View File

@@ -336,10 +336,6 @@ spec:
- '2048'
- --max-total-tokens
- '4096'
- --max-batch-total-tokens
- '65536'
- --max-batch-prefill-tokens
- '4096'
volumeMounts:
- mountPath: /data
name: model-volume

4
ChatQnA/benchmark/oob/without_rerank/single_gaudi/oob_single_gaudi_without_rerank.yaml
View File

@@ -336,10 +336,6 @@ spec:
- '2048'
- --max-total-tokens
- '4096'
- --max-batch-total-tokens
- '65536'
- --max-batch-prefill-tokens
- '4096'
volumeMounts:
- mountPath: /data
name: model-volume

4
ChatQnA/benchmark/oob/without_rerank/two_gaudi/oob_two_gaudi_without_rerank.yaml
View File

@@ -336,10 +336,6 @@ spec:
- '2048'
- --max-total-tokens
- '4096'
- --max-batch-total-tokens
- '65536'
- --max-batch-prefill-tokens
- '4096'
volumeMounts:
- mountPath: /data
name: model-volume