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

removed benchmark template

Browse Source
This commit is contained in:
Zhenzhong1
2024-10-23 09:30:03 +03:00
parent 2876677214
commit 1046aad26f
25 changed files with 2 additions and 1212 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ChatQnA/benchmark/performance/helm_charts/deployment.py
View File

@@ -59,7 +59,7 @@ def generate_yaml(num_nodes, mode="oob", with_rerank="True"):
else None
),
{"name": "llm-dependency-deploy", "resources": {"limits": {"habana.ai/gaudi": 1}}},
{"name": "retriever-deploy", "resources": {"requests": {"cpu": "16", "memory": "8000Mi"}}},
{"name": "retriever-deploy", "resources": {"requests": {"cpu": "8", "memory": "8000Mi"}}},
]
replicas = [replica for replica in replicas if replica]
@@ -72,7 +72,7 @@ def generate_yaml(num_nodes, mode="oob", with_rerank="True"):
{"name": "--model-id", "value": "$(LLM_MODEL_ID)"},
{"name": "--max-input-length", "value": 1280},
{"name": "--max-total-tokens", "value": 2048},
{"name": "--max-batch-total-tokens", "value": 35536},
{"name": "--max-batch-total-tokens", "value": 65536},
{"name": "--max-batch-prefill-tokens", "value": 4096},
],
},