valknar/runpod-ai-orchestrator
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: adjust VRAM allocation for concurrent Llama+BGE
All checks were successful
Build and Push RunPod Docker Image / build-and-push (push) Successful in 13s
Details

Browse Source 
- Llama: 85% GPU, 8K context (model needs ~15GB base)
- BGE: 10% GPU (1.3GB model)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Sebastian Krüger
2025-11-30 20:16:00 +01:00
parent f668e06228
commit b2de3b17ee
2 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
services/vllm/config_bge.yaml
View File

@@ -2,6 +2,6 @@ model: BAAI/bge-large-en-v1.5
host: "0.0.0.0" host: "0.0.0.0"
port: 8002 port: 8002
uvicorn-log-level: "info" uvicorn-log-level: "info"
gpu-memory-utilization: 0.15 gpu-memory-utilization: 0.10
dtype: float16 dtype: float16
task: embed task: embed

4
services/vllm/config_llama.yaml
View File

@@ -2,8 +2,8 @@ model: meta-llama/Llama-3.1-8B-Instruct
host: "0.0.0.0" host: "0.0.0.0"
port: 8001 port: 8001
uvicorn-log-level: "info" uvicorn-log-level: "info"
gpu-memory-utilization: 0.70 gpu-memory-utilization: 0.85
max-model-len: 16384 max-model-len: 8192
dtype: auto dtype: auto
enforce-eager: false enforce-eager: false
enable-auto-tool-choice: true enable-auto-tool-choice: true