fix: remove vllm embedding

services/vllm/config_bge.yaml

@@ -1,5 +0,0 @@
-model: BAAI/bge-large-en-v1.5
-host: "0.0.0.0"
-port: 8002
-uvicorn-log-level: "info"
-gpu-memory-utilization: 0.1

services/vllm/config_llama.yaml

@@ -2,5 +2,7 @@ model: meta-llama/Llama-3.1-8B-Instruct
 host: "0.0.0.0"
 port: 8001
 uvicorn-log-level: "info"
-gpu-memory-utilization: 0.9
+gpu-memory-utilization: 0.95
-max-model-len: 20480
+max-model-len: 32768
+dtype: auto
+enforce-eager: false

supervisord.conf

@@ -73,23 +73,6 @@ environment=HF_HOME="./.cache/vllm",HF_TOKEN="%(ENV_HF_TOKEN)s"
 priority=200
 stopwaitsecs=30
 
-# vLLM BGE Embedding Server (Port 8002)
-[program:bge]
-command=services/vllm/venv/bin/vllm serve --config services/vllm/config_bge.yaml
-directory=.
-autostart=false
-autorestart=true
-startretries=3
-stderr_logfile=.logs/bge.err.log
-stdout_logfile=.logs/bge.out.log
-stdout_logfile_maxbytes=50MB
-stdout_logfile_backups=10
-stderr_logfile_maxbytes=50MB
-stderr_logfile_backups=10
-environment=HF_HOME="./.cache/vllm",HF_TOKEN="%(ENV_HF_TOKEN)s"
-priority=201
-stopwaitsecs=30
-
 
 # AudioCraft Studio Service
 [program:audiocraft]
@@ -112,7 +95,7 @@ programs=comfyui,webdav-sync
 priority=100
 
 [group:vllm]
-programs=llama,bge
+programs=llama
 priority=200
 
 [group:audiocraft]