Spaces:
Build error
Build error
File size: 2,348 Bytes
5612054 cc3c56d 53b6f17 cc3c56d 55913e8 53b6f17 e2b0585 53b6f17 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 |
#!/bin/bash
# Configuración
MODEL_PATH="/home/agent_h/data/starvector-1b-im2svg"
CHAT_TEMPLATE="configs/chat-template.jinja"
MAX_MODEL_LEN=8192
# SSH Tunnels (opcionales)
bash -c "$SSH_TUNNEL_CMD_1" &
bash -c "$SSH_TUNNEL_CMD_2" &
bash -c "$SSH_TUNNEL_CMD_3" &
bash -c "$SSH_TUNNEL_CMD_4" &
# Controller
python -m starvector.serve.vllm_api_gradio.controller --host 0.0.0.0 --port 10000 &
# Esperar a que el controller inicie
sleep 5
# ====== AÑADIR ESTOS COMANDOS - Iniciar servidores vLLM ======
echo "Iniciando servidores vLLM..."
vllm serve $MODEL_PATH \
--chat-template $CHAT_TEMPLATE \
--trust-remote-code \
--port 8000 \
--max-model-len $MAX_MODEL_LEN &
vllm serve $MODEL_PATH \
--chat-template $CHAT_TEMPLATE \
--trust-remote-code \
--port 8001 \
--max-model-len $MAX_MODEL_LEN &
vllm serve $MODEL_PATH \
--chat-template $CHAT_TEMPLATE \
--trust-remote-code \
--port 8002 \
--max-model-len $MAX_MODEL_LEN &
vllm serve $MODEL_PATH \
--chat-template $CHAT_TEMPLATE \
--trust-remote-code \
--port 8003 \
--max-model-len $MAX_MODEL_LEN &
# Esperar a que vLLM inicie (puede tardar varios minutos)
echo "Esperando a que vLLM cargue el modelo (esto puede tardar 2-3 minutos)..."
sleep 180
# ====== FIN DE LA MODIFICACIÓN ======
# Model Workers
python -m starvector.serve.vllm_api_gradio.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40000 --worker http://localhost:40000 --model-name $MODEL_PATH --vllm-base-url http://localhost:8000 &
python -m starvector.serve.vllm_api_gradio.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40001 --worker http://localhost:40001 --model-name $MODEL_PATH --vllm-base-url http://localhost:8001 &
python -m starvector.serve.vllm_api_gradio.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40002 --worker http://localhost:40002 --model-name $MODEL_PATH --vllm-base-url http://localhost:8002 &
python -m starvector.serve.vllm_api_gradio.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40003 --worker http://localhost:40003 --model-name $MODEL_PATH --vllm-base-url http://localhost:8003 &
# Gradio Web Server
python -m starvector.serve.vllm_api_gradio.gradio_web_server --controller http://localhost:10000 --model-list-mode reload --port 7860 |