Agent 2.0 exdev 서버 배포 스택

- server-dev start/stop/deploy 및 Gitea push 자동 배포 - local-dev 로컬 개발 환경 Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-21 22:57:30 +09:00
commit 4b86b2a660
344 changed files with 45787 additions and 0 deletions
@@ -0,0 +1,78 @@
+=========== embedding ===========
+
+docker run -d --gpus all \
+  --name tei-embedding \
+  -p 16001:80 \
+  -v /DATA/exlink/models:/data \
+  --pull never \
+  --restart always \
+  ghcr.io/huggingface/text-embeddings-inference:hopper-1.8 \
+  --model-id /data/qwen3-embedding-8b \
+  --port 80 \
+  --max-batch-tokens 40960
+
+
+curl 127.0.0.1:16001/embed \
+  -X POST \
+  -d '{"inputs":"안녕하세요, 임베딩 테스트 중입니다."}' \
+  -H 'Content-Type: application/json'
+=========== embedding ===========
+
+=========== reranker ===========
+
+docker rm -f vllm-reranker
+
+docker rm -f vllm-reranker
+
+docker run -d --gpus all \
+  --name vllm-reranker \
+  -p 16002:8000 \
+  -v /DATA/exlink/models:/data \
+  --ipc=host \
+  --restart always \
+  -e HF_HUB_OFFLINE=1 \
+  -e TRANSFORMERS_OFFLINE=1 \
+  vllm/vllm-openai:latest \
+  python3 /data/reranker_server.py
+
+curl http://127.0.0.1:16002/v1/rerank \
+  -X POST \
+  -H "Content-Type: application/json" \
+  -d '{
+"model": "/data/qwen3-reranker-4b",
+"query": "사과는 어떤 과일인가요?",
+"documents": [
+"사과는 빨간색 과일이며 비타민이 풍부합니다.",
+"서울은 대한민국의 수도입니다."
+]
+}'
+=========== reranker ===========
+
+
+=========== LLM ===========
+
+docker run -d --gpus all \ 
+  --name vllm-server \ 
+  --memory 32g \ 
+  -p 16000:8000 \ 
+  -v /DATA/exlink/models:/data \ 
+  --ipc=host \ --restart always \ 
+  --network bridge \ 
+  -e HF_HUB_OFFLINE=1 \ 
+  -e HF_HOME=/data/cache \ 
+  -e TRANSFORMERS_OFFLINE=1 \ 
+  --entrypoint python3 \ 
+  vllm/vllm-openai:latest \ 
+  -m vllm.entrypoints.openai.api_server \ 
+  --model /data/Qwen3.5-27B-FP8 \ 
+  --host 0.0.0.0 \ 
+  --port 8000 \ 
+  --gpu-memory-utilization 0.65 \ 
+  --max-model-len 5120 \ 
+  --max-num-seqs 8 \ 
+  --enable-chunked-prefill \ 
+  --max-num-batched-tokens 4096 \ 
+
+  =========== LLM ===========
+
+
@@ -0,0 +1,80 @@
+import math
+import os
+import torch
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel
+from typing import List
+from vllm import LLM, SamplingParams
+from vllm.inputs.data import TokensPrompt
+from transformers import AutoTokenizer
+
+# 오프라인 환경 변수 강제 설정
+os.environ["HF_HUB_OFFLINE"] = "1"
+os.environ["TRANSFORMERS_OFFLINE"] = "1"
+
+app = FastAPI()
+
+# 1. 모델 경로 (컨테이너 내부 경로 기준)
+MODEL_PATH = "/data/qwen3-reranker-4b"
+
+# 토크나이저 및 모델 로드
+# vllm-openai 이미지에는 이미 transformers, vllm, fastapi가 들어있습니다.
+tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH, local_files_only=True)
+model = LLM(
+    model=MODEL_PATH,
+    gpu_memory_utilization=0.10, # H100 80GB 중 24GB 점유 (남은 공간은 80B 모델용)
+    max_model_len=4096,
+    max_num_seqs=20,    
+    trust_remote_code=True,
+    dtype="float16",
+    enforce_eager=True # 오프라인 환경에서 불필요한 커널 컴파일 방지
+)
+
+# 토큰 설정
+true_token = tokenizer("yes", add_special_tokens=False).input_ids[0]
+false_token = tokenizer("no", add_special_tokens=False).input_ids[0]
+suffix = "<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n"
+suffix_tokens = tokenizer.encode(suffix, add_special_tokens=False)
+
+class RerankRequest(BaseModel):
+    query: str
+    documents: List[str]
+
+@app.post("/rerank")
+async def rerank(request: RerankRequest):
+    task = 'Given a web search query, retrieve relevant passages that answer the query'
+    prompts = []
+    
+    for doc in request.documents:
+        messages = [
+            {"role": "system", "content": "Judge whether the Document meets the requirements based on the Query and the Instruct provided. Note that the answer can only be \"yes\" or \"no\"."},
+            {"role": "user", "content": f"<Instruct>: {task}\n\n<Query>: {request.query}\n\n<Document>: {doc}"}
+        ]
+        token_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=False)
+        # 길이 제한 및 suffix 추가
+        token_ids = token_ids[:8192 - len(suffix_tokens)] + suffix_tokens
+        prompts.append(TokensPrompt(prompt_token_ids=token_ids))
+
+    sampling_params = SamplingParams(
+        temperature=0, max_tokens=1, logprobs=20, 
+        allowed_token_ids=[true_token, false_token]
+    )
+
+    outputs = model.generate(prompts, sampling_params, use_tqdm=False)
+    
+    results = []
+    for i, output in enumerate(outputs):
+        final_logits = output.outputs[0].logprobs[-1]
+        t_logit = final_logits[true_token].logprob if true_token in final_logits else -10.0
+        f_logit = final_logits[false_token].logprob if false_token in final_logits else -10.0
+        
+        t_score = math.exp(t_logit)
+        f_score = math.exp(f_logit)
+        score = t_score / (t_score + f_score)
+        results.append({"index": i, "score": score})
+    
+    return {"results": sorted(results, key=lambda x: x['score'], reverse=True)}
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=8000)