From 0516da2b33d8a739d3e997bddf29a8784f79eaae Mon Sep 17 00:00:00 2001
From: Jaewook Lee <11328376+jaewooklee93@users.noreply.github.com>
Date: Tue, 16 Jul 2024 21:45:26 +0900
Subject: [PATCH] add translator example

---
 gemma-2-translator/README.md | 103 +++++++++++++++++++++++++++++++++++
 1 file changed, 103 insertions(+)
 create mode 100644 gemma-2-translator/README.md

diff --git a/gemma-2-translator/README.md b/gemma-2-translator/README.md
new file mode 100644
index 0000000..a84fdfd
--- /dev/null
+++ b/gemma-2-translator/README.md
@@ -0,0 +1,103 @@
+## `app.py`
+```python
+import json, re, requests
+
+example = '''# LLaMA.cpp HTTP Server
+
+Fast, lightweight, pure C/C++ HTTP server based on [httplib](https://github.com/yhirose/cpp-httplib), [nlohmann::json](https://github.com/nlohmann/json) and **llama.cpp**.
+
+Set of LLM REST APIs and a simple web front end to interact with llama.cpp.
+
+**Features:**
+ * LLM inference of F16 and quantum models on GPU and CPU
+ * [OpenAI API](https://github.com/openai/openai-openapi) compatible chat completions and embeddings routes
+ * Parallel decoding with multi-user support
+ * Continuous batching
+ * Multimodal (wip)
+ * Monitoring endpoints
+ * Schema-constrained JSON response format
+
+The project is under active development, and we are [looking for feedback and contributors](https://github.com/ggerganov/llama.cpp/issues/4216).'''
+
+prompt = {
+    'filename': '/ggerganov/llama.cpp/master/examples/server/README.md',
+    'en-US': {'src': example},
+    'ko-KR': '%PLACEHOLDER%'
+}
+prompt = json.dumps(prompt)
+prompt = re.sub(' "%PLACEHOLDER%(.*)', '', prompt)
+print('\n\tPrompt:\n')
+print(prompt)
+print('\n---\n')
+
+response = requests.post(
+    f'http://localhost:8080/completion',
+    stream=True,
+    json=dict(
+        prompt=prompt,
+        stream=True,
+        json_schema=dict(
+            type='object',
+            properties=dict(
+                src={'type': 'string'}
+            )
+        )
+    )
+)
+
+print('\tResponse:\n')
+content = ''
+for line in response.iter_lines():
+    if line := re.match('data: (.*)', line.decode()):
+        line = json.loads(line.group(1))
+        print(line['content'], end='', flush=True)
+        
+        content += line['content']
+
+result = json.loads(content)['src']
+print('\n---\n')
+
+print('\tMarkdown:\n')
+try:
+    # IPython 환경인지 확인
+    ipython = get_ipython()
+    from IPython.display import Markdown
+    display(Markdown(result))
+except:
+    print(result)
+```
+
+```sh
+$ python app.py 
+
+        Prompt:
+
+{"filename": "/ggerganov/llama.cpp/master/examples/server/README.md", "en-US": {"src": "# LLaMA.cpp HTTP Server\n\nFast, lightweight, pure C/C++ HTTP server based on [httplib](https://github.com/yhirose/cpp-httplib), [nlohmann::json](https://github.com/nlohmann/json) and **llama.cpp**.\n\nSet of LLM REST APIs and a simple web front end to interact with llama.cpp.\n\n**Features:**\n * LLM inference of F16 and quantum models on GPU and CPU\n * [OpenAI API](https://github.com/openai/openai-openapi) compatible chat completions and embeddings routes\n * Parallel decoding with multi-user support\n * Continuous batching\n * Multimodal (wip)\n * Monitoring endpoints\n * Schema-constrained JSON response format\n\nThe project is under active development, and we are [looking for feedback and contributors](https://github.com/ggerganov/llama.cpp/issues/4216)."}, "ko-KR":
+
+---
+
+        Response:
+
+{"src": "# LLaMA.cpp HTTP 서버\n\n[httplib](https://github.com/yhirose/cpp-httplib), [nlohmann::json](https://github.com/nlohmann/json) 및 **llama.cpp** 기반의 빠르고 가벼운 순수 C/C++ HTTP 서버입니다.\n\nLLM REST API 세트 및 llama.cpp와 상호 작용하기 위한 간단한 웹 프런트 엔드입니다.\n\n**특징:**\n * GPU 및 CPU에서 F16 및 양자 모델의 LLM 유추\n * [OpenAI API](https://github.com/openai/openai-openapi) 호환되는 채팅 완성 및 잠재형 엔드포인트\n * 다중 사용자 지원을 가진 병렬 디코딩\n * 지속적인 배치 처리\n * 다중 모달 (준비 중)\n * 모니터링 엔드포인트\n * 스키마 제약 JSON 응답 형식\n\n이 프로젝트는 적극적으로 개발 중이며, [피드백 및 기여자를 찾고 있습니다](https://github.com/ggerganov/llama.cpp/issues/4216)."}
+
+---
+
+
+        Markdown:
+```
+# LLaMA.cpp HTTP 서버
+
+[httplib](https://github.com/yhirose/cpp-httplib), [nlohmann::json](https://github.com/nlohmann/json) 및 **llama.cpp** 기반의 빠르고 가벼운 순수 C/C++ HTTP 서버입니다.
+
+LLM REST API 세트 및 llama.cpp와 상호 작용하기 위한 간단한 웹 프런트 엔드입니다.
+
+**특징:**
+ * GPU 및 CPU에서 F16 및 양자 모델의 LLM 유추
+ * [OpenAI API](https://github.com/openai/openai-openapi) 호환되는 채팅 완성 및 잠재형 엔드포인트
+ * 다중 사용자 지원을 가진 병렬 디코딩
+ * 지속적인 배치 처리
+ * 다중 모달 (준비 중)
+ * 모니터링 엔드포인트
+ * 스키마 제약 JSON 응답 형식
+
+이 프로젝트는 적극적으로 개발 중이며, [피드백 및 기여자를 찾고 있습니다](https://github.com/ggerganov/llama.cpp/issues/4216).
\ No newline at end of file