호출하다 error나고 호출도 안한 함수 잘못 쓰고 있다, 에러라는 attribute가 없다 등의 이상한 에러를 내뱉어서구글링 해보니 이게 다 버전문제라고해서chatopenai 최신버전으로 업뎃하고, import도 warnning메세지를 보고 고쳤더니 된다. open AI 에러메세지는 아직 직관적이지 않아서, 난해한 편인듯. from langchain.chat_models import ChatOpenAIfrom langchain.llms import OpenAIfrom langchain_openai import ChatOpenAImodel_name = "gpt-3.5-turbo"temperature = 0.0llm = ChatOpenAI(model_name=model_name, ..
https://learn.microsoft.com/ko-kr/credentials/certifications/azure-ai-engineer/ Microsoft Certified: Azure AI Engineer Associate - Certifications Microsoft Azure AI 엔지니어는 Azure AI를 최대한 활용하는 AI 솔루션을 구축, 관리, 배포합니다. learn.microsoft.com https://learn.microsoft.com/ko-kr/credentials/certifications/exams/ai-102/ 시험 AI-102: Microsoft Azure AI 솔루션 설계 및 구현 - Certifications 시험 AI-102: Microsoft Azure AI 솔..
https://learn.microsoft.com/ko-kr/credentials/certifications/resources/study-guides/ai-900 시험 AI-900: Microsoft Azure AI 기본 사항 학습 가이드 시험 AI-900: Microsoft Azure AI 기본 사항 학습 가이드 | Microsoft Docs learn.microsoft.com https://learn.microsoft.com/ko-kr/credentials/certifications/exams/ai-900/ 시험 AI-900: Microsoft Azure AI Fundamentals - Certifications 시험 AI-900: Microsoft Azure AI Fundamentals learn...
![](http://i1.daumcdn.net/thumb/C148x148/?fname=https://blog.kakaocdn.net/dn/eVtEkA/btsFHnhzyxC/RBYJWx6IdEv0pVK53TaKiK/img.png)
개요 자연어 처리 영역에서 LLM의 성능을 평가하는 것은 중요한 측면입니다. 이 평가를 위한 주요 도구 중 하나는 ROUGE(Recall-Oriented Understudy for Gisting Evaluation) 측정항목입니다. ROUGE는 주로 LLM의 텍스트 생성 품질을 평가하는 데 사용됩니다. GPT-2와 같은 LLM은 종종 텍스트 완성이나 요약과 같은 작업에 참여합니다. 생성된 텍스트의 효율성은 확장성 및 일관성 문제로 인해 사람의 판단만으로는 효과적으로 측정할 수 없습니다. 예를 들어, 아래 코드를 실행하여 다음 프롬프트에 따라 텍스트를 생성하세요. 여기에 어떤 점수를 할당할 수 있는지 생각해 보고 다양한 텍스트에 점수를 매기는 표준화된 측정 기준을 생각해 보세요. # Generate tex..
한빛출판사 프리뷰어스 사이트에 기고한 내용입니다. https://www.notion.so/modeling-chatGPT-04fc8db66274499abcc2687f9891d89e chatGPT의 지식에 해당하는 부분은 말그대로 FACT, 사실이 기재되어있어야 한다. 정확하지 않은 정보가 아닌 경우, 사용자에게 신뢰도는 매우 하락할 수 있다. 또한 chatGPT에 사용할 LLM모델을 직접 만드는 경우, 대량의 문서를 학습시켜 만들지만, 일반적으로 규모가 작은 개발회사에서 적은 비용으로 LLM modeling을 바로 할 수가 없다. GPU보드 주문부터, 또는 Cloud서비스를 통해 instance를 갖추고 그 안에 LLM Modeling할 인프라부터 설계해야 된다. chatGPT를 위한 LLM생성 기본적인 ..
https://github.com/iarai/science4cast GitHub - iarai/science4cast: Details for the Science4Cast competition Details for the Science4Cast competition. Contribute to iarai/science4cast development by creating an account on GitHub. github.com 토치 심플 모델 #책 #파이토치 딥러닝 마스터 소스코드 https://github.com/deep-learning-with-pytorch/dlwpt-code GitHub - deep-learning-with-pytorch/dlwpt-code: Code for the book Deep..
https://www.cloudskillsboost.google/course_templates/536 Introduction to Generative AI - 한국어 | Google Cloud Skills Boost 생성형 AI란 무엇이고 어떻게 사용하며 전통적인 머신러닝 방법과는 어떻게 다른지 설명하는 입문용 마이크로 학습 과정입니다. 직접 생성형 AI 앱을 개발하는 데 도움이 되는 Google 도구에 대 www.cloudskillsboost.google All Readings: Introduction to Generative AI (G-GENAI-I) Here are the assembled readings on generative AI: ● Ask a Techspert: What is generat..
pip install numpy pillow matplotlib ipython tensorflow tensorflow_hub imageio Activation Function 활성화 함수 뉴런은 0 또는 1 두가지 값만 출력할 수 있는 기계 작은 숫자를 입력받으면 0에 가까운 숫자를 출력하고, 어느 정도 큰 숫자를 입력받으면 1에 가까운 숫자를 출력함 시그모이드 함수: 0~1 사이로 하이퍼볼릭 탄젠트 함수 -1~1 사잇값 탄젠트는 원점을 기준으로 대칭 렐루 ReLu : 0보다 작은 값은 0으로 만들고, 0보다 큰 값은 그대로 출력. 단일 뉴런 보다는 여러 개의 뉴런으로 구성된 신경 다발의 정보전달 형상. 기술기 소실문제 해결에도 사용 리키 렐루> 0 보다 작은 값을 버리지는 않고 원본 그대로 출력하지 않..
- Total
- Today
- Yesterday
- Paper
- mongodb
- linkedin-skill-assessments-quizzes
- pytorch
- tag hello
- 알고리즘
- 자바
- amazon
- LECTURE
- ML
- Shell
- quiz
- 통계
- geometry
- Neo4j
- 파이썬
- 기본
- LeetCode
- extract archive multiple files
- MachineLearning
- nvidia #gan
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
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 |