<div id="readability-page-1" class="page"><div>
                <div id="SE-9bff080a-dce6-4913-9d96-5943ad0e5f26">
                            <p id="SE-56005efa-ca3f-4f76-91c7-dcf0d6cf9169"><span id="SE-65746f54-1b4f-401c-b961-8dd4b55724b2">가장먼저, 블로그 포스트의 고유값인 log_no를 수집합니다.</span></p>
                        </div>                <div id="SE-3e98de26-d176-473d-9326-6ab06eaeb4ba">
                                <p><a href="#" onclick="return false;" data-linktype="img" data-linkdata="{&quot;id&quot; : &quot;SE-3e98de26-d176-473d-9326-6ab06eaeb4ba&quot;, &quot;src&quot; : &quot;https://mblogthumb-phinf.pstatic.net/MjAyNDA2MjNfMjcw/MDAxNzE5MTIxNjM5MjQ2.kpfuRu0TD0o2qXnFfW_7aQDu7T_dzh0uSLmHkulJ2Kkg.2XnC0JU1xnzmqK_76XnnjW5L5R3lkoTGLQsexoUrXXcg.PNG/%EC%8A%A4%ED%81%AC%EB%A6%B0%EC%83%B7_2024-06-23_12.53.17.png&quot;, &quot;originalWidth&quot; : &quot;2226&quot;, &quot;originalHeight&quot; : &quot;1404&quot;, &quot;linkUse&quot; : &quot;false&quot;, &quot;link&quot; : &quot;&quot;}">
                                        <img src="https://mblogthumb-phinf.pstatic.net/MjAyNDA2MjNfMjcw/MDAxNzE5MTIxNjM5MjQ2.kpfuRu0TD0o2qXnFfW_7aQDu7T_dzh0uSLmHkulJ2Kkg.2XnC0JU1xnzmqK_76XnnjW5L5R3lkoTGLQsexoUrXXcg.PNG/%EC%8A%A4%ED%81%AC%EB%A6%B0%EC%83%B7_2024-06-23_12.53.17.png?type=w800" data-width="693" data-height="437" alt="">
                                    </a>
                                </p>
                        </div>
                <div id="SE-bdbe2aa2-4923-44ab-8f1b-5e49d602b7ab">
                                    <p id="SE-aff940f8-dafb-4fbf-a894-5d8c62a227f9"><span id="SE-32f5d51e-4a37-4b79-bff1-24ad010e7d83">​</span></p><p id="SE-19b87129-ac68-4d79-a649-75a1a7d1ea67"><span id="SE-72ecd7c9-2026-4bc0-aa3e-9e35bf8adff1">​</span></p><p id="SE-66f38fd8-5de7-42e4-bad4-5435a5971df3"><span id="SE-e26ce8d5-76b5-4943-8940-4e658e735c4b">두 번째로 각 페이지를 임베딩하여 결과를 chroma에 저장합니다. 포스트에 필요없는 정보들을 제거하지 않고 페이지의 모든내용을 임베딩했습니다. </span></p>
                            </div>                <div id="SE-5d4966e7-fea6-44d1-bacb-2aa0132c778c">
                                <p><a href="#" onclick="return false;" data-linktype="img" data-linkdata="{&quot;id&quot; : &quot;SE-5d4966e7-fea6-44d1-bacb-2aa0132c778c&quot;, &quot;src&quot; : &quot;https://mblogthumb-phinf.pstatic.net/MjAyNDA2MjNfMTA1/MDAxNzE5MTIxNjYwMTA4.A37qKTo5UgLGCsltXAxKCiDPo7i3EsVoEu6ktTVEgC8g.M__MVx5mSUMLuXp0m8zXym3WqLGoikZVQqnvnTsYX0og.PNG/%EC%8A%A4%ED%81%AC%EB%A6%B0%EC%83%B7_2024-06-23_13.25.31.png&quot;, &quot;originalWidth&quot; : &quot;1586&quot;, &quot;originalHeight&quot; : &quot;599&quot;, &quot;linkUse&quot; : &quot;false&quot;, &quot;link&quot; : &quot;&quot;}">
                                        <img src="https://mblogthumb-phinf.pstatic.net/MjAyNDA2MjNfMTA1/MDAxNzE5MTIxNjYwMTA4.A37qKTo5UgLGCsltXAxKCiDPo7i3EsVoEu6ktTVEgC8g.M__MVx5mSUMLuXp0m8zXym3WqLGoikZVQqnvnTsYX0og.PNG/%EC%8A%A4%ED%81%AC%EB%A6%B0%EC%83%B7_2024-06-23_13.25.31.png?type=w800" data-width="693" data-height="261" alt="">
                                    </a>
                                </p>
                        </div>
                <div id="SE-7fb19585-75bb-4f73-b11f-0133720fe81a">
                                    <p id="SE-760f3ebc-e555-42e7-84f9-6c1ebf717ab2"><span id="SE-519b7182-251d-41f4-af91-005d115b2e2b">약 1,500개의 포스팅 정보를 embedding 했습니다.</span></p><p id="SE-bca0caa0-13ad-4fcd-93be-8f14e0148309"><span id="SE-bcea47c3-33c3-4206-87cb-e685b15b1c3f">​</span></p><p id="SE-625eba58-c6f1-42c9-ad8b-3ba9f9c336a8"><span id="SE-c2ac521e-ac37-4df2-be6c-7b44db0f0fea">​</span></p><p id="SE-003bed48-818a-4861-9036-7055e0929772"><span id="SE-4f8d0b4d-94c3-4a56-b7a9-6ed8e8b1f60a">마지막으로 RAG를 기반으로 LLM을 사용해보았습니다.</span></p>
                            </div>                <div id="SE-f94e943c-8fbf-4aa3-836b-ecf415c56906">
                                <p><a href="#" onclick="return false;" data-linktype="img" data-linkdata="{&quot;id&quot; : &quot;SE-f94e943c-8fbf-4aa3-836b-ecf415c56906&quot;, &quot;src&quot; : &quot;https://mblogthumb-phinf.pstatic.net/MjAyNDA2MjNfMjUw/MDAxNzE5MTIxNjg5MjQ1.jWMG7YWvOV4NCw0QvsABAlihOwelbzqnAz04KvcVBaEg.IXsjl6bqAHtxy_DmoPV4X7Kwg1pwcMDF5YCijcUdYYAg.PNG/%EC%8A%A4%ED%81%AC%EB%A6%B0%EC%83%B7_2024-06-23_14.45.34.png&quot;, &quot;originalWidth&quot; : &quot;1330&quot;, &quot;originalHeight&quot; : &quot;468&quot;, &quot;linkUse&quot; : &quot;false&quot;, &quot;link&quot; : &quot;&quot;}">
                                        <img src="https://mblogthumb-phinf.pstatic.net/MjAyNDA2MjNfMjUw/MDAxNzE5MTIxNjg5MjQ1.jWMG7YWvOV4NCw0QvsABAlihOwelbzqnAz04KvcVBaEg.IXsjl6bqAHtxy_DmoPV4X7Kwg1pwcMDF5YCijcUdYYAg.PNG/%EC%8A%A4%ED%81%AC%EB%A6%B0%EC%83%B7_2024-06-23_14.45.34.png?type=w800" data-width="693" data-height="243" alt="">
                                    </a>
                                </p>
                        </div>
                <div id="SE-1a84c5f9-31bd-4094-bdd7-6b2256fc6e91">
                                    <p id="SE-5e5f8f0a-d3f0-45f9-8714-06aef718cc3e"><span id="SE-f9356513-af10-4715-9947-0b4ec5f0adc6">​</span></p><p id="SE-c54004d7-5cd0-4ab2-90f4-42f857ad419f"><span id="SE-84bb4a7c-b8c5-46c7-86c2-cc23e055945e"><a href="https://github.com/pjt3591oo/blog-rag/blob/main/README.md" target="_blank">https://github.com/pjt3591oo/blog-rag/blob/main/README.md</a></span></p>
                            </div>                <div id="SE-cbfb7637-c6be-48dc-8e09-d7352e69fe4c">
                                <p><a href="https://github.com/pjt3591oo/blog-rag/blob/main/README.md" target="_blank">
                                    <img src="https://dthumb-phinf.pstatic.net/?src=%22https%3A%2F%2Fopengraph.githubassets.com%2F5a4e6103b22a1148a721c04c29f49e3fb27abcf5cc0ec1958aab2ebf8eaeba5c%2Fpjt3591oo%2Fblog-rag%22&amp;type=ff500_300" alt="">
                                </a></p><a href="https://github.com/pjt3591oo/blog-rag/blob/main/README.md" target="_blank">
                                    <div>
                                        <p><strong>blog-rag/README.md at main · pjt3591oo/blog-rag</strong></p><p>Contribute to pjt3591oo/blog-rag development by creating an account on GitHub.</p>
                                        <p>github.com</p>
                                    </div>
                                </a>
                            </div>                <div id="SE-d19f91bb-ced1-4f09-b5fc-dbc59c5cbf6c">
                                    <p id="SE-72450fdc-cdfa-4588-bb71-69ff80505ba5"><span id="SE-7cfdd18c-dd1f-48d8-89dc-c163f0508112">​</span></p><p id="SE-2414828f-3d73-45ab-8f8b-f9f5cd18de56"><span id="SE-9e0e712f-c8ca-4055-aebe-5f67495ac09d">간단하게 RAG를 다뤄보았으니 조만간 이를 주제로 포스팅을 할 예정~</span></p>
                            </div>    </div></div>

삽질하는 멍개씨

가장먼저, 블로그 포스트의 고유값인 log_no를 수집합니다. 두 번째로 각 페이지를 임베딩하여 결과를 chroma에 저장합니다. 포스트에 필요없는 정보들을 제거하지 않고 페이지의 모든내용을 임베딩했습니다. 약 1,500개의 포스팅 정보를 embedding 했습니다. 마지막으로 RAG를 기반으로 LLM을 사용해보았습니다. https://github.com/pjt3591oo/blog-rag/blob/main/README.md 간단하게 RAG를 다뤄보았으니 조만간 이를 주제로 포스팅을 할 예정~

[LLM] 내 블로그를 기반으로 RAG를 테스트해보았다.

댓글