An open API service indexing awesome lists of open source software.

https://github.com/ziweek/tec-chat

πŸ† Chief of Staff of the Republic of Korea Army Award, in 11th Republic of Korea Army Entrepreneurship Competition - πŸ€–πŸͺ– Guide Chatbot to Field Training Utilizing Generative AI
https://github.com/ziweek/tec-chat

huggingface langchain langsmith mistral-7b nextjs retrieval-augmented-generation transformers

Last synced: about 1 year ago
JSON representation

πŸ† Chief of Staff of the Republic of Korea Army Award, in 11th Republic of Korea Army Entrepreneurship Competition - πŸ€–πŸͺ– Guide Chatbot to Field Training Utilizing Generative AI

Awesome Lists containing this project

README

          

# Hello, Tech_Chat


πŸ† Chief of Staff of the Army Award, in 11th ROKA Entrepreneurship Competition πŸ†




KOREAN
Β |Β 
ENGLISH




Guide Chatbot to Field Training Utilizing Generative AI




































Check out prototypes in the badge below
















# 1. Introduction

> [!NOTE]
>
> - This idea is a project that won the Chief of Staff of the Army Award (Excellence Award) at the [2024 11th Korea Army Startup Competition](https://www.army-startup.co.kr/) hosted by the Republic of Korea Army Headquarters.
> - This idea involves developing an intelligent chatbot service powered by Large Language Models (LLMs) to overcome the limitations of traditional booklet-style field manuals. The project aims to research and develop an intelligent electronic manual platform that enables quick and accurate access to essential information for operating military equipment.

https://github.com/ziweek/tec-chat/assets/99459331/1265acf4-f164-467d-b7c6-bc51e887ddad




On-premise LLM ensemble architecture





Realtime 3D rendering of equipments





Project Application


















# 2. Implementation

핡심기λŠ₯

#### 1. On-Premise ν™˜κ²½μ˜ LLM 앙상블 ꡬ쑰







이 κΈ°λŠ₯은 둜컬 ν™˜κ²½μ— μ—¬λŸ¬ μ–Έμ–΄ λͺ¨λΈμ„ κ΅¬μΆ•ν•˜κ³  μ—°λ™ν•˜μ—¬ λ³΄μ•ˆμ„±μ„ κ°•ν™”ν•˜λ©°, λŒ€κ·œλͺ¨ 데이터λ₯Ό μ‹€μ‹œκ°„μœΌλ‘œ μ²˜λ¦¬ν•˜λŠ” 데 ν•„μš”ν•œ μ„±λŠ₯을 μ œκ³΅ν•©λ‹ˆλ‹€. 이λ₯Ό 톡해 μ‹ μ†ν•œ 응닡 및 μ •ν™•ν•œ μ •λΉ„ 지원을 κ°€λŠ₯μΌ€ ν•©λ‹ˆλ‹€.




#### 2. 운용μž₯λΉ„ κ΅¬μ‘°λ„μ˜ μ‹€μ‹œκ°„ 3D λ Œλ”λ§







이 κΈ°λŠ₯은 μš΄μš©λ˜λŠ” μž₯λΉ„μ˜ ꡬ쑰도λ₯Ό μ‹€μ‹œκ°„μœΌλ‘œ 3D λ Œλ”λ§ν•˜μ—¬ μ‚¬μš©μžμ—κ²Œ μ œκ³΅ν•©λ‹ˆλ‹€. 이λ₯Ό 톡해 μ‚¬μš©μžλŠ” λ³΅μž‘ν•œ μž₯λΉ„ ꡬ쑰λ₯Ό λͺ…ν™•ν•˜κ²Œ 이해할 수 있으며, 효율적으둜 μ •λΉ„λ₯Ό 마칠 수 μžˆμŠ΅λ‹ˆλ‹€.




#### 3. λ©€ν‹°λͺ¨λ‹¬ μ§€μ›μœΌλ‘œ μ‚¬μš©μž νŽΈμ˜μ„± κ°œμ„ 







ν•΄λ‹Ή κΈ°λŠ₯은 ν…μŠ€νŠΈ, μŒμ„±, 이미지 λ“± λ‹€μ–‘ν•œ λ©€ν‹°λͺ¨λ‹¬ μžμ›μ„ λŒ€μƒμœΌλ‘œ μ§ˆμ˜μ–΄ μž…λ ₯을 μ§€μ›ν•˜μ—¬ μ‚¬μš©μžμ˜ νŽΈμ˜μ„±μ„ λ†’μž…λ‹ˆλ‹€. μ‚¬μš©μžλŠ” μžμ‹ μ—κ²Œ κ°€μž₯ νŽΈν•œ λ°©μ‹μœΌλ‘œ μƒν˜Έμž‘μš©ν•  수 있으며, 이λ₯Ό 톡해 μ‚¬μš©μž νŽΈμ˜μ„±μ„ κ·ΉλŒ€ν™”ν•  수 μžˆμŠ΅λ‹ˆλ‹€.






μ•„ν‚€ν…μ²˜

#### ν”„λ‘œλ•νŠΈ μ•„ν‚€ν…μ²˜









λ³Έ ν”„λ‘œμ νŠΈμ˜ μ•„ν‚€ν…μ²˜λŠ” 데이터 μ „μ²˜λ¦¬ μˆ˜ν–‰ μ„œλ²„(μ΄ˆκ³ ν•΄μƒλ„ 이미지 κ°œμ„  λ”₯λŸ¬λ‹ λͺ¨λΈ), 핡심 κΈ°λŠ₯ μˆ˜ν–‰ μ„œλ²„(μ†Œν˜• 객체 식별 λ”₯λŸ¬λ‹ λͺ¨λΈ), 그리고 μ›Ή μ–΄ν”Œλ¦¬μΌ€μ΄μ…˜(ν”„λ‘ νŠΈμ—”λ“œμ™€ λ²‘μ—”λ“œ 및 λ°μ΄ν„°λ² μ΄μŠ€)으둜 κ΅¬μ„±λ˜μ–΄ μžˆμŠ΅λ‹ˆλ‹€.






# 3. νŒ€μ›