Riffusion
🔥🔥🔥😱 Really? Yup. This is the v1.5 stable diffusion model with no modifications, just fine-tuned on images of spectrograms paired with text. Audio processing happens downstream of the model. It can generate infinite variations of a prompt by varying the seed. All the same web UIs and techniques like img2img, inpainting, negative prompts, and interpolation work out of the box. Code: https://git
OSSベクトルDBã®Chromaを使ã£ã¦Q&AボットをLangChainã§ä½œæˆã™ã‚‹ï½œmah_lab / 西見 å…¬å®
新興ã§å‹¢ã„ã®ã‚るベクトルDBã«Chromaã¨ã„ã†OSSãŒã‚ã‚Šã€ã‚ªãƒ³ãƒ¡ãƒ¢ãƒªã®ãƒ™ã‚¯ãƒˆãƒ«DBã¨ã—ã¦æ°—軽ã«è©¦ã›ã¾ã™ã€‚ LangChainã‚„LlamaIndexã¨ã®ã‚¤ãƒ³ãƒ†ã‚°ãƒ¬ãƒ¼ã‚·ãƒ§ãƒ³ãŒã‚¦ãƒªã®OSSã§ã™ãŒã€ä»Šå›žã¯å˜ç´”ã«ãƒ™ã‚¯ãƒˆãƒ«DBã¨ã—ã¦ä½¿ã†æ„Ÿã˜ã§è©¦ã—ã¦ã¿ã¾ã—ãŸã€‚ データをChromaã«ç™»éŒ²ã™ã‚‹ä»Šå›žã¯LangChainã®ãƒ‰ã‚ュメントをChromaã«ç™»éŒ²ã—ã€LangChainã®Q&AãŒã§ãるよã†ãªãƒœãƒƒãƒˆã‚’作æˆã—よã†ã¨æ€ã„ã¾ã™ã€‚ ã—ã‹ã—LangChainã®ãƒ‰ã‚ュメントã¯ã»ã¨ã‚“ã©ãŒJupyter Notebookå½¢å¼ãªã®ã§ã€ãƒ™ã‚¯ãƒˆãƒ«DBã¸å–ã‚Šè¾¼ã¿ã‚„ã™ã„よã†ã«ãƒ•ãƒ©ãƒƒãƒˆãƒ†ã‚ストã«ã—ã¦ã‚ã’ã‚‹å¿…è¦ãŒã‚ã‚Šã¾ã™ã€‚ 以下ã®é–¢æ•°ã¯Jupyter Notebookå½¢å¼ï¼ˆJSON）ã®ãƒ•ã‚¡ã‚¤ãƒ«ã‚’分解ã—ã¦Markdownå½¢å¼ã«å¤‰æ›ã—ã€ãã®å¾ŒUnstructured.ioã®Markdownスプリッタを利用ã—ã¦ã‚³ãƒ³ãƒ†ãƒ³ãƒ„ã‚’ãƒãƒ£ãƒ³ã‚¯ã«åˆ†å‰²
Building LLM applications for production
[Hacker News discussion, LinkedIn discussion, Twitter thread] A question that I’ve been asked a lot recently is how large language models (LLMs) will change machine learning workflows. After working with several companies who are working with LLM applications and personally going down a rabbit hole building my applications, I realized two things: It’s easy to make something cool with LLMs, but ver
ã€LangChain】外部ã®æ–‡æ›¸ã‚’解釈ã—ãŸQAbotを作りã€ãã®å—ã‘ç”ãˆãŒé©åˆ‡ã‹ã‚’確ã‹ã‚ã‚‹
LangChainã§æ–‡æ›¸ãƒ‡ãƒ¼ã‚¿ã‚’èªã¿è¾¼ã‚“ã Q&Abotã®è©•ä¾¡ã‚’行ㆠLangChainを使ã†ã“ã¨ã§ã€å¤–部ã®æ–‡æ›¸ã‚’解釈ã•ã›ã¦ãã®å†…容ã«å‰‡ã—ãŸQAbotを作るã¨ã„ã£ãŸã“ã¨ãŒç°¡å˜ã«å®Ÿç¾ã§ãã¾ã™ã€‚今回ã¯ã€ãã†ã—ã¦å‡ºæ¥ãŸQ&AbotãŒé©åˆ‡ã«æ–‡æ›¸ã‚’解釈ã—ã¦ã„ã‚‹ã®ã‹ã‚’確ã‹ã‚る方法を紹介ã—ã¾ã™ã€‚ 今回ã®æ–¹æ³•ã‚’応用ã™ã‚Œã°ã€ ①外部ã®æ–‡æ›¸ã‚’解釈ã•ã›ãŸAIを作る â‘¡ãã®çŸ¥è˜ã‚’利用ã—ãŸChatbotを作る â‘¢ãã®ChatbotãŒé©åˆ‡ã«å‹•ä½œã™ã‚‹ã‹æ¤œè¨¼ã™ã‚‹ ã¨ã„ã£ãŸã“ã¨ãªã©ãŒå¯èƒ½ã§ã™ã€‚ ソース：https://langchain.readthedocs.io/en/latest/use_cases/evaluation/data_augmented_question_answering.html 最終的ãªã‚¢ã‚¦ãƒˆãƒ—ット 最終的ã«ã¯ä»¥ä¸‹ã®ã‚ˆã†ãªã‚¢ã‚¦ãƒˆãƒ—ットを目指ã—ã¾ã™ã€‚ 文書ã®å†…容ã«å‰‡ã—ãŸè³ªå•ã«å¯¾ã—ã¦ã€ã€Œå®Ÿéš›ã®ç”ãˆã€ã¨ã„
ReAct: Synergizing Reasoning and Acting in Language Models (言語モデルã«ãŠã‘る推論ã¨è¡Œå‹•ã®ç›¸ä¹—効果)｜npaka
ReAct: Synergizing Reasoning and Acting in Language Models (言語モデルã«ãŠã‘る推論ã¨è¡Œå‹•ã®ç›¸ä¹—効果) 1. ã¯ã˜ã‚ã«è¿‘å¹´ã€æ€è€ƒã®é€£éŽ– (CoT: Chain of Thought) を促ã—ãŸè¨€èªžãƒ¢ãƒ‡ãƒ«ã¯ã€ã•ã¾ã–ã¾ãªç®—è¡“ã€å¸¸è˜æŽ¨è«–タスクã«å„ªã‚Œã€å‰µç™ºçš„ãªèƒ½åŠ›ã‚’発æ®ã—ã¾ã™ã€‚ãŸã ã—ã€ç‹¬è‡ªã®å†…部表ç¾ã‚’使用ã™ã‚‹ãŸã‚ã€ãƒªã‚¢ã‚¯ãƒ†ã‚£ãƒ–ã«æŽ¢ç´¢ãƒ»æŽ¨è«–ã—ãŸã‚Šã€çŸ¥è˜ã‚’æ›´æ–°ã—ãŸã‚Šã™ã‚‹èƒ½åŠ›ãŒåˆ¶é™ã•ã‚Œã¾ã™ã€‚ 一方ã€æœ€è¿‘ã®ç ”究ã§ã¯ã€äº‹å‰å¦ç¿’済ã¿è¨€èªžãƒ¢ãƒ‡ãƒ«ã§ã€Œãƒ†ã‚ストコンテã‚ストã€ã‚’「テã‚スト行動ã€ã«ãƒžãƒƒãƒ”ングã™ã‚‹ã“ã¨ã§ã€ã•ã¾ã–ã¾ãªã‚¤ãƒ³ã‚¿ãƒ©ã‚¯ãƒ†ã‚£ãƒ–環境 (例: テã‚ストゲームã€Webナビゲーションã€ãƒãƒœãƒ†ã‚£ã‚¯ã‚¹) ã§è¡Œå‹•ã‚’決定ã—ã¦ã„ã¾ã™ã€‚ãŸã ã—ã€ãれらã¯å¤šæ®µéšŽã®ç›®æ¨™ã«ã¤ã„ã¦æŠ½è±¡çš„ã«æŽ¨è«–ã—ãŸã‚Šã€ä½œæ¥è¨˜æ†¶ã‚’ç¶æŒã—ãŸã‚Šã—ã¾ã›ã‚“。 「ReAct: Synergizing R
What are Diffusion Models?
Date: July 11, 2021 | Estimated Reading Time: 32 min | Author: Lilian Weng [Updated on 2021-09-19: Highly recommend this blog post on score-based generative modeling by Yang Song (author of several key papers in the references)]. [Updated on 2022-08-27: Added classifier-free guidance, GLIDE, unCLIP and Imagen. [Updated on 2022-08-31: Added latent diffusion model. [Updated on 2024-04-13: Added prog
1
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
{{#tags}}- {{label}}
{{/tags}}