Introduction The deployment of machine learning models is the process for making your models available in production environments, where they can provide predictions to other software systems. It is only once models are deployed to production that they start adding value, making deployment a crucial step. However, there is complexity in the deployment of machine learning models. This post aims to
Model Spec (2024/05/08)
May 08, 2024 Overview This is the first draft of the Model Spec, a document that specifies desired behavior for our models in the OpenAI API and ChatGPT. It includes a set of core objectives, as well as guidance on how to deal with conflicting objectives or instructions. Our intention is to use the Model Spec as guidelines for researchers and data labelers to create data as part of a technique cal
TechCrunch
Tesla has officially revealed a new Performance variant of the recently refreshed Model 3 sedan as the company looks to fight off receding demand. The new version of the Model 3, which starts at $52,9
CentOS Stream: Building an innovative future for enterprise Linux
On June 30, 2024, CentOS Linux 7 will reach End of Life (EOL). Explore Red Hat’s options to help ease your migration, including Red Hat Enterprise Linux 7 for Third Party Linux Migration. Find out more In September 2019, we announced CentOS Stream, an upstream development platform designed for CentOS community members, Red Hat partners, ecosystem developers, and many other groups to more quickly a
Fast and Portable Llama2 Inference on the Heterogeneous Edge
Fast and Portable Llama2 Inference on the Heterogeneous Edge • 12 minutes to read The Rust+Wasm stack provides a strong alternative to Python in AI inference. Compared with Python, Rust+Wasm apps could be 1/100 of the size, 100x the speed, and most importantly securely run everywhere at full hardware acceleration without any change to the binary code. Rust is the language of AGI. We created a very
GitHub - yandex/YaLM-100B: Pretrained language model with 100B parameters
YaLM 100B is a GPT-like neural network for generating and processing text. It can be used freely by developers and researchers from all over the world. The model leverages 100 billion parameters. It took 65 days to train the model on a cluster of 800 A100 graphics cards and 1.7 TB of online texts, books, and countless other sources in both English and Russian. Training details and best practices o
openai-cookbook/techniques_to_improve_reliability.md at main · openai/openai-cookbook
Techniques to improve reliability When GPT-3 fails on a task, what should you do? Search for a better prompt that elicits more reliable answers? Invest in thousands of examples to fine-tune a custom model? Assume the model is incapable of the task, and move on? There is no simple answer - it depends. However, if your task involves logical reasoning or complexity, consider trying the techniques in
如何にして第二次大戦後に経済学が間違った方向に転換したか、および、経済学の新たな理論的枠組み - himaginary’s diary
と題したエントリ（原題は「How economics took a wrong turn post World War II and an alternative theoretical framework for economics」）でMostly Economicsが、Meir Kohnダートマス大教授がケイトー研究所のCato Journal秋号に掲載した論文「An Alternative Theoretical Framework for Economics」を紹介している。以下はその論文の冒頭。 As a profession, economics is thriving. The number of economists is large and growing. The volume of their output is exploding—more articles ar
GitHub - google-deepmind/alphafold: Open source code for AlphaFold.
This package provides an implementation of the inference pipeline of AlphaFold v2. For simplicity, we refer to this model as AlphaFold throughout the rest of this document. We also provide: An implementation of AlphaFold-Multimer. This represents a work in progress and AlphaFold-Multimer isn't expected to be as stable as our monomer AlphaFold system. Read the guide for how to upgrade and update co
Author: Neo Chiu Background We use Prisma + Postgres from prototype and start to migrate all data with RLS ( Row Level Security ) last year. We are managing multi-tenants data in one database, and we don't want data be accessed cross tenants. RLS restricts data with security policies at database engine level to prevent any unexpected access from client side. You can find more details in Japanese a
Stability AI launches SDXL 0.9: A Leap Forward in AI Image Generation — Stability AI
Stability AI launches SDXL 0.9: A Leap Forward in AI Image Generation Today, Stability AI announces SDXL 0.9, the most advanced development in the Stable Diffusion text-to-image suite of models. Following the successful release of Stable Diffusion XL beta in April, SDXL 0.9 produces massively improved image and composition detail over its predecessor. The model can be accessed via ClipDrop today,
ICLR 2022 — A Selection of 10 Papers You Shouldn’t Miss
Image by Zeta Alpha.The International Conference in Learning Representations (ICLR) will be held online (for the third year in a row!) from Monday, April 25th through Friday, April 29th. It’s one of the biggest and most beloved conferences in the world of Machine Learning Research, and this year is no exception: it comes packed with more than a thousand papers on topics ranging from ML theory, Rei
How to OCR with Tesseract in Python with Pytesseract and OpenCV?
In this blog post, we will try to explain the technology behind the widely used Tesseract Engine, which was upgraded with the latest knowledge researched in optical character recognition. This article will also serve as a how-to guide/ tutorial on how to implement PDF OCR in python using the Tesseract engine. We will be walking through the following modules: Tesseract OCR FeaturesPreprocessing for
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
Tracking covid-19 excess deaths across countriesIn many parts of the world, official death tolls undercount the total number of fatalities AS COVID-19 has spread around the world, people have become grimly familiar with the death tolls that their governments publish each day. Unfortunately, the total number of fatalities caused by the pandemic may be even higher, for several reasons. First, the of
統計的観点から見た実証マクロ経済学とDSGEモデル作成 - himaginary’s diary
というプレプリントがarXivに上がっている（H/T beさんツイート；著者の一人のShaliziが「君の好きなDSGEはダメダメ（Your Favorite DSGE Sucks）」と題した自ブログエントリ*1で内容を解説し、ツイートに流している）。原題は「Empirical Macroeconomics and DSGE Modeling in Statistical Perspective」で、著者はDaniel J. McDonald（ブリティッシュコロンビア大学バンクーバー校）、Cosma Rohilla Shalizi（カーネギーメロン大学）。 以下はその結論部（Discussion）の前半。 As we said in the introduction, there are very few who will defend the forecasting record of
There are many different ways to build with LLMs, including training models from scratch, fine-tuning open-source models, or using hosted APIs. The stack we’re showing here is based on in-context learning, which is the design pattern we’ve seen the majority of developers start with (and is only possible now with foundation models). The next section gives a brief explanation of this pattern; experi
GitHub - bentoml/OpenLLM: Operating LLMs in production
OpenLLM is an open-source platform designed to facilitate the deployment and operation of large language models (LLMs) in real-world applications. With OpenLLM, you can run inference on any open-source LLM, deploy them on the cloud or on-premises, and build powerful AI applications. Key features include: 🚂 State-of-the-art LLMs: Integrated support for a wide range of open-source LLMs and model ru
This post introduces a new feature called Profiles, which allows you to create a specific Kubernetes application platform to meet your business needs. We show how you can enable machine learning operations or MLOps with specific Profiles for two different types of Kubernetes instances - EKS and Kubernetes with Firekube. If you ask an application developer what they want from Kubernetes, the answer
The Illustrated GPT-2 (Visualizing Transformer Language Models)
Jay Alammar Visualizing machine learning one concept at a time. @JayAlammar on Twitter. YouTube Channel This year, we saw a dazzling application of machine learning. The OpenAI GPT-2 exhibited impressive ability of writing coherent and passionate essays that exceed what we anticipated current language models are able to produce. The GPT-2 wasn’t a particularly novel architecture – it’s architectur
「BERT/GPT-3/DALL-E 自然言語処理・画像処理・音声処理 人口知能プログラミング実践入門」を読んで、リクルートのAI研究機関「Megagon Labs」提供の「GINZA」という日本語の自然言語処理ライブラリがあることを知りました。 ※書籍へのリンクも記載していますが、このnoteは書籍の内容に従わずにあくまでも勝手に最新バージョンで試したことに対する内容です 興味を惹かれBERTくらいしか自然言語処理ライブラリの名前を知らなかったため興味を惹かれたのですが、書籍内のGINZAのバージョンは4.0.5であり少し古いバージョンでした。2021/08/26にv5がリリースされているようで、2021/10/01時点では最新は5.0.2 (2021/09/06)となっていました。 試そうとするもせっかく試すならば最新で試したいと思ったところ、v4からv5になった際にbraking c
Updates: Dec 25, 2021: Added Auto Encoder and variational Encoder Dec 25, 2020: Added Ensemble Methods Download the updated version of the cheat sheets from http://cheatsheets.aqeel-anwar.com/ A couple of years ago I started applying for internships in the area of Machine Learning and ML system design. I had been studying and actively researching in the area of ML for a few years then. I was famil
Zig And Rust
Zig And Rust Mar 26, 2023 This post will be a bit all over the place. Several months ago, I wrote Hard Mode Rust, exploring an allocation-conscious style of programming. In the ensuing discussion, @jamii name-dropped TigerBeetle, a reliable, distributed, fast, and small database written in Zig in a similar style, and, well, I now find myself writing Zig full-time, after more than seven years of Ru
Announcing Optuna 2.0 - Preferred Networks Research & Development
We are pleased to announce the second major version of Optuna, a hyperparameter optimization (HPO) framework in Python, is now available on PyPI and conda-forge. See the release notes on GitHub for the list of changes. Starting from January this year when the first major version was released, we have seen tremendous effort from the community in terms of pull requests, issues, use cases beyond the
AI Agents are disrupting automation: Current approaches, market solutions and recommendations
The mainstreaming of AI tools has ignited hope for dramatic productivity improvements for knowledge workers and consumers alike. Transformer-based Large Language Models (LLMs) have demonstrated AI capabilities that are transforming workflows with new automation approaches. In the article below, we trace the automation journey in the age of AI and dig into some of the current and evolving platforms
{CausalImpact}を使う上での注意点を簡単にまとめてみた - 渋谷駅前で働くデータサイエンティストのブログ
実はこのネタは元々別のところでやり取りのあった話題だったりします。 色々な都合があってここ最近{CausalImpact}に触れる機会が自分に限らず周囲でも増えているのですが、若い人たちから「そもそも{CausalImpact}って何をしているんですか？使う際は何に気を付けたら良いですか？」などと聞かれることがちょくちょくあるので、備忘録も兼ねてまとめてみることにしました。いつもながらですが、内容に不備や誤解や理解不足がありましたらどしどしご指摘くださいm(_ _)m なお、{CausalImpact}パッケージそのものの簡潔な説明は随分昔に書きました。単純に使いたいだけならこちらの記事をお読みいただければ十分かと思います。 {bsts}をバックエンドとして動く、統計的因果推論を目的としたcounterfactual modelingフレームワーク どのようにcounterfactual
初めに、論文発表元のGithubからソースコードを取得します %cd /content !git clone https://github.com/CompVis/latent-diffusion.git 次にライブラリをインストールします。 %cd /content !git clone https://github.com/CompVis/taming-transformers !pip install -e ./taming-transformers !pip install omegaconf>=2.0.0 pytorch-lightning>=1.0.8 torch-fidelity einops import sys sys.path.append(".") sys.path.append('./taming-transformers') from taming.models
![[Latent Diffusion] AIでテキストから画像を生成する](https://cdn-ak-scissors.b.st-hatena.com/image/square/6706790eb9410b1ef92e43f8816afc6e44db04b1/height=288;version=1;width=512/https%3A%2F%2Fblogger.googleusercontent.com%2Fimg%2Fb%2FR29vZ2xl%2FAVvXsEirH4niWQQXpBqo7BhSe5lUHIdTCKpaUe2kqFwsL6FSbvn07DkGNxn2zM0yh5MZ85kuwMeR2YlsjIxxRy5zrb-hHoB0LrQEc9bDVC3S43_xXxgv4RnLrEYsMVQxgWkv9K1YAT-aKdoAjDmjFrQo0WZINi63fUdoqzgov_wVqCLOB54vVp_67PzETkUeHg%2Fw1200-h630-p-k-no-nu%2F%25E3%2582%25A2%25E3%2582%25A4%25E3%2582%25AD%25E3%2583%25A3%25E3%2583%2583%25E3%2583%2581.png)
皆さんこんにちは。株式会社cocoの高橋です。最近、弊社サービスの新規受注金額（アップセル含む）が1年で4.5倍にまで増えました。しかもこの間セールスの人件費は殆ど同じで、広告費も殆ど変えていません。単純な計算で1年で生産性が4.5倍になりました。 そこで今回、その成長の過程での弊社の工夫と改善の結果生まれたセールススタイル＝「スクラムセールス」を皆様にご紹介します。 ※本記事は、2022年8月に公開され、2023年4月に内容を追記しています 四半期ごとの受注金額（新規契約 + アップセル）の推移全員・全顧客・全プロセス対応のスクラムセールス全員で全顧客全プロセスに対応できる体制を作っています「全員で全部やる」という感じがスクラムっぽかったので「スクラムセールス」という名前をつけました。各商談先や顧客の「担当」を明記せず、また各プロセス担当も置かず、「全員で、全顧客のあらゆる課題の解決に向
Design for the iPadOS pointer - WWDC20 - Videos - Apple Developer
Streaming is available in most browsers, and in the WWDC app. Bring the power of the pointer to your iPad app: We'll show you how Apple's design team approached designing the iPadOS pointer to complement touch input, and how you can customize and refine pointer interactions in your app to make workflows more efficient and gratifying. Discover how the pointer's adaptive precision enables people to
Building AI Trading Systems
Two years ago I wrote a post about applying Reinforcement Learning to financial markets. A few people asked me what became of it. This post covers some high-level things I’ve learned. It’s more of a rant than an organized post. Over the past few years I’ve built four and a half trading systems. The first one failed to be profitable. The second one I never finished because I realized early on that
Transformer Debugger (TDB) is a tool developed by OpenAI's Superalignment team with the goal of supporting investigations into specific behaviors of small language models. The tool combines automated interpretability techniques with sparse autoencoders. TDB enables rapid exploration before needing to write code, with the ability to intervene in the forward pass and see how it affects a particular
本記事は日本オラクルが運営する下記Meetupで発表予定の内容になります。発表までに今後、内容は予告なく変更される可能性があることをあらかじめご了承ください。当日は記事内容以外にデモンストレーションも実施する予定です。 以下の記事内容とセットで実施する予定です。以下の記事がメインでこちらの記事がサブというアジェンダとなります。 はじめに 2022年暮れ、ChatGPTの登場以降、あらゆる企業がDXの在り方を問われはじめ、大規模言語モデルの仕組みをどのように業務に取り入れるかを検討されていると思います。 その検討の一つとして、「GPT(LLM)が学習していない企業内のデータや最新のデータも有効活用すべき」 という点は非常に大きな論点なのではないでしょうか。 ご存じの通り、LLMとはインターネット上に存在するドキュメントデータをクローリングにより大量に収集し、それを学習データとして機械学習にか
Image edited by Author for the icon taken from the official site of PyTorch and KerasThe war between Deep Learning Frameworks is still on fire, which one that will gain more masses, it will be the next game-changer for the deep learning community in future. The loser one will fade in if they can’t survive by giving the best solution for the deep learning community and the world. The first framewor
GitHub - openai/gpt-3: GPT-3: Language Models are Few-Shot Learners
arXiv link Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from
CLIP: Connecting text and images
We’re introducing a neural network called CLIP which efficiently learns visual concepts from natural language supervision. CLIP can be applied to any visual classification benchmark by simply providing the names of the visual categories to be recognized, similar to the “zero-shot” capabilities of GPT-2 and GPT-3. Although deep learning has revolutionized computer vision, current approaches have se
What I Worked On
February 2021 Before college the two main things I worked on, outside of school, were writing and programming. I didn't write essays. I wrote what beginning writers were supposed to write then, and probably still are: short stories. My stories were awful. They had hardly any plot, just characters with strong feelings, which I imagined made them deep. The first programs I tried writing were on the
営業の仕組みづくりにまず成功したのは米国だった。東京と異なり、米国は広大だ。かつては、米国でもリード（見込み客）の獲得から受注までを営業が担当していたが、営業が製品を持ってまだ製品の存在さえ知らない顧客を求めて国中を飛び回るのはあまりに効率が悪い。 まず、リードの獲得という役割をマーケティング部が担うようになった。すると、マーケティングが持ってきたリードがゆるい、商談に結びつかないと営業が文句を言うようになる。 だから、マーケと営業の間にインサイドセールスという役割の部署がおかれた。リードを商談につなげるように製品やサービスの効用を知らせ、顧客の興味を喚起する役割だ。リード作りをマーケティングが、商談作りをインサイドセールスが、提案や契約を営業が担うようにした。 「分業」のメリット、デメリット「THE MODEL」の本質は、生産性向上を目的とした営業プロセスの分業だと筆者は考えています。一
入力したテキストに則した画像を生成したり、ある画像から別の画像を生成したりすることができる画像生成AIの「Stable Diffusion」を使い、さまざまなインテリアデザインをAIに生成してもらうという試みをKaren X. ChengさんとJustin Alveyさんが行っています。 Using AI for design inspiration We used Stable Diffusion Depth to Image to get the consistency - collab with @justLV See below for our process#ArtificialIntelligence #stablediffusion #interiordesign pic.twitter.com/teImanZsZF— Karen X. Cheng (@karenxcheng
IBM, Red Hat and Free Software: An old maddog’s view Copyright 2023 by Jon “maddog” Hall Licensed under Creative Commons BY-SA-ND Photo: © Santiago Ferreira Litowtschenko Several people have opined on the recent announcement of Red Hat to change their terms of sales for their software. Here are some thoughts from someone who has been around a long time and been in the midst of a lot of what occur
Making LLMs even more accessible with bitsandbytes, 4-bit quantization and QLoRA
LLMs are known to be large, and running or training them in consumer hardware is a huge challenge for users and accessibility. Our LLM.int8 blogpost showed how the techniques in the LLM.int8 paper were integrated in transformers using the bitsandbytes library. As we strive to make models even more accessible to anyone, we decided to collaborate with bitsandbytes again to allow users to run models
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
How to Deploy Machine Learning Models
Custom Prisma Client for RLS - Beatrust techBlog
What are Diffusion Models?
Tracking covid-19 excess deaths across countries
Emerging Architectures for LLM Applications | Andreessen Horowitz
MLOps on Kubernetes with portable Profiles
NLP | GINZA v5で固有表現抽出のルール追加を試してみた｜Koji Iino
Cheat Sheets for Machine Learning Interview Topics
[Latent Diffusion] AIでテキストから画像を生成する
新規受注額を1年で4.5倍にした「スクラムセールス」｜高橋俊介@株式会社cocoの代表取締役
GitHub - openai/transformer-debugger
【ChatGPT】ファインチューニングをわかりやすく解説 - Qiita
Why I Switch From Keras to PyTorch
THE MODELなマーケターが見るKPIはリード獲得単価、商談獲得単価、受注単価｜松本健太郎
画像生成AI「Stable Diffusion」を使って新しいインテリアデザインを作成しまくる試み
IBM, Red Hat and Free Software: An old maddog’s view