Argilla (@argilla

Pinned

Jun 13, 2024

Today is a big day for Argilla and the Open Source AI community: We’re joining @huggingface🤗 ! Time to double down on community and data-centric, open source AI Read all about it and ask questions:

@dvilasuero on Hugging Face: "Today is a huge day in Argilla’s history. We couldn’t be more excited...

🚀 Open-source AI strikes again! Announcing Notux 8x7B, a fine-tune of Mixtral Instruct with high-quality chat data and DPO. Notux now the top ranked MoE on the Open LLM leaderboard. huggingface.co/argilla/notux-…

Argilla

@argilla_io

Feb 26, 2024

🚀🧙🏼‍♂️Introducing OpenHermesPreferences: the largest open dataset for RLHF & DPO Built together with the @huggingface H4 team, it's a 1M preferences dataset on top of the amazing @teknium's dataset. huggingface.co/datasets/argil… Let's dive in! 🧵

Argilla

@argilla_io

Nov 21, 2021

Build a news classifier from scratch with weak supervision 1. Programatically label 38.000 examples with rules and Snorkel. 2. Train a downstream classifier with scikit-learn to achieve 0.81 macro avg. f1-score. Tutorial link below 👇 #nlproc #ml #datascience #opensource

Argilla

@argilla_io

Dec 1, 2023

🔥Open-source, open-science, and data curation for the win! Meet Notus 7B, a new LLM tuned with DPO on a new curated UltraFeedback dataset, surpassing Zephyr and Claude 2 on AlpacaEval. Built on the shoulders of giants: 🙌@huggingface Alignment Handbook

Meet Notus-7B: Data Curation and Open Science go a long way in shaping AI's future

🔥 More is less for DPO, high quality matters! 📢 Dropping our first open dataset and LLM of the year: 💾Meet distilabel Orca Pairs DPO, an improved version of the now famous dataset from @intel 🏛️And a new OpenHermes model outperforming baselines with 54% less DPO pairs 🧵

Argilla

@argilla_io

Feb 10, 2024

🔥 Introducing a new open dataset for the Open Source AI community: OpenHermes2.5-dpo-binarized-alpha built atop the amazing dataset by @teknium huggingface.co/datasets/argil… This time we use OSS models for everything, even for the preference step! 🧵

Argilla

@argilla_io

Nov 28, 2021

Training a text classifier without labelled data using @PyTorch End-to-end weak supervision with Weasel & @huggingface transformers Guide and more details below 👇 #nlproc #datascience #python #opensource

Argilla

@argilla_io

Feb 7, 2023

🚀Data labeling from the @huggingface Hub is here. No more excuses to build great NLP datasets! argilla.io/blog/launching…

Argilla

@argilla_io

Feb 27, 2024

🤖Yesterday, we shared a large AI feedback dataset 👩‍💻Today, @argilla_io & @huggingface are thrilled to release a high-quality human feedback dataset, built with and for the community! 10K Prompts Ranked: +14K human ratings from +300 contributors! huggingface.co/datasets/DIBT/…

Argilla

@argilla_io

Jan 31, 2024

🚀 The OSS AI community needs more open datasets for improving LLMs: 🎁 Excited to ship a new open DPO dataset for boosting chat models: ⚗️ distilabel capybara-dpo, a multi-turn preference dataset built atop the awesome dataset by @ldjconfirmed huggingface.co/datasets/argil… 🧵

Argilla

@argilla_io

Jan 7, 2024

⚗️ distilabel, our OSS framework for building LLM finetuning & alignment datasets just crossed 300 stars Let's do a quick overview of its current features

GitHub - argilla-io/distilabel: Distilabel is a framework for synthetic data and AI feedback for...

We're building a high-quality prompt dataset together with the community. The result will be published with an open, commercial-friendly license, anyone can use to build eval, sft and DPO datasets. We need your help! This is how simple is to contribute: huggingface.co/spaces/DIBT/pr…

00:00

Argilla

@argilla_io

Jan 4, 2024

🔦 In this paper released by Apple, they introduced an efficient LLM inference model for devices with limited memory, showing inference speeds of 4-5x in CPU, and 20-25x in GPU. Argilla's GitHub: buff.ly/3tmZOmm distilabel: buff.ly/41DzWzk #nlproc #llms