🔥WILD! @AnthropicAI released Citations API for Claude last week -- a game-changing feature that will allow you to build trustworthy AI systems. Here's why this is HUGE... 🧵

1/ Build reliable RAG systems. AI should prove its work, not just give "trust me bro" responses. Anthropic's Citations API solves this elegantly: feed docs to Claude, enable citations, and get perfect source references! 🎯 @AnthropicAI 2/ What makes Citations special:

- Zero prompt engineering needed
- Automatic text chunking
- Works with PDFs & plain text
- Precise page/character citations
- Super efficient token usage
- Built for streaming
- No more complex citation systems! 🚀

🆕 SwiftEdit: Lightning Fast Text-guided Image Editing via One-step Diffusion

Make your edit in just 0.23 seconds !!! 🤯

User-friendly, lightning-fast editing gradio-app that enables instant edits through simple prompts, delivering precise results.👇 Learn more about the project SwiftEdit
> You can edit facial Identity and expressions via flexible text prompts

Welcome, Gradio 5 👋

We’ve been hard at work over the past few months, and we are excited to announce today the stable release of Gradio 5!

With more than 2 million users every month (and >470,000 apps on Hugging Face Spaces), Gradio has become the default way to build, share, and use machine learning applications.

Our goal with Gradio 5 was to address the most common pain points that we’ve heard from Gradio developers about taking these apps to production. Here are 5 new things in Gradio 5 (including a new way to build Gradio apps without writing any code!) Does Gradio load too slowly for you?

Well, not anymore! Gradio 5 ships with major performance improvements, including the ability to serve apps via server-side rendering (SSR) which loads Gradio apps almost instantaneously in the browser. No more loading spinner! 🏎️💨

🔥 🔥 Generate 3D characters with CharacterGen (SIGGRAPH'24)

- Generated 3D characters have high-quality shapes and textures
- Useful for downstream applications such as animation and game dev
- Code, demo, and pretrained weights available now on 🤗 🚀 Generate 3D characters with CharacterGen (SIGGRAPH'24)

👉 Official Demo on @huggingface Spaces: huggingface.co/spaces/VAST-AI…

🚀Introducing LLaVA-NeXT Interleave: Now AI can understand and reason with multiple images at once

- This opens up multi-image scenarios like multi-frame videos, multi-view 3D, and multiple inter-leaved images.
- An all round LMM that can understand videos, images, and 3D
More⬇️ LLaVA-NeXT-Interleave🔥
- Interleave data format unifies different tasks.
- New datasets on 🤗Hub:
1️⃣M4-Instruct, high-quality dataset, 1.1M samples from domains: multi-image, video, 3D & single-image
2️⃣LLaVA-Interleave Bench - Set of tasks to evaluate multi-image capabilities

🔥Introducing VILA: The next-gen VLM for Video and inter-leaved image understanding with Apache-2.0 license!

Is VILA the best in-class small VLM for edge-deployment? Keep reading🧵 👀 VILA uses interleaved image-text pretraining and it unlocks cool capabilities:

- Video & multi-image understanding 🎬
- In-context learning 🧠, and
- Visual chain-of-thought 💡

Experience the future -- Edge deployment 📱of VLMs! Try the demo now! 👇

🤯 MimicBrush might be the coolest AI app on 🤗 Spaces right now!

In-the-wild results are really something!🚀 🎨MimicBrush: New way to edit images with reference imitation! 🖼️🔍

- Zero-shot editing🎯
- Use in-the-wild references for inspiration🤩
- Learns semantic correspondence through self-supervised learning 🧠
- Outperforms existing alternatives in various test cases 📈
- Demo👇

🔥 SD3 has finally arrived on Huggingface!

Learn everything you'd need to get started👇 🌟SD3-Medium: A very powerful t2i model

- Multimodal Diffusion Transformer (MMDiT) architecture🧠
- Trained on 1B images, 30M aesthetic & 3M preference data🤯
- Improved image quality, typography & prompt understanding💡
- Non-commercial allowed, commercial license required📜

🎉Exciting news for the open-source world of AI! Google has released 𝐏𝐚𝐥𝐢𝐆𝐞𝐦𝐦𝐚, a family of vision-language models (pre-trained & fine-tuned)

🌟PaliGemma can take an image and text as input and generate bounding boxes, segmentation masks, do visual Q&A, and much more📸 💪 PaliGemma has 3 flavors: Pretrained (pt), Mix, and Fine-tuned (ft) models, each with different resolutions and precisions.
🧩 All models are available on the @huggingface Hub with nice model cards and with 🔥 𝐓𝐫𝐚𝐧𝐬𝐟𝐨𝐫𝐦𝐞𝐫𝐬 𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐢𝐨𝐧 🤗

🤩 Gradio project has been trending on Github, 🔥 alongside TEN GenAI projects that are leveraging Gradio for their development!!

Projects are from across the board, more in 🧵👇:
✅Voice generation
✅Video editing
✅Image generation
✅Cool LLMs
✅Multimodal
✅Audio generation ✅ Applio - VITS-based Voice Conversion focused on simplicity, quality & performance
🔗

✅ FunClip - Open source, accurate and convenient video slicing tool
🔗 github.com/IAHispano/Appl…
github.com/alibaba-damo-a…

Congratulations to AUTOMATIC1111: A Stable Diffusion Web-UI built using @Gradio for 100k @github stars! 🎉🥳

🖼️Automatic1111 is the Ultimate Creative Playground!🚀

Let's explore in this thread what all is possible with this feature-packed Web UI 👇🧵 1️⃣ One-click install & run script for easy setup📥
2️⃣ Outpainting, inpainting, color sketching – unleash your artistic vision🎨
3️⃣ Control attention with ((image-element)) format for focus👁️
4️⃣ Loopback & X/Y/Z plots for dynamic image transformations📊
5️⃣ Generate variations🔄

BIG NEWS 🥳🎈

Building Chatbots apps just got wayyy easier: announcing the new 𝙲𝚑𝚊𝚝𝙸𝚗𝚝𝚎𝚛𝚏𝚊𝚌𝚎 class 🙌

The *fastest* way to build to build a Chatbot UI in Python -- including streaming, undo/retry, API, all out of the box!

Let's take a look at a few examples... The 𝙲𝚑𝚊𝚝𝙸𝚗𝚝𝚎𝚛𝚏𝚊𝚌𝚎 only has 1 required parameter: a function that takes in:

- an input message (string)
- a history (list of list of strings)

and returns:
- a response (string)

Here's a toy example: (for the example above!)

💫 We're excited to launch a NEW Python library 💫

The 𝚐𝚛𝚊𝚍𝚒𝚘_𝚌𝚕𝚒𝚎𝚗𝚝 library lets you run any Gradio app as an API 🚀

See a cool Hugging Face Space? Use it programmatically instantly: You can connect to private Spaces by passing in your Hugging Face token

Gradio 3.7 is out! Everything that is new in 3.7 🔽

1. Gradio now supports *batched* function. You can specify a batch size and Gradio will automatically batch incoming requests so that your demo runs on a lot faster on Spaces! 2. A new upload() event for Video, Audio, Image, and File components:

Excited to release 3 big features as part of 𝚐𝚛𝚊𝚍𝚒𝚘 𝟹.𝟹 🧡

1. Iterative outputs: you can now `yield` a sequence of outputs and they will show up sequentially in your Gradio demo: 2. The 𝚐𝚛𝚊𝚍𝚒𝚘.𝙰𝚌𝚌𝚘𝚛𝚍𝚒𝚘𝚗 class, which allows you to toggle on/off the visibility of components (think advanced options) in your demo very easily:

Now it's time for our favorite part of the Blocks Party: awarding the top machine learning demos 🏆

We had more than 50 incredibly creative and complex demos, built entirely in Python💫

Here are the top 5, as voted by your likes... 🧡 In 5th place, with 22 likes, is: StyleGAN-NADA 🥳

Built by @RinonGalt, he demo allows you to edit specific attributes in your face, and generate a new picture with specific styles turned on or off. It's scary good!

Link: huggingface.co/spaces/Gradio-…