yianchen

Funny :3

[そう｜Claude Codeで始めるAI自動化 @so_ainsight]

無料でここまでやるのか、というツール。GitHubで早くもスター1万超え。

「YouTube動画まるごと多言語ふきかえマシン」みたいなものが公開されています。名前は Voice-Pro。

・海外の動画から人の声だけを抜き出す
・しゃべりを文字起こしして100以上の言語へ翻訳
・元の話者の声を真似て、別の言語でしゃべり直させる

など、動画の音声まわりを丸ごとAIに任せられる仕組みです👇

[Rory Flynn @Ror_Fly]

KLING TRY-ON → GRAPHIC TEE WORKFLOW.

Midjourney + Illustrator + Kling.

A few extra steps...but worth it.

PROCESS:
1. MJ: Generate Model
2. Kling: Virtual Try-On
3. MJ: Generate Design
4. Illustrator: Mockup
5. Kling: Animate

🧵Details in thread

#klingai #midjourney @Kling_ai

[el.cine @EHuanglu]

This AI filmmaking workflow is incredible!

RenderNet AI’s Canvas has everything we need - Image & Video Generator, lip-sync, pose lock and more.

And it's insane that it can do character consistency with just one image.

I made this music video in just a few hours.

Here’s how:

[AK @_akhaliq]

New Kling, Runway, Luma competitor?

Open text and image video generation model

Pyramidal Flow Matching for Efficient Video Generative Modeling

[Andrew Ng @AndrewYNg]

This week, Google announced a doubling of Gemini Pro 1.5's input context window from 1 million to 2 million tokens, and OpenAI released GPT-4o, which generates tokens 2x faster and 50% cheaper than GPT-4 Turbo and natively accepts and generates multimodal tokens. I view these developments as the latest in an 18-month trend. Given the improvements we've seen, best practices for developers have changed as well.

Since the launch of ChatGPT in Nov 2022, with key milestones that include the releases of GPT-4, Gemini 1.5 Pro, Claude 3 Opus, and Llama 3-70B, many model providers have improved their capabilities in two important ways: (i) reasoning, which allows LLMs to think through complex concepts and and follow complex instructions; and (ii) longer input context windows.

The reasoning capability of GPT-4 and other advanced models makes them quite good at interpreting complex prompts with detailed instructions. Many people are used to dashing off a quick, 1- to 2-sentence query to an LLM. In contrast, when building applications, I see sophisticated teams frequently writing prompts that might be 1 to 2 pages long (my teams call them “mega-prompts”) that provide complex instructions to specify in detail how we’d like an LLM to perform a task. I still see teams not going far enough in terms of writing detailed instructions. For an example of a moderately lengthy prompt, take a look at Claude 3’s system prompt. It’s detailed and gives clear guidance on how Claude should behave.

This is a very different style of prompting than we typically use with LLMs’ web user interfaces, where we might dash off a quick query and, if the response is unsatisfactory, clarify what we want through repeated conversational turns with the chatbot.

Further, the increasing length of input context windows has added another technique to the developer’s toolkit. GPT-3 kicked off a lot of research on few-shot in-context learning. For example, if you’re using an LLM for text classification, you might give a handful — say 1 to 5 examples — of text snippets and their class labels, so that it can use those examples to generalize to additional texts. However, with longer input context windows — GPT-4o accepts 128,000 input tokens, Claude 3 Opus 200,000 tokens, and Gemini 1.5 Pro 1 million tokens (2 million just announced in a limited preview) — LLMs aren’t limited to a handful of examples. With many-shot learning, developers can give dozens, even hundreds of examples in the prompt, and this works better than few-shot learning.

When building complex workflows, I see developers getting good results with this process:
- Write quick, simple prompts and see how it does.
- Based on where the output falls short, flesh out the prompt iteratively. This often leads to a longer, more detailed, prompt, perhaps even a mega-prompt.
- If that’s still insufficient, consider few-shot or many-shot learning (if applicable) or, less frequently, fine-tuning.
- If that still doesn’t yield the results you need, break down the task into subtasks and apply an agentic workflow.

I hope a process like this will help you build applications more easily. If you’re interested in taking a deeper dive into prompting strategies, I recommend Microsoft's Medprompt paper (Nori et al., 2023), which lays out a complex set of prompting strategies that can lead to very good results.

[Original text (with links): https://t.co/UOtLDza1Vh ]

[Soumith Chintala @soumithchintala]

No More GIL!
the Python team has officially accepted the proposal.

Congrats @colesbury on his multi-year brilliant effort to remove the GIL, and a heartfelt thanks to the Python Steering Council and Core team for a thoughtful plan to make this a reality.

https://t.co/58QK2yctRD

[Hugging Face @huggingface]

Llama 2: Now on Hugging Chat 🤗🦙

Try out the 70B Chat model for free with super fast inference, web search, and powered by open-source tools!

👉 https://t.co/TQ0ZaVZcdi

[Sebastian Raschka @rasbt]

Looks like the first open source equivalent of ChatGPT has arrived: https://t.co/4vQ83pcX2H

I.e., an implementation of RLHF (Reinforcement Learning with Human Feedback) on top of Google’s 540 billion parameter PaLM architecture

[François Fleuret @francoisfleuret]

🎶 Something in the Air 🎶

https://t.co/fC4njKAWsn

https://t.co/hQ4RuS2Gum

[Google AI @GoogleAI]

Ensembles of neural models are simple, yet valuable tools for #ML developers. We demonstrate that such ensembles can match or exceed the accuracy of state-of-the-art single models, while being considerably more efficient. Learn more at https://t.co/XJCVdGoSKO

[AK @_akhaliq]

Deep Learning Tools for Audacity: Helping Researchers Expand the Artist’s Toolkit
abs: https://t.co/oaw9r1w5ni
project page: https://t.co/j1FLnmkVow

[Google AI @GoogleAI]

Developing efficient algorithms for distributed computation on very large graphs is an ongoing challenge. Check out a new model that augments the capabilities of #MapReduce, providing a way to solve distributed graph algorithms in fewer computation rounds. https://t.co/8nE2XZ48IS

[AK @_akhaliq]

Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention
pdf: https://t.co/BG4tUJ4MBJ
abs: https://t.co/1KJxItlY39
github: https://t.co/YtHwXlpkfY

[Papers with Code @paperswithcode]

🎉 Introducing... Datasets! We are now indexing 3000+ research datasets from machine learning. Find datasets by task and modality, compare usage over time, browse benchmarks, and much more! Explore the catalogue here: https://t.co/sST84xRiSd

[机器之心|Synced @SyncedTech]

微软工程师用PyTorch实现图注意力网络，可视化效果惊艳 https://t.co/lMMVX19782

[yianchen @cattail1976]

@affige_yang 強大

[KDnuggets @kdnuggets]

Top Python Libraries for Deep Learning, Natural Language Processing & Computer Vision - KDnuggets https://t.co/VFWyiTAsMm

[yianchen @cattail1976]

awesome package!
It's easy to reproduce the result and apply to your own problems

[AK @_akhaliq]

Pytorch implementation of Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
github: https://t.co/k68yBz4JhH

[Jennifer Stirrup #MBA Topics: #AI #Data #Strategy @jenstirrup]

The latest Data Preparation Daily! https://t.co/zAuxWV718S Thanks to @edugil39 #machinelearning #tamr

[yianchen @cattail1976]

#ThisAttackComeFromTaiwan