正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

💡 Teach models repeatable behavior with dataset distillation. Train smaller models on clean, structured examples to stabilize parser outputs.

Google for Developers

2,577,312 subscribers

17,457 次观看 • 2 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

From sketch to SVG in seconds. Arrow 1.1 vectorizes technical drawings with precision, producing clean, structured outputs that stay true to the original.

From sketch to SVG in seconds. Arrow 1.1 vectorizes technical drawings with precision, producing clean, structured outputs that stay true to the original.

QuiverAI

11,778 次观看 • 4 个月前

Xue et al., "DepthART: Scaling Foundation Monocular Depth to Tiny Models" Well-strategized distillation with camera-conditioned fine-tuning. 1000 FPS on RTX A6000, 200 FPS on Jetson Orin NX.

Xue et al., "DepthART: Scaling Foundation Monocular Depth to Tiny Models" Well-strategized distillation with camera-conditioned fine-tuning. 1000 FPS on RTX A6000, 200 FPS on Jetson Orin NX.

Kwang Moo Yi

16,667 次观看 • 1 个月前

Was really interesting to hear John, Beren, and Charlie speculate about why Sonnet 5 and Opus 5 feel like worse models than GLM 5.3 (despite the fact that Anthropic can do raw logit distillation from Fable, and can also train Sonnet/Opus on the environments from which Fable was trained). Led to some interesting thoughts about value of distillation, what it takes to do distillation effectively, and what kinds of model behaviors are hard to extract from distillation.

Was really interesting to hear John, Beren, and Charlie speculate about why Sonnet 5 and Opus 5 feel like worse models than GLM 5.3 (despite the fact that Anthropic can do raw logit distillation from Fable, and can also train Sonnet/Opus on the environments from which Fable was trained). Led to some interesting thoughts about value of distillation, what it takes to do distillation effectively, and what kinds of model behaviors are hard to extract from distillation.

Dwarkesh Patel

127,381 次观看 • 6 天前

Claude distillation has been a big topic this week while I am (coincidentally) writing Chapter 8 on model distillation. In that context, I shared some utilities to generate distillation data from all sorts of open-weight models via OpenRouter and Ollama:

Claude distillation has been a big topic this week while I am (coincidentally) writing Chapter 8 on model distillation. In that context, I shared some utilities to generate distillation data from all sorts of open-weight models via OpenRouter and Ollama:

Sebastian Raschka

62,838 次观看 • 6 个月前

Today, we're introducing our document parser built specifically for RAG. The parser combines the best vision, OCR, and vision language models to deliver unmatched accuracy. Try it for free today—the first 500+ pages are on us! 🧵 1/

Today, we're introducing our document parser built specifically for RAG. The parser combines the best vision, OCR, and vision language models to deliver unmatched accuracy. Try it for free today—the first 500+ pages are on us! 🧵 1/

Douwe Kiela

1,308,680 次观看 • 1 年前

Introducing Subtask Annotations in LeRobot Datasets. LeRobot now supports subtasks as a native dataset field, so you can label task progression frame-by-frame, train hierarchical policies, and build stronger reward models with stage-aware supervision. Available today, with an annotation space + tooling to easily add subtasks and high-level dialogue to your dataset.

Introducing Subtask Annotations in LeRobot Datasets. LeRobot now supports subtasks as a native dataset field, so you can label task progression frame-by-frame, train hierarchical policies, and build stronger reward models with stage-aware supervision. Available today, with an annotation space + tooling to easily add subtasks and high-level dialogue to your dataset.

LeRobot

16,524 次观看 • 7 个月前

Generate images and video directly from your terminal with Claude + our new genmedia CLI. • Search and run fal models from CLI • Generate structured commands with Claude • Run async media workflows • Download outputs locally • Use installable agent skills Watch the full tutorial:

Generate images and video directly from your terminal with Claude + our new genmedia CLI. • Search and run fal models from CLI • Generate structured commands with Claude • Run async media workflows • Download outputs locally • Use installable agent skills Watch the full tutorial:

fal

11,549 次观看 • 4 个月前

5️⃣ object detection and image segmentation use cases with Ultralytics 🎉 ✅ Tomato detection ✅ Teeth segmentation ✅ Pill detection ✅ Cabbage detection ✅ Fire detection 💡All models were trained on the custom dataset using Ultralytics YOLO11. #ml #research #usecases #ai

5️⃣ object detection and image segmentation use cases with Ultralytics 🎉 ✅ Tomato detection ✅ Teeth segmentation ✅ Pill detection ✅ Cabbage detection ✅ Fire detection 💡All models were trained on the custom dataset using Ultralytics YOLO11. #ml #research #usecases #ai

Muhammad Rizwan Munawar

12,501 次观看 • 1 年前

Andrej Karpathy: Internet training data is terrible, so big models end up compressing "memory" instead of doing cognitive work Use intelligent models to filter to the cognitive core With cleaner data, smaller models, likely distilled from a stronger one, are enough

Andrej Karpathy: Internet training data is terrible, so big models end up compressing "memory" instead of doing cognitive work Use intelligent models to filter to the cognitive core With cleaner data, smaller models, likely distilled from a stronger one, are enough

Haider.

384,349 次观看 • 11 个月前

✨✨ Announcing Open Source Dataset from excavators in real construction sites retrofitted by Flywheel (YC S25)! ✨✨ This 100hrs of observation+action data enables training autonomy models for excavators. We were able to train a small task model from 6 hours of dataset on a Kubota U17 on Y Combinator demo day! Link in comments!

✨✨ Announcing Open Source Dataset from excavators in real construction sites retrofitted by Flywheel (YC S25)! ✨✨ This 100hrs of observation+action data enables training autonomy models for excavators. We were able to train a small task model from 6 hours of dataset on a Kubota U17 on Y Combinator demo day! Link in comments!

Jash Mota

36,617 次观看 • 11 个月前

Anthropic CEO Dario Amodei says big, powerful AI models will spawn and orchestrate smaller models to assist with tasks, creating a swarm intelligence that will decrease the need for human input

Anthropic CEO Dario Amodei says big, powerful AI models will spawn and orchestrate smaller models to assist with tasks, creating a swarm intelligence that will decrease the need for human input

Tsarathustra

184,332 次观看 • 2 年前

Introducing Unsloth Studio ✨ A new open-source web UI to train and run LLMs. • Run models locally on Mac, Windows, Linux • Train 500+ models 2x faster with 70% less VRAM • Supports GGUF, vision, audio, embedding models • Auto-create datasets from PDF, CSV, DOCX • Self-healing tool calling and code execution • Compare models side by side + export to GGUF GitHub: Blog and Guide: Available now on Hugging Face, NVIDIA, Docker and Colab.

Introducing Unsloth Studio ✨ A new open-source web UI to train and run LLMs. • Run models locally on Mac, Windows, Linux • Train 500+ models 2x faster with 70% less VRAM • Supports GGUF, vision, audio, embedding models • Auto-create datasets from PDF, CSV, DOCX • Self-healing tool calling and code execution • Compare models side by side + export to GGUF GitHub: Blog and Guide: Available now on Hugging Face, NVIDIA, Docker and Colab.

Unsloth AI

1,740,792 次观看 • 6 个月前

Train Once, Earn Limitless. Introducing Moonbase Your All-in-one AI Model Marketplace, only on FLock. Train AI models, create AI agents with standard API, and earn when AI agents use your models. Coming soon. Join the waitlist:

Train Once, Earn Limitless. Introducing Moonbase Your All-in-one AI Model Marketplace, only on FLock. Train AI models, create AI agents with standard API, and earn when AI agents use your models. Coming soon. Join the waitlist:

FLock.io

277,143 次观看 • 1 年前

New Short Course: Getting Structured LLM Output! Learn how to get structured outputs from your LLM applications in this course, built in partnership with .txt, and taught by Will Kurt, a Founding Engineer, and , Developer Relations Engineer. It's challenging for software to automatically parse through an LLM's freeform text outputs. Structured outputs—like JSON—solve this by converting natural language into consistent, clear, data that a machine can read and process. This course teaches you how to generate structured outputs while building several use cases, including a social media analysis agent. You’ll learn about structured outputs and efficient ways to generate outputs in your defined schema or format. You’ll begin by using structured output APIs, then use re-prompting libraries like “instructor” to generate structured output. Finally, you’ll learn how constrained decoding works; this is a very clever technique in which constraints are applied on each subsequent token generated, blocking any tokens that don’t fit your defined schema. In detail, you’ll: - Learn why structured outputs are important, how they allow for scalable software development, and the different approaches to generate them, including vendor-provided APIs, re-prompting libraries, and structured generation. - Build a simple social media agent using OpenAI’s structured output API, learn how to define a model's desired structured output using Pydantic, and perform basic programming with your outputs, such as importing structured data into a data frame using pandas. - Learn how to use the open-source library "instructor," which checks the structured output of the model and re-prompts the model until it validates the desired output, and explore the limitations of this approach. - Understand how structured generation by the “outlines” library works by modifying LLM logits, on a per-generated-token basis based on the desired format, to give a particular output structure. - Learn how regular expressions, which outlines works with, are represented as finite-state machines, and how they can be used to develop a range of structured outputs beyond JSON. By the end of this course, you’ll have broadened your knowledge of the approaches you can use to get structured outputs from your LLM applications. Please sign up here:

New Short Course: Getting Structured LLM Output! Learn how to get structured outputs from your LLM applications in this course, built in partnership with .txt, and taught by Will Kurt, a Founding Engineer, and , Developer Relations Engineer. It's challenging for software to automatically parse through an LLM's freeform text outputs. Structured outputs—like JSON—solve this by converting natural language into consistent, clear, data that a machine can read and process. This course teaches you how to generate structured outputs while building several use cases, including a social media analysis agent. You’ll learn about structured outputs and efficient ways to generate outputs in your defined schema or format. You’ll begin by using structured output APIs, then use re-prompting libraries like “instructor” to generate structured output. Finally, you’ll learn how constrained decoding works; this is a very clever technique in which constraints are applied on each subsequent token generated, blocking any tokens that don’t fit your defined schema. In detail, you’ll: - Learn why structured outputs are important, how they allow for scalable software development, and the different approaches to generate them, including vendor-provided APIs, re-prompting libraries, and structured generation. - Build a simple social media agent using OpenAI’s structured output API, learn how to define a model's desired structured output using Pydantic, and perform basic programming with your outputs, such as importing structured data into a data frame using pandas. - Learn how to use the open-source library "instructor," which checks the structured output of the model and re-prompts the model until it validates the desired output, and explore the limitations of this approach. - Understand how structured generation by the “outlines” library works by modifying LLM logits, on a per-generated-token basis based on the desired format, to give a particular output structure. - Learn how regular expressions, which outlines works with, are represented as finite-state machines, and how they can be used to develop a range of structured outputs beyond JSON. By the end of this course, you’ll have broadened your knowledge of the approaches you can use to get structured outputs from your LLM applications. Please sign up here:

Andrew Ng

89,792 次观看 • 1 年前

Ted Chiang's essay made some good points about art and intentionality, but it is premature to debate about AI making art. For now, AI doesn't exist autonomously. Humans make art with AI models trained on humans. We train our own models on our own datasets. This project took 18 months to prompt. I see making models as art.

Ted Chiang's essay made some good points about art and intentionality, but it is premature to debate about AI making art. For now, AI doesn't exist autonomously. Humans make art with AI models trained on humans. We train our own models on our own datasets. This project took 18 months to prompt. I see making models as art.

Holly Herndon

66,466 次观看 • 2 年前

Watch this 45 min video to learn how to create synthetic datasets and train tiny (100M params) local language models that expertise on narrow tasks. Code, datasets, models, harnesses all in comments.

Watch this 45 min video to learn how to create synthetic datasets and train tiny (100M params) local language models that expertise on narrow tasks. Code, datasets, models, harnesses all in comments.

AVB

41,103 次观看 • 3 个月前