Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Can Google Gemma DiffusionGemma help fix broken OCR? In theory, denoising tokens in parallel could work better for OCR correction since context is seen upfront? Pointed it at 19th-century newspaper OCR. It corrected better than the autoregressive baseline — at ~8x the speed.

Daniel van Strien

5,732 subscribers

36,105 views • 10 days ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

i compared the top open-source OCR solutions and found which one works the best. featuring: > DeepSeek OCR > Datalab Chandra > Qwen3-VL > Dots OCR > Granite Docling also created an app that lets you run all of these OCR models in one place. 100% open-source.

i compared the top open-source OCR solutions and found which one works the best. featuring: > DeepSeek OCR > Datalab Chandra > Qwen3-VL > Dots OCR > Granite Docling also created an app that lets you run all of these OCR models in one place. 100% open-source.

Akshay 🚀

45,427 views • 5 months ago

Inspired by Brian Roemmele, I set up DeepSeek-OCR on colab. Even with a T4 GPU and 4-bit quantization, it scans a page in about 45 seconds. In this video, you can see that it compressed 527 text tokens to 249 image tokens. DeepSeek-OCR is trained on nearly 100 languages, they say in their paper. So, I'm going to try using it on some old manuscripts written in indic languages soon.

Inspired by Brian Roemmele, I set up DeepSeek-OCR on colab. Even with a T4 GPU and 4-bit quantization, it scans a page in about 45 seconds. In this video, you can see that it compressed 527 text tokens to 249 image tokens. DeepSeek-OCR is trained on nearly 100 languages, they say in their paper. So, I'm going to try using it on some old manuscripts written in indic languages soon.

Ashraff Hathibelagal

226,767 views • 8 months ago

I just built my own OCR app powered by Google's Gemma 3. It runs 100% locally on my computer using Ollama and extracts text from images as structured markdown. The code is open-source, and you can find it in the next tweet.

I just built my own OCR app powered by Google's Gemma 3. It runs 100% locally on my computer using Ollama and extracts text from images as structured markdown. The code is open-source, and you can find it in the next tweet.

Akshay 🚀

477,662 views • 1 year ago

OCR can process characters but it doesn’t understand pixels. OCR has no way to reason about the headers, totals, or checkboxes found in tables, invoices, or forms. In our course with LandingAI, "Document AI: From OCR to Agentic Doc Extraction," we build agents to address these failure modes by breaking documents into pieces, applying the right tools, and mapping information to expected formats. Learn more and enroll today:

OCR can process characters but it doesn’t understand pixels. OCR has no way to reason about the headers, totals, or checkboxes found in tables, invoices, or forms. In our course with LandingAI, "Document AI: From OCR to Agentic Doc Extraction," we build agents to address these failure modes by breaking documents into pieces, applying the right tools, and mapping information to expected formats. Learn more and enroll today:

DeepLearning.AI

21,490 views • 4 months ago

Oracle consensus is at the core of the Chainlink Network. How the Offchain Reporting Protocol (OCR) is advancing oracle consensus ↓

Oracle consensus is at the core of the Chainlink Network. How the Offchain Reporting Protocol (OCR) is advancing oracle consensus ↓

Chainlink

21,836 views • 6 months ago

Sharada manuscripts are among the most difficult Indic texts for OCR. Their lines are often long, curved, and rich in diacritics. We built CURIO, a curvature-aligned OCR to read them. 🧵👇

Sharada manuscripts are among the most difficult Indic texts for OCR. Their lines are often long, curved, and rich in diacritics. We built CURIO, a curvature-aligned OCR to read them. 🧵👇

Ravi Kiran S

36,239 views • 3 months ago

Today's autoregressive models generate one token at a time. Mercury 2 generates tokens in parallel. Over 1,000 tok/sec on standard GPUs, at comparable quality to speed-optimized models. Since launch, the community has been showing what diffusion LLMs can unlock. Thanks to the team at Clyep for the breakdown.

Today's autoregressive models generate one token at a time. Mercury 2 generates tokens in parallel. Over 1,000 tok/sec on standard GPUs, at comparable quality to speed-optimized models. Since launch, the community has been showing what diffusion LLMs can unlock. Thanks to the team at Clyep for the breakdown.

Inception

21,104 views • 1 month ago

Mistral OCR 3 now returns confidence scores, at page level or word level. Know exactly how sure the model is about what it extracted. Available now via API and to test in Mistral Studio.

Mistral OCR 3 now returns confidence scores, at page level or word level. Know exactly how sure the model is about what it extracted. Available now via API and to test in Mistral Studio.

Mistral AI for Developers

28,579 views • 2 months ago

Deepseek OCR now on MLX :) Thanks for the patience! I finally found and fixed the bug on the LM part that was stopping it from understanding the Vision Tokens. New release later today, probably alongside batch infer to allow you to process N documents at once.

Deepseek OCR now on MLX :) Thanks for the patience! I finally found and fixed the bug on the LM part that was stopping it from understanding the Vision Tokens. New release later today, probably alongside batch infer to allow you to process N documents at once.

Prince Canuma

56,286 views • 7 months ago

Bengaluru-based startup Sarvam AI is drawing global attention with its latest AI models, Sarvam Vision and Bulbul V3. The company claims its technology has outperformed global giants like Google Gemini and ChatGPT in key optical character recognition (OCR) benchmarks. If verified, the breakthrough could mark a major moment for India’s growing AI ecosystem. What makes Sarvam Vision different? And how does Bulbul V3 push the boundaries of OCR performance? #SarvamAI

Bengaluru-based startup Sarvam AI is drawing global attention with its latest AI models, Sarvam Vision and Bulbul V3. The company claims its technology has outperformed global giants like Google Gemini and ChatGPT in key optical character recognition (OCR) benchmarks. If verified, the breakthrough could mark a major moment for India’s growing AI ecosystem. What makes Sarvam Vision different? And how does Bulbul V3 push the boundaries of OCR performance? #SarvamAI

NDTV

51,372 views • 4 months ago

MinerU-Diffusion A 2.5B diffusion-based OCR model that replaces slow autoregressive decoding with parallel block-wise diffusion, achieving up to 3.2x faster inference while improving robustness on complex documents with tables, formulas, and layouts.

MinerU-Diffusion A 2.5B diffusion-based OCR model that replaces slow autoregressive decoding with parallel block-wise diffusion, achieving up to 3.2x faster inference while improving robustness on complex documents with tables, formulas, and layouts.

DailyPapers

15,304 views • 2 months ago

I made my own visual novel OCR app that overlays text directly on top of the game. It also has context-aware dictionary results and translations. Also Anki support.

I made my own visual novel OCR app that overlays text directly on top of the game. It also has context-aware dictionary results and translations. Also Anki support.

funorange

223,065 views • 12 days ago

I'm about to piss off a 13 billion dollar industry - the OCR market. Here's why: Monday: I optimized my PDF-to-Excel converter landing page for SEO. Tuesday: I started ranking #1 organically for "pdf to excel ai", a phrase with 8K+ monthly searches and growing. The Google Ads competition is insane. Wednesday: I was previously using a basic OCR API and I replaced it with the Gemini 2.0 Flash model, which is insanely good at PDF parsing. Thursday: I bought the domain pdftoexcelai[dot]com.

I'm about to piss off a 13 billion dollar industry - the OCR market. Here's why: Monday: I optimized my PDF-to-Excel converter landing page for SEO. Tuesday: I started ranking #1 organically for "pdf to excel ai", a phrase with 8K+ monthly searches and growing. The Google Ads competition is insane. Wednesday: I was previously using a basic OCR API and I replaced it with the Gemini 2.0 Flash model, which is insanely good at PDF parsing. Thursday: I bought the domain pdftoexcelai[dot]com.

David Bressler 📊

12,417 views • 1 year ago

Let's talk parsing charts 📊📈. Last week we released ParseBench, the first document OCR benchmark for AI agents. New in ParseBench: ChartDataPointMatch. Most document look at a chart and OCR the caption. Agents need the actual numbers. That's the gap between "OCR'd the text around the chart" and "actually read the chart." More about ParseBench, the GitHub code, Hugging Face dataset, and scientific paper→

Let's talk parsing charts 📊📈. Last week we released ParseBench, the first document OCR benchmark for AI agents. New in ParseBench: ChartDataPointMatch. Most document look at a chart and OCR the caption. Agents need the actual numbers. That's the gap between "OCR'd the text around the chart" and "actually read the chart." More about ParseBench, the GitHub code, Hugging Face dataset, and scientific paper→

LlamaIndex 🦙

13,987 views • 2 months ago

Multimodal AI is here 🤯 GPT-4 can now turn your images into a text file in a snap with the new code interpreter model. Witness the OCR magic in action 🔥

Multimodal AI is here 🤯 GPT-4 can now turn your images into a text file in a snap with the new code interpreter model. Witness the OCR magic in action 🔥

Shubham Saboo

727,655 views • 3 years ago

Ever wonder what’s behind your OCR? It’s Donald.

Ever wonder what’s behind your OCR? It’s Donald.

Reducto

10,691 views • 2 months ago

People are overlooking Google Gemini Realtime models for computer use It gave me sub 100ms latency with computer use. It has a much larger context window and is much cheaper as well Combine that with local OCR and local screen detection model based on Omniparser by Microsoft it works under 100ms action taking when combined with Cua I also put in a harness for Nous Research Hermes with it. You can access it all at your tip of your cursor. You can draw on your screen to give your agents a context And I am making it Open Source! Link in the Comments Sundar Pichai Min-Liang Tan Mojtaba Seyedhosseini

People are overlooking Google Gemini Realtime models for computer use It gave me sub 100ms latency with computer use. It has a much larger context window and is much cheaper as well Combine that with local OCR and local screen detection model based on Omniparser by Microsoft it works under 100ms action taking when combined with Cua I also put in a harness for Nous Research Hermes with it. You can access it all at your tip of your cursor. You can draw on your screen to give your agents a context And I am making it Open Source! Link in the Comments Sundar Pichai Min-Liang Tan Mojtaba Seyedhosseini

Milind S

37,839 views • 20 days ago

We used DeepSeek OCR to extract every dataset from tables/charts across 500k+ AI arXiv papers for $1000 🚀 See which benchmarks are trending and discover datasets you didn't know existed Doing the same task with Mistral OCR would've cost $7500 👀

We used DeepSeek OCR to extract every dataset from tables/charts across 500k+ AI arXiv papers for $1000 🚀 See which benchmarks are trending and discover datasets you didn't know existed Doing the same task with Mistral OCR would've cost $7500 👀

alphaXiv

285,001 views • 8 months ago

launching our open source OCR tool today! try it out with some terrible pdfs and let me know how it goes:

launching our open source OCR tool today! try it out with some terrible pdfs and let me know how it goes:

Tyler Maran

342,540 views • 1 year ago

We raised $1.5m to launch the world’s first LLM built for developer tasks. Run OCR, web scraping, STT, and classification that you can rely on for any dev tasks. Available in Beta starting today!

We raised $1.5m to launch the world’s first LLM built for developer tasks. Run OCR, web scraping, STT, and classification that you can rely on for any dev tasks. Available in Beta starting today!

Yoeven

93,421 views • 8 months ago