TLDRocket
Sign in
Latest 8 insights from LogicMonitor’s Elevate event: AI-driven observability... — SiliconANGLE When it came to Cal AI’s viral success, a small army of influencers pa... — TechCrunch Meta releases 816B-parameter Muse Spark Open for coding and agent work — SiliconANGLE AI #190: Scion of the Times — Zvi (Don't Worry About the Vase) Vinci targets faster chip design and simulation with physics-driven AI — SiliconANGLE Hackers hijacked 1,900 GitHub accounts. Then they went after 30,000 re... — The New Stack Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint — Hugging Face You Have No Idea How Close the AI Bubble Is to Popping — The Algorithmic Bridge

The AI intelligence platform

Every AI story that matters — and the intelligence behind it.

TLDRocket reads all relevant sources, removes duplicate coverage, and publishes a short neutral summary of every story, linking back to the original. Free, no spam, unsubscribe anytime.

Add to Slack

Every story also updates live profiles event timelines weekly rankings the AI Market Index

Monday, 22 August 2022

Pre-Train BERT with Hugging Face Transformers and Habana Gaudi

Hugging Face 4 years ago 53

Hugging Face released a tutorial for pre-training BERT-base from scratch using Habana Gaudi accelerators on AWS, leveraging the Transformers and Optimum Habana libraries with masked-language modeling. The training ran for 100,000 steps with a global batch size of 256 over approximately 12.5 hours on a DL1 instance with 8 HPU-cores. Users can now train custom BERT models on Gaudi hardware by following the guide's four steps: dataset preparation, tokenizer training, dataset preprocessing, and distributed model pre-training.

Read the full story

Stable Diffusion with 🧨 Diffusers

Hugging Face 4 years ago 50

Stability AI and collaborators created Stable Diffusion, a text-to-image model trained on 512x512 images from the LAION-5B dataset, which can be run through the Diffusers library with just a few lines of code. The model uses latent diffusion with a default of 50 inference steps and achieves an 8x8 spatial compression ratio, reducing memory requirements so that 512×512 images can be generated on 16GB GPUs. Users can customize outputs by adjusting parameters like guidance_scale (7 to 8.5 recommended), num_inference_steps, and image dimensions, with quality improving as step count increases but generation time slowing proportionally.

Read the full story

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.