✷ The archive · 11 labs · 3,433 dispatches
The labs
No more opening a dozen official blogs every morning. First-hand releases from OpenAI, Anthropic, DeepMind and the rest, each with its substance pulled out.
Hugging Face Director of Machine Learning Insights Part 4
Four Machine Learning Directors share industry-specific insights on the impact, challenges, and integration pitfalls of ML in e-commerce, engineering, education, and SaaS.
Hugging Face Inference Solutions Overview November 2022
Hugging Face announced a suite of free and paid inference options—including a widget, API, Inference Endpoints, and Spaces—to simplify model testing, deployment, and production scaling.
Hugging Face Accelerating Document AI
Hugging Face provides a comprehensive guide to using open-source multimodal models to automate document classification, parsing, and visual question answering for enterprise workflows.
Sentiment Analysis on Encrypted Data with Homomorphic Encryption
Hugging Face demonstrates how to use the Concrete-ML library to perform sentiment analysis on encrypted data using a combination of BERT transformers and XGBoost with Fully Homomorphic Encryption (FHE).
Hugging Face and arXiv Integration for Machine Learning Demos
Hugging Face has integrated Hugging Face Spaces with arXivLabs to provide interactive machine learning demos directly on arXiv paper abstract pages.
OFA: Towards Building a One-For-All Model
OFA is a unified multimodal pretrained model that unifies understanding and generation tasks across modalities into a single framework using instruction-based multitask pretraining.
Hugging Face Pricing Update November 2022
Hugging Face has transitioned to a compute-based monetization model, sunsetting the Paid tier of the Inference API in favor of Inference Endpoints and hardware upgrades for Spaces.
Contrastive Search for Human-Level Text Generation in Transformers
Hugging Face has integrated Contrastive Search into the transformers library, a decoding method that prevents model degeneration and maintains semantic coherence across 16 languages using off-the-shelf models.
Dreambooth Stable Diffusion fine‑tuning guide with Diffusers
Hugging Face released detailed recommendations for training Stable Diffusion with Dreambooth using the Diffusers library, showing that low learning rates, enough steps, prior preservation for faces, and text‑encoder fine‑tuning yield the highest quality results.
Measuring Progress on Scalable Oversight for Large Language Models
Anthropic researchers propose a framework for empirically studying scalable oversight to ensure humans can supervise AI systems that may eventually outperform them on complex tasks.
OpenAI DALL·E API public beta announcement
OpenAI announced that the DALL·E image generation API is now available in public beta, letting developers integrate state‑of‑the‑art image creation and built‑in moderation into their apps.
Fine-Tuning Whisper for Multilingual ASR with Hugging Face Transformers
Hugging Face provides a comprehensive guide on fine-tuning OpenAI's Whisper model for multilingual automatic speech recognition (ASR), demonstrating a 31.5% absolute WER improvement on Hindi using only 8 hours of data.
Hugging Face Optimum Intel and OpenVINO Integration
Hugging Face has integrated Intel OpenVINO into Optimum Intel, enabling accelerated inference and quantization for Transformer models on Intel hardware.
Evaluating Language Model Bias with 🤗 Evaluate
Hugging Face added bias metrics—toxicity, language polarity, and HONEST—to the 🤗 Evaluate library, enabling systematic measurement of harmful language in causal language models.
Distributed Training with PyTorch DDP, Accelerate, and Transformers Trainer
Hugging Face explains how to implement distributed training using three levels of abstraction: native PyTorch DDP, the Accelerate library, and the high-level Transformers Trainer API.
OpenAI Scaling Laws for Reward Model Overoptimization
OpenAI identifies that optimizing against an imperfect proxy reward model eventually degrades ground truth performance, following a predictable scaling law based on model parameters and dataset size.
MTEB: Massive Text Embedding Benchmark
Hugging Face introduced MTEB, a massive and multilingual benchmark consisting of 56 datasets across 8 tasks to evaluate the performance of text embedding models.
Hugging Face Inference Endpoints
Hugging Face Inference Endpoints is a managed service that allows users to deploy machine learning models from the Hugging Face Hub to scalable, secure cloud infrastructure with a few clicks.
Stable Diffusion JAX and Flax Integration
Hugging Face Diffusers version 0.5.1 introduces support for Flax, enabling high-speed Stable Diffusion inference on Google TPUs via JAX.
Hugging Face BLOOM Inference Optimization
Hugging Face achieved a 5x reduction in latency and a 50x increase in throughput for the BLOOM model by transitioning from Pipeline Parallelism to Tensor Parallelism and implementing custom CUDA kernels.
Hugging Face Introduces DOI Support for Models and Datasets
Hugging Face now lets users generate Digital Object Identifiers (DOIs) for Hub models and datasets, providing permanent, citable links that persist across versions.
Japanese Stable Diffusion release by rinna
rinna released Japanese Stable Diffusion, a Japanese‑language fine‑tuned version of Stable Diffusion that generates culturally appropriate images from Japanese prompts.
Hugging Face Zero-Shot Evaluation on the Hub
Hugging Face has introduced zero-shot evaluation for causal language models on the Hub, enabling users to benchmark models up to 66 billion parameters without writing code.
OpenAI DALL·E Beta Now Available Without Waitlist
OpenAI has removed the waitlist for the DALL·E beta, allowing immediate sign-up for a system that currently supports over 1.5 million users creating more than 2 million images daily.
Hugging Face AutoTrain Image Classification
Hugging Face has added Image Classification to AutoTrain, enabling users to train custom image categorization models without writing code or configuring hyperparameters.
Hugging Face Accelerate: Running Large Models with PyTorch
Hugging Face Accelerate enables the execution of massive AI models on consumer hardware by leveraging PyTorch's meta device and sharded checkpoints to manage memory across GPUs, CPU RAM, and disk.
SetFit: Efficient Few-Shot Learning Without Prompts
Hugging Face introduces SetFit, a prompt-free framework for few-shot fine-tuning of Sentence Transformers that achieves high accuracy with minimal labeled data.
Hugging Face Ethics and Society Newsletter #1
Hugging Face introduces its Ethics and Society newsletter and outlines a decentralized, value-driven approach to operationalizing AI ethics through collaboration, transparency, and responsibility.
OpenAI Whisper Release
OpenAI has released Whisper, an automatic speech recognition (ASR) system trained on 680,000 hours of multilingual and multitask supervised data to improve robustness across accents, noise, and languages.
Incredibly Fast BLOOM Inference with DeepSpeed and Accelerate
Hugging Face demonstrates sub‑millisecond per‑token generation for the 176B‑parameter BLOOM model using DeepSpeed‑Inference tensor parallelism and Accelerate pipeline parallelism on 8×80 GB A100 GPUs.
Anthropic Toy Models of Superposition Research
Anthropic researchers use small ReLU networks to demonstrate that models can represent more features than they have dimensions through a phenomenon called superposition, provided those features are sparse.
Diffusers 0.3 release adds image‑to‑image, textual inversion, inpainting, GPU optimizations, Mac MPS, ONNX support and new docs
Hugging Face announced Diffusers version 0.3, introducing image‑to‑image, textual inversion, experimental inpainting, smaller‑GPU optimizations, Mac MPS support, an ONNX exporter, expanded documentation, and a wave of community projects.
Training Decision Transformers for Offline Reinforcement Learning
Hugging Face provides a guide and implementation for training an offline Decision Transformer from scratch to solve the HalfCheetah environment using the transformers Trainer and a custom data collator.
Training Language Models with Megatron-LM
Hugging Face provides a guide on using NVIDIA's Megatron-LM framework to efficiently pre-train large language models on GPUs, including integration with the Transformers library.
DALL·E Outpainting Feature Release
OpenAI has introduced Outpainting for DALL·E, allowing users to extend images beyond their original borders using natural language descriptions while maintaining visual consistency.
OpenRAIL: Towards open and responsible AI licensing frameworks
Hugging Face announced OpenRAIL, a set of AI‑specific licenses that combine open access with use‑based restrictions to promote responsible deployment of machine‑learning models.
OpenAI Announces Alignment Strategy Using RL from Human Feedback and InstructGPT
OpenAI detailed its alignment approach, highlighting RL from human feedback and InstructGPT as effective, low‑cost methods for making language models follow human intent while acknowledging current limitations and future goals for AI‑driven alignment research.
Visualize proteins on Hugging Face Spaces
Hugging Face provides a guide on integrating 3Dmol.js into Hugging Face Spaces via Gradio to enable 3D protein structure visualization in the browser.
Pre-training BERT with Hugging Face Transformers and Habana Gaudi
Hugging Face demonstrates how to pre-train BERT-base from scratch using Habana Gaudi DL1 instances on AWS, achieving a 25% cost reduction compared to NVIDIA V100 GPU-based training.
Stable Diffusion with Diffusers
Hugging Face announced how to run the text-to-image Stable Diffusion model using the Diffusers library, providing code, licensing details, and explanations of the model’s latent diffusion architecture.
Anthropic Red Teaming Language Models Report – Methods, Scaling Behaviors, and Lessons Learned
Anthropic released a detailed study showing that reinforcement‑learning‑from‑human‑feedback (RLHF) models become harder to red‑team as they scale, while other model types show flat red‑teamability, and they published a 38,961‑attack dataset to help the community improve safety.
Deploying Hugging Face ViT on Vertex AI
Hugging Face demonstrates how to deploy a Vision Transformer (ViT) model on Google Cloud's Vertex AI platform to achieve Kubernetes-level scalability with significantly less code.
Vision Transformers on Hugging Face Optimum Graphcore
Hugging Face and Graphcore enable the efficient fine-tuning of Vision Transformer (ViT) models on Intelligence Processing Units (IPUs) using the Optimum library, demonstrated through a multi-label chest X-ray classification task.
Hugging Face integrates LLM.int8() 8-bit matrix multiplication into Transformers and Accelerate
Hugging Face released an integration of the LLM.int8() 8-bit quantization technique into the Transformers and Accelerate libraries, enabling inference of very large language models with no performance degradation and up to 2× memory savings.
Hugging Face's TensorFlow Philosophy
Hugging Face outlines its technical approach to TensorFlow integration, emphasizing deep alignment with Keras and XLA to optimize performance and deployment for transformer models.
Skops library enables scikit-learn model hosting, documentation, and collaboration on Hugging Face Hub
Hugging Face announced Skops, a library that lets users package, document, and push scikit-learn models to the Hub with automatic model cards and inference widgets.
Deploying Hugging Face ViT on Kubernetes with TF Serving
Hugging Face provides a technical guide on scaling Vision Transformer (ViT) deployments using Docker and Kubernetes with TensorFlow Serving to enable production-grade autoscaling and resource management.
OpenAI Moderation Endpoint Release
OpenAI has released a free Moderation endpoint providing GPT-based classifiers to help developers detect and block prohibited content in their applications.
Hugging Face Sentence Transformers Training Guide (historical reference)
Hugging Face announced an outdated tutorial for training and fine-tuning Sentence Transformers models, explaining model architecture, dataset formats, loss functions, and how to upload models, while directing users to newer training APIs.
Hugging Face Deep RL Class Unit 8: Proximal Policy Optimization (PPO) Explained
Hugging Face introduced a detailed walkthrough of Proximal Policy Optimization (PPO), showing how clipping the policy‑update ratio stabilizes training and providing a from‑scratch PyTorch implementation tested on CartPole‑v1 and LunarLander‑v2.