# forck.live — 25–31 July 2022 > AI lab announcements, news reports and podcasts. forck.live follows what companies announce about their own products. Press may corroborate an official story; it never leads and never appears as an item of its own. A missing date stays missing. A title is the publisher's words — where a company does not publish in English, an English rendering is listed separately as `Title in English` and named as this site's translation, never as theirs. This file is one issue of the weekly archive: every confirmed first-party announcement of ISO week 2022-W30, 25–31 July 2022, newest first. The window never moves, so this URL always names the same seven days — but a post published inside them and detected later is added when it arrives, so a closed week can still gain entries. Every entry is one first-party announcement, and carries the same eight fields in this order before any others: Company · Announced · Category · Coverage · Announcement · Group · Source · Record `Source` is the company's own page the announcement was traced to and `Record` is this site's page for it. Both are safe to cite. `Coverage` is a number of outlets, or `not counted` — which means this record reached no count of its own, never that nobody wrote about it. `Category` is this tracker's own analysis, and is authoritative wherever it is stated. `not stated` means either that nothing has analysed the post yet or that the analysis placed it outside this taxonomy; neither is a verdict about the announcement. `Announcement` is whether the post is news of something the company did, rather than a how-to, a customer story, a changelog listing or a staffing note. `Group` is which of this record's groups the entry falls in, carried on the entry itself so it holds where the groups are headings and where they are not. One of: `models`, `covered`, `announcements`, `routine`. Anything after `Record` is held detail that only some announcements carry. - This week's page: https://forck.live/week/2022-W30 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Replit — Get Replit Famous - Company: Replit (replit.com) - Announced: 2022-07-28 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/replit-famous - Record: https://forck.live/items/18082-get-replit-famous - Subject: Replit Agent Replit introduced a Following feature that lets users follow friends and creators, with their social activity appearing in a new Following feed on the homepage. The company also plans to experiment with monetization options for creators, such as a native app store, community extensions, or consumer-facing ads. ### Hugging Face — Introducing new audio and vision documentation in 🤗 Datasets - Company: Hugging Face (huggingface.co) - Announced: 2022-07-28 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/datasets-docs-update - Record: https://forck.live/items/2172-introducing-new-audio-and-vision-documentation-in-datasets - Subject: Platform Hugging Face announced new documentation and features for the 🤗 Datasets library, including an updated Quickstart with audio and image dataset examples, dedicated modality-specific guides, and an ImageFolder dataset builder that simplifies loading image datasets for various tasks without needing a custom loading script. ### Replit — The New Game: Engaging 15M Next Generation Developers - Company: Replit (replit.com) - Announced: 2022-07-27 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replit.com/blog/company-profiles - Record: https://forck.live/items/18141-the-new-game-engaging-15m-next-generation-developers - Subject: Replit Agent Replit announced partnerships with 19 companies, providing over 35 templates for its users. The platform now has over 15 million users, 20 million websites and apps created and hosted, and 10 billion monthly visits. Replit also introduced a verified badge for companies. ### LG AI Research — [AAAI 2022] Learning Parameterized Task Structure for Generalization to Unseen Entities - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-27 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=235 - Record: https://forck.live/items/4938-aaai-2022-learning-parameterized-task-structure-for-generalization-to-unseen - Subject: EXAONE LG AI Research introduces Parameterized Subtask Graph Inference (PSGI), a method for inferring hierarchical and compositional task structures in a first-order logic manner, enabling generalization to unseen entities. The method is tested on cooking, mining, and simulated domains, showing improved efficiency and generalizability over prior work. ### Hugging Face — Faster Text Generation with TensorFlow and XLA - Company: Hugging Face (huggingface.co) - Announced: 2022-07-27 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/tf-xla-generate - Record: https://forck.live/items/2173-faster-text-generation-with-tensorflow-and-xla - Subject: Platform The Hugging Face transformers library now supports XLA compilation for text generation with TensorFlow, achieving up to 100x speedup and surpassing PyTorch performance. ### Hugging Face — Deploying TensorFlow Vision Models in Hugging Face with TF Serving - Company: Hugging Face (huggingface.co) - Announced: 2022-07-25 - Category: developer-tool-release - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/tf-serving-vision - Record: https://forck.live/items/2174-deploying-tensorflow-vision-models-in-hugging-face-with-tf-serving - Subject: Platform To deploy TensorFlow vision models from Hugging Face Transformers using TensorFlow Serving, with a focus on a Vision Transformer (ViT) for image classification, including steps for saving the model, preprocessing, and postprocessing.