# forck.live — 6–12 February 2023 > 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 2023-W06, 6–12 February 2023, 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/2023-W06 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Replit — Altimeter Capital: Supporting builders in AI with Bounties - Company: Replit (replit.com) - Announced: 2023-02-11 - Category: not stated - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/bounties-altimeter-case-study - Record: https://forck.live/items/18128-altimeter-capital-supporting-builders-in-ai-with-bounties - Subject: Replit Agent Bounties is a marketplace where anyone can connect with and contract top software creators from the Replit community. These developers are known as Bounty Hunters. The Bounty Hunter community on Replit is global and includes thousands of vetted developers ranging from young prodigy hackers to senior software engineers based in Silicon Valley. About the Bounty Poster Altimeter is a leading technology-focused investment firm built by founders for founders. Their mission is to help visionary entrepreneurs build iconic companies, disrupt markets, and improve lives through all stages of growth. Altimeter manages a variety of venture and public funds and serves as an expert long-term partner to companies as they enter the public markets. About the Bounty OpenAI’s launch of ChatGPT in November 2022 led to a massive increase in attention and engagement towards LLMs and their ability to power AI-integrated products. The Altimeter team wanted to find a creative way to support builders in the space and explore the latest capabilities of LLMs and libraries like LangChain . To find top developer talent, they turned to Replit Bounties. … ### Hugging Face — Parameter-Efficient Fine-Tuning using 🤗 PEFT - Company: Hugging Face (huggingface.co) - Announced: 2023-02-10 - Category: developer-tool-release - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/peft - Record: https://forck.live/items/2098-parameter-efficient-fine-tuning-using-peft - Subject: Platform Hugging Face introduces the PEFT library for parameter-efficient fine-tuning of billion-scale models on low-resource hardware. It supports methods like LoRA, Prefix Tuning, Prompt Tuning, and P-Tuning, and is integrated with Transformers and Accelerate, enabling fine-tuning on consumer hardware with tiny checkpoints. ### LG AI Research — [NeurIPS 2022] Transferring Pre-trained Multimodal Representations with Cross-modal Similarity Matching - Company: LG AI Research (lgresearch.ai) - Announced: 2023-02-08 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=301 - Record: https://forck.live/items/4909-neurips-2022-transferring-pre-trained-multimodal-representations-with-cross - Subject: EXAONE LG AI Research introduced BeamCLIP, a method for transferring representations from large pre-trained multimodal models (e.g., CLIP) to smaller models (e.g., ResNet-18), using Cross-modal Similarity Matching (CSM) and Context-based Prompt Augmentation (CPA), achieving 66.2% linear probe accuracy on ImageNet-1K. ### Hugging Face — Speech Synthesis, Recognition, and More With SpeechT5 - Company: Hugging Face (huggingface.co) - Announced: 2023-02-08 - Category: model-update - Coverage: not counted - Announcement: yes - Group: models - Source: https://huggingface.co/blog/speecht5 - Record: https://forck.live/items/2099-speech-synthesis-recognition-and-more-with-speecht5 - Subject: Platform Hugging Face announces that SpeechT5, a unified speech model from Microsoft Research Asia, is now available in the Transformers library, supporting text-to-speech, speech-to-text, and speech-to-speech tasks. ### Replicate — Introducing LoRA: A faster way to fine-tune Stable Diffusion - Company: Replicate (replicate.com) - Announced: 2023-02-07 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replicate.com/blog/lora-faster-fine-tuning-of-stable-diffusion - Record: https://forck.live/items/2576-introducing-lora-a-faster-way-to-fine-tune-stable-diffusion - Subject: Platform Replicate introduces LoRA, a faster method to fine-tune Stable Diffusion. LoRA trains in minutes, produces 5MB models, and allows combining multiple concepts. It is available now on Replicate. ### Hugging Face — Introducing ⚔️ AI vs. AI ⚔️ a deep reinforcement learning multi-agents competition system - Company: Hugging Face (huggingface.co) - Announced: 2023-02-07 - Category: developer-tool-release - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/aivsai - Record: https://forck.live/items/2101-introducing-ai-vs-ai-a-deep-reinforcement-learning-multi-agents-competition - Subject: Platform Hugging Face introduces AI vs. AI, an open-source deep reinforcement learning multi-agent competition system hosted on Spaces, using ELO rating for model ranking and leveraging Datasets and Leaderboard. ### Hugging Face — Generating Stories: AI for Game Development #5 - Company: Hugging Face (huggingface.co) - Announced: 2023-02-07 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/ml-for-games-5 - Record: https://forck.live/items/2100-generating-stories-ai-for-game-development-5 - Subject: Platform A tutorial blog post about using AI, specifically ChatGPT, for generating story content in game development, including process, limitations, and tips. ### Hugging Face — Accelerating PyTorch Transformers with Intel Sapphire Rapids - part 2 - Company: Hugging Face (huggingface.co) - Announced: 2023-02-06 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/intel-sapphire-rapids-inference - Record: https://forck.live/items/2102-accelerating-pytorch-transformers-with-intel-sapphire-rapids-part-2 - Subject: Platform This is a technical blog post about accelerating PyTorch Transformers inference on Intel Sapphire Rapids CPUs using the Optimum Intel library. It benchmarks models like distilbert-base-uncased, bert-base-uncased, and roberta-base on Ice Lake and Sapphire Rapids servers, measuring latency for short and long token sequences. It does not announce a new product, model, API, or any of the other specific categories.