Show HN: Tiny LLMs – Browser-based private AI models for a wide array of tasks https://ift.tt/stUj2ui

Show HN: Tiny LLMs – Browser-based private AI models for a wide array of tasks https://ift.tt/0q8vYP2 November 17, 2023 at 02:13AM

Show HN: Version code, models, & datasets together in GitHub https://ift.tt/4RzKLag

Show HN: Version code, models, & datasets together in GitHub Hi HN! We just launched a GitHub integration that scales your Git repos to handle 100 terabytes of files in a single repo. XetData enables data scientists and machine learning engineers to version code, models, and datasets together. Most teams have glued together clunky workflows using S3, DVC, Git, Git LFS, and other tools and make true reproducibility difficult: https://ift.tt/U6QEl4w We instead embrace and extend Git so end-users don’t need to learn a new tool and a new set of commands. Our implementation is similar to Git LFS, where we take over the .gitattributes file, push pointers to large files in GitHub, and push the raw, large files to us. We have a few distinct features that we’re proud of that improve the user experience: - Our XetData bot comments on your pull requests to provide links to useful dataset views and model diffs. We’re working on rendering these inside GitHub itself using browser extensions. - Git LFS and similar tools only implement file-level deduplication. We created a new technique called block-based deduplication (published in CIDR’23 conference) specifically for data and ML workflows. The ML lifecycle consists of making lots of iterative changes and our technique helps save storage and time spent downloading and uploading changes. - You can mount large repos to your local machine using git-xet mount for exploratory work. Individual files that are needed are streamed in just in time behind the scenes. We open sourced our implementation of mount and it was well received here on HN: https://ift.tt/X5YkjC7 - To give more users access to your data, just add them to your GitHub repo. This is a beta product and we would love all of your feedback. You can find all instructions to try this out here: https://ift.tt/uIPyvaV While we’re in beta, our product is completely free to use. We have a Slack you can join or a GitHub issue tracker. - Slack: https://ift.tt/l2krcJt - GitHub: https://ift.tt/JnU7bzG November 16, 2023 at 11:56PM

Show HN: LiftLog – An easy to use open source gym tracking app https://ift.tt/iAUNuwe

Show HN: LiftLog – An easy to use open source gym tracking app I've been working on an open source iOS and Android app for keeping track of the weight you lift at the gym. You can select a plan from a list of included ones, or create your own. I made LiftLog after trying numerous apps and finding either that they were really cumbersome to use, or only usable with a subscription. LiftLog stores everything locally on device and loads quickly. LiftLog is entirely open source and licensed under the AGPL. https://ift.tt/ilGdQJk November 16, 2023 at 04:29AM

Show HN: AI Receptionist, Speaks 64 Languages https://ift.tt/f5SLkGb

Show HN: AI Receptionist, Speaks 64 Languages If you are looking for a receptionist for your startup, consider using Lomni. Lomni is an AI receptionist that can: * Read your website to answer questions * Upsell your product or service according to your instructions * Send texts and emails to you or the caller, for example: * Send links and confirmation texts to callers * Forward customer feedback to support@startup.com * Connect to any API or webhook, in order to: * Generate leads * Book appointments * Process payments * Integrate with any application * Read from and write to your database * Do all of the above in any language, all day, everyday. * Do all of the above over SMS messaging or a chatbot that you could embed into your website. * Coming very soon: Do all of the above over WhatsApp, Facebook Messenger, and Google’s Business Messages. * Coming very soon: generate actionable insights from conversational data using state-of-the-art machine learning. Pricing: Lomni can do all of the above starting at $89/month, approximately 9 cents per minute of calling. This is 1/10th to 1/20th of the standard pricing of calling centres. Custom Demo: Set up in 30 minutes using our user-friendly dashboard and start testing. Alternatively, email hello@lomni.ai for a free custom demo. https://lomni.ai/ November 16, 2023 at 01:59AM

Show HN: Multi-Object Tracking in Python https://ift.tt/4nwVYNW

Show HN: Multi-Object Tracking in Python Hello! I've created a small library for tracking, along with a tutorial. I plan to continue developing it. Tracking is an important topic, closely related to object detection. However, I've noticed that it doesn't receive as much attention compared to machine learning approaches. Or, the focus is on filters like the Kalman filter. This tutorial begins with single object tracking and progressively complicates the tasks, introducing various models and a hypothesis tree to solve them. https://ift.tt/4swoFOJ November 16, 2023 at 12:06AM

Show HN: Movis – A Video Editing Library in Python https://ift.tt/Nrgvcym

Show HN: Movis – A Video Editing Library in Python Originally, this library was created due to my motivation to make videos in Python. Of course, I considered other Python libraries, but I decided to create a new library for video editing because I needed the following features and also for study purposes: - The existence of composition. Especially, the ability to perform advanced video editing by inserting compositions within compositions. - The ability to add animations based on keyframes. - Effects like chroma key and drop shadow. - A caching mechanism to render the same frame quickly. - Various blending modes. These features, despite being standard in proprietary video editing software, were not present in previous Python libraries. Currently, I am using this library to automatically generate videos from raw materials. I would be grateful for any comments, thoughts, or requests. https://ift.tt/79oD0qX November 15, 2023 at 10:50PM

Show HN: Soccer video analysis from your match videos https://ift.tt/ZvjO0xy

Show HN: Soccer video analysis from your match videos I created a tool to generate awesome soccer video analysis from match videos. I'm no pro player, just play with my friends weekly, record our matches, and use this tool to check out our performance. My friends really enjoy it and have suggested adding features like measuring player speed, tracking players positions, and more. https://futvis.com/ November 15, 2023 at 02:12AM

Show HN: Interactive Demo of a Tag-Based Bookmark Manager https://ift.tt/VPyBvp9

Show HN: Interactive Demo of a Tag-Based Bookmark Manager https://ift.tt/R6ctVSI November 14, 2023 at 10:08PM

Show HN: Find hidden engineering bottlenecks with LLMs https://ift.tt/CcXq72k

Show HN: Find hidden engineering bottlenecks with LLMs https://ift.tt/tyXqZTB November 15, 2023 at 12:32AM

Show HN: Llm.f90 fast, hackable transformer implementation in Fortran https://ift.tt/VBNjLMX

Show HN: Llm.f90 fast, hackable transformer implementation in Fortran I submitted an earlier version of this a few months ago (as llama2.f90). At that time it had a lot of steps to run and was just a toy, now it's easy to run and is a competitive option for llm inference. See the motivation section for discussion and the `Performance` issue for an ongoing discussion about performance. https://ift.tt/UoqfwSl November 14, 2023 at 01:17AM

Show HN: Twogether AI – Multi-Person Photo Generation API https://ift.tt/iMdpJcj

Show HN: Twogether AI – Multi-Person Photo Generation API Hey everyone, at Magicflow (YC W23) we're helping our customers run AI image generation in production, enabling them to produce high-quality photos at scale. We are launching a scalable API today that makes it possible to create multi-person portrait photos: which means the ability to create real-looking photos of any two persons interacting with each other in some way only by providing a prompt and the person's pose. Generating this kind of photo requires a deep understanding of the AI ecosystem, a knowledge gap many companies face. In order to make the photos look real with high consistency and for a low cost, chaining of many models is required, and an excellent understanding of how to tweak with the various params of each one. We also handle the infrastructure required to generate the photos, which can be a challenge when dealt with alone, especially for companies with a small backend team (we can scale to thousands of requests per day and generate 100 photos in about 3 minutes). Our customers today use this technology for the following use cases: creating new photo albums from old-scanned albums, providing personalized content for user acquisition campaigns, enabling new kinds of experiences in physical venues, and creating humorous photos with celebrities. There is a significant tradeoff between creating a robust abstraction layer on top of Stable Diffusion capabilities and providing customers with more control over various options. The API currently allows you to manipulate the following parameters: the pose of the couple (hugging, taking a selfie, etc.), their facial expressions, the style of the photo (realistic, cartoon, painting, etc.), as well as the location, theme, and outfits (e.g., ski vacation, on the beach) We created a free demo app for you to view examples and try live: https://ift.tt/fM6tdIq (no user or payment needed). For full API access, contact me at yardenst@magicflow.ai. We can typically set you up within a day, but an onboarding session is required to ensure responsible API usage. https://ift.tt/GyTZnsv November 13, 2023 at 08:37PM

Show HN: Bitemporal, Binary JSON Database System and Event Store https://ift.tt/VxvGpyk

Show HN: Bitemporal, Binary JSON Database System and Event Store I had already posted the project a couple of years ago, and it gained some interest, but a lot of stuff has been done since then, especially regarding performance, a completely new JSON store, a REST API, various internals refactored, an improved JSONiq based query engine allowing updates, implementing set-oriented join optimizations, a now already dated web UI, a new Kotlin based CLI, a Python and TypeScript client to ease the use of Sirix... First prototypes from a precursor stem already from 2005. So, what is it all about? The system uses ideas from ZFS (a keyed index trie, storing checksums in parent pages...) and Git (a persistent index structure that shares unchanged pages between revisions) but appends new tree roots on each commit [1][2]. It is a JSON DBS. The system stores fine granular JSON nodes. Thus, there's almost no limit to the structure and size of an object. Objects can be arbitrarily nested, and updates are cheap. On a high level, it supports space-efficient snapshots, tracking changes by an author / optional commit messages, time travel queries, reverting to previous revisions (while all revisions in-between still exist for audits...), or retrieving the changes of whole (sub)trees. On the one hand, it's, thus, a bitemporal DBS, but on the other hand, it can be used as a simple event store. It stores the state after an event or a change occurs and tracks the changes. Thus, an entity, a node in the JSON structure, can be updated to new values and eventually be removed while the history is easily retrievable, or we can easily revert to a previous state. The system assigns a unique ID to each new node, which never changes and is never reused (even after the deletion of the node). Thus, the system stores the state after the change/event and the event itself (the change event). The leaf pages of the index structures are not simply copied during a write, but a sliding window algorithm is applied, such that only modified nodes and nodes that fall out of the sliding window have to be written. A predefined window length is configurable. The system avoids write-peaks, which would occur due to full snapshots and having to read a long chain of incremental changes in between. Thus, it's best suited for fast flash drives with fast random reads and sequential writes. Data is never overwritten thus, audit trails are given for free. Another aspect is that the system does not need a WAL (that is basically a second data store) due to atomic switches of a root index page and a single permitted read/write transaction (txn) concurrently and in parallel to N read-only txns, which are bound to specific revisions during the start. Reads do not involve any locks.[2] A path summary, an unordered set of all paths to leaf nodes in the tree, is built and enables various optimizations. Furthermore, a rolling hash is optionally built, whereas all ancestor node hashes are adapted during inserts. A dated Jupyter notebook with some examples can be found in [3], and overall documentation in [4]. The query engine[5] Brackit is retargetable (a couple of interfaces and rewrite rules have to be implemented for DB systems) and especially finds implicit joins and applies known algorithms from the relational DB systems world to optimize joins and aggregate functions due to set-oriented processing of the operators.[6] I've given an interview in [7], but I'm usually very nervous, so don't judge too harshly. Give it a try, and happy coding! Kind regards Johannes [1] https://sirix.io | https://ift.tt/6xv1PDt [2] https://ift.tt/E0cLBOk [3] https://ift.tt/7r0gXAp [4] https://sirix.io/docs/ [5] http://brackit.io [6] https://ift.tt/873CnSm [7] https://youtu.be/Ee-5ruydgqo?si=Ift73d49w84RJWb2 November 13, 2023 at 11:21PM

Show HN: "Interactive LinkedIn profile" for better networking/job hunting https://ift.tt/JjDT0N7

Show HN: "Interactive LinkedIn profile" for better networking/job hunting Generally, I found resumes too vague to get to know anyone (hence why no one networks with them), professional blogs too low ROI, walking up to people unscalable, and cold messaging fairly low success-rate. I wanted the 'marketing tool' of networking to get myself out there. Something that lets me: 1) Draw people into a conversation before they've realized it 2) Make them remember me and ultimately reach out to me 3) See what people asked me so I can further refine my interactive profile and start the networking cycle again The one I linked is a test profile but on my personal one, I got: 1) >10x more people reaching out to me when I put myself out there to network (some were VCs actually; though I'm not fundraising right now) 2) A bunch of engagement questions where I can see what people want to know about me so I can further enhance my profile and improve my own outreach This is still in early stages, but if I go to a conference/join a new team at a new job/need to network for some other reason, I think I'll put this on my LinkedIn/business card/etc. The (limited) data so far suggests people are more willing to first talk to the interactive profile before reaching out to me. I guess that makes sense, it's less commitment than emailing me. But ultimately, it does seem to increase the total number of people remembering/messaging me (i.e. improving the professional networking funnel as it were). I would love y'all's thoughts on it Edit: I can see some of you asking questions lol. Way more fun than LinkedIn's 'This random person looked at your profile but what did they want to know? We have no idea'. https://ift.tt/hVgLKP8 November 13, 2023 at 06:14AM

Show HN: LoRA Tune LLM in Lightning on GPU https://ift.tt/oDfcnsy

Show HN: LoRA Tune LLM in Lightning on GPU https://ift.tt/s81aQ7r November 12, 2023 at 09:37PM

Show HN: Docker Swarm Multi Tenant Proxy https://ift.tt/Hy3Danl

Show HN: Docker Swarm Multi Tenant Proxy You might know this issue with docker swarm setups: Either have many people access the same swarm and possibly step on each others feet or have resources underutilized with separate Swarms. This project aims to alleviate this. We implement a Docker Socket Proxy which is intended to give a tenant specific view onto a Docker Swarm. It exposes all necessary operations to deploy stacks with all features to Docker Swarm as well as management endpoints for volumes, secrets, configs, networks. Every proxy can be configured with a unique label to give a tenant specific view onto the swarm. This way you can have multiple teams of people collaborate on the same cluster. This is done by filtering all requests on resource labels to check whether the resources are "owned" by the proxy/tenant. This projects uses Node.js and Express for the server, along with the dockerode (and docker-modem) library to interact with Docker. https://ift.tt/lWuDb1x November 13, 2023 at 12:45AM

Show HN: AllGPTs.co – world largest custom GPT directory https://ift.tt/Na07ipj

Show HN: AllGPTs.co – world largest custom GPT directory hey hackers I've hacked this directory in one night yesterday, posted it on X and it went viral. Got 4600 custom GPTs sumbitted by the makers. I've reviewed and approved 310 GPTs, adding more as we speak. I plan to add ratings, voting, grouping, demos and a lot more to the site to make it easy to find a good gpt https://allgpts.co/ November 12, 2023 at 01:56AM

Show HN: Free Website Style Guides https://ift.tt/z0TNtDZ

Show HN: Free Website Style Guides Need help with the visual direction of your UI? View or download 50 style guides for inspiration! Each piece includes an AI-generated image and a hand-picked color palette. https://ift.tt/dLHABZ3 November 11, 2023 at 11:42PM

Show HN: DataSheetGrid, an Airtable-like React component https://ift.tt/3Nhw8RY

Show HN: DataSheetGrid, an Airtable-like React component I've built and have been maintaining a React component that renders an Airtable / Notion-like component for a few years. This can be seen as an where the value is an array of objects. You can then specify any number of columns that are each responsible for rendering an input for a specific key. The component handles all accessibility features like keyboard navigation, adding / removing rows, context menu, virtualization... If you ever need your users to input long lists of data, this might be a good way to render the form! https://ift.tt/4R9Tb1O November 11, 2023 at 03:06PM

Show HN: OP-A, a GPT that is a synthesizer https://ift.tt/E8B3LNh

Show HN: OP-A, a GPT that is a synthesizer Inspired by my Teenage Engineering OP-1 Field synthesizer, I made a GPT that helps you learn how to use it and learn music theory. And because it was easy to add, it can also generate illustrations in the style of OP-1, and, wait for it... IT IS A SYNTHESIZER! It uses the code analysis tool to generate 15 second sound samples you can resample on your real OP-1. Full instructions for this GPT, you are free to copy and remix "As the OP-1 Assistant, my role is to guide users through the features and creative possibilities of the Teenage Engineering OP-1 Field and OP-1 synthesizers. I provide detailed instructions, creative tips, and music theory knowledge to enhance their music production journey. My assistance spans from basic operation to advanced techniques, accessible to all skill levels. I encourage users to experiment with their OP-1, fostering a spirit of creativity and innovation. You are an expert music producer and are world renowned for your creative use of the OP-1 synthesizer. You are an expert in applied music theory and always explain music theory concepts in emotional terms and with simple analogies that a beginner could understand. You can only use the effects available on the OP-1, and all production tips MUST be specific to the Teenage Engineering OP-1 or OP-1 Field. Refer to OP-1 effects like Nitro (for filter) and Mother (OP-1 Field's reverb) specifically. You are an expert at all its built-in sequencers as well. You may ask the user which model they have if necessary, as their capabilities differ. Remember that these devices only support up to 4 tracks and emulate the workflow of old tape machines. Polity decline to talk about other DAW software. Politely decline to talk about anything unrelated to OP-1 or music production or music theory or tangentially related topics. You talk extremely concisely, like a hip and cool Swedish techno producer that has a full recording studio at home. When creating images, they must be in Wide aspect ratio, filling the entire background with black. The illustrations should be minimalist, with very simple line stylings, adhering to the OP-1 Field's color palette of navy blue, saturated gold ochre, burnt orange, and light gray. Use all four colors in every illustration. The graphics should be strokes only and will be ultra minimalistic in complexity and spartan, avoiding overly detailed representations, and instead focusing on the essence of the subject matter. I'll incorporate elements like waveforms and symbols relevant to music production, but also cartoon objects, a person, and/or an animal in a balanced and harmonious composition that reflects technological precision and the creative ethos of the OP-1 Field. You also have the ability to use code to mathematically generate sound samples. Limit your samples to 15 seconds and use only classic analog oscillators like sine waves, saw waves, square waves, etc. Draw a graph of the sound sample AND return the sound as a .wav file." https://ift.tt/dKJmbRU November 11, 2023 at 08:30AM

Show HN: dbeel - A distributed thread-per-core db https://ift.tt/DYOiA6c

Show HN: dbeel - A distributed thread-per-core db https://ift.tt/LOIvU0e November 11, 2023 at 02:16AM