Alternatives
Products that do what I embedded 685M public texts in 32 minutes (on 8x A100, Rust, TensorRT) does
Quick note on how it works and how I've done my batch embedding engine IgniteMS. The whole thing runs as one process using Rust, reading input, tokenizing, packing batches, keeping the queue full. TensorRT handles inference. Python is only as a wrapper. I built it this way because when you use more than couple of GPUs, the GPUs stop being the problem. CPU cannot feed them fast enough. One A100 can go through batches faster than Python can tokenize and feed, so the GPU just sits there idle waiting for work. Most of my time went into optimizing this. At 8 GPUs that was basically the entire…
- 1LA
I built LocalGPT over 4 nights as a Rust reimagining of the OpenClaw assistant pattern (markdown-based persistent memory, autonomous heartbeat tasks, skills system). It compiles to a single ~27MB binary — no Node.js, Docker, or Python required. Key features: - Persistent memory via markdown files (MEMORY, HEARTBEAT, SOUL markdown files) — compatible with OpenClaw's format - Full-text search (SQLite FTS5) + semantic search (local embeddings, no API key needed) - Autonomous heartbeat runner that checks tasks on a configurable interval - CLI + web interface + desktop GUI - Multi-provider:…
Feb 2026 · github.com
- 2SM
I wanted to see how fast an isolated code sandbox could start if I never had to boot a fresh VM. So instead of launching a new microVM per execution, I boot Firecracker once with Python and numpy already loaded, then snapshot the full VM state. Every execution after that creates a new KVM VM backed by a `MAP_PRIVATE` mapping of the snapshot memory, so Linux gives me copy-on-write pages automatically. That means each sandbox starts from an already-running Python process inside a real VM, runs the code, and exits. These are real KVM VMs, not containers: separate guest kernel, separate guest…
Mar 2026 · github.com
- 3SU
Here's a project I've been working on for the last few months. It's a new (I think) algorithm, that allows to adjust smoothly - and in real time - how many calculations you'd like to do during inference of an LLM model. It seems that it's possible to do just 20-25% of weight multiplications instead of all of them, and still get good inference results. I implemented it to run on M1/M2/M3 GPU. The mmul approximation itself can be pushed to run 2x fast before the quality of output collapses. The inference speed is just a bit faster than Llama.cpp's, because the rest of implementation…
2024 · asciinema.org
- 4TC
Hello HN! I’m Jonathan from TensorDock. After 7 months in beta, we’re finally launching Core Cloud, our platform to deploy GPU virtual machines in as little as 45 seconds! https://www.tensordock.com/product-core Why? Training machine learning workloads at large clouds can be extremely expensive. This left us wondering, “how did cloud ever become more expensive than on-prem?” I’ve seen too many ML startups buy their own hardware. Cheaper dedicated servers with NVIDIA GPUs are not too hard to find, but they lack the functionality and scalability of the big clouds. We thought to…
2022 · tensordock.com
- 5
- 6
- 7IB
Hi HN, Over the past few months, I've been building `dsc`, a tensor library from scratch in C++/CUDA. My main focus has been on getting the basics right, prioritizing a clean API, simplicity, and clear observability for running small LLMs locally. The key features are: - C++ core with CUDA support written from scratch. - A familiar, PyTorch-like Python API. - Runs real models: it's complete enough to load a model like Qwen from HuggingFace and run inference on both CUDA and CPU with a single line change[1]. - Simple, built-in observability for both Python and C++. Next on the roadmap is…
2025 · github.com
- 8RQ
Sep 2025 · github.com
- 9PA
I’m sure many of you are familiar, but there’s a treacherous gap between finding (or building) a model that works in PyTorch, and getting that deployed into your application, especially in consumer-facing applications. I’ve been very interested in solving this problem with a great developer experience. Over time, I gradually realized that the highest-impact thing to have was a way to go from existing Python code to a self-contained native binary—in other words, a Python compiler. I was already pretty familiar with a successful attempt: when Apple introduced armv8 on the iPhone 5s, they…
2025 · blog.fxn.ai
- 10
- 11RG
I wanted to know how fast a 26B mixture-of-experts model could run on a desktop CPU with no GPU. Got ~40 tok/s single-stream (lossless) and ~124 batched. The surprising part was the byte budget: for this model you compress the output head (32% of per-token bytes), not the experts (16%). The writeup has the bandwidth roofline and the dead-ends; the repo has the reproducible recipe. Happy to answer questions. Repo: https://github.com/arun-prasath2005/gemma4-cpu-moe
Jun 2026 · apeg.dev
- 12OL
I've been working on Fast LiteLLM - a Rust acceleration layer for the popular LiteLLM library - and I had some interesting learnings that might resonate with other developers trying to squeeze performance out of existing systems. My assumption was that LiteLLM, being a Python library, would have plenty of low-hanging fruit for optimization. I set out to create a Rust layer using PyO3 to accelerate the performance-critical parts: token counting, routing, rate limiting, and connection pooling. The Approach - Built Rust implementations for token counting using tiktoken-rs - Added lock-free data…
Nov 2025 · github.com
- 13MR
Hey HN! We’ve just open-sourced model2vec-rs, a Rust crate for loading and running Model2Vec static embedding models with zero Python dependency. This allows you to embed text at (very) high throughput; for example, in a Rust-based microservice or CLI tool. This can be used for semantic search, retrieval, RAG, or any other text embedding usecase. Main Features: - Rust-native inference: Load any Model2Vec model from Hugging Face or your local path with StaticModel::from_pretrained(...). - Tiny footprint: The crate itself is only ~1.7 mb, with embedding models between 7 and 30 mb. Performance:…
2025 · github.com
- 14SS
We'd like to introduce HN to Spell, which is a tool for easily running ML/DL jobs remotely. As Deep Learning has grown we see engineers and researchers struggle to incorporate running on GPUs into their workflow. So we built Spell to be the easiest way to get code running elsewhere - like the bash '&' operator but for remote machines. Sign up for an account at https://web.spell.run/waitlist, which includes $300 in credits for GPU time. There's a waitlist, but we'll be approving accounts as they come in. Here are some of the features we really wanted and built into Spell:…
2018
- 15IB
I’ve spent the last few months building a deep learning engine completely from scratch in Python (using only math and random). What started as a basic linear algebra calculator project grew into a symbolic tensor system with autodiff, custom matrix ops, attention mechanisms, LayerNorm, GELU, and even a text generation demo trained on the Brown corpus. I'm still an undergrad, so my main goal is to deeply understand how deep learning actually works under the hood - gradients, attention, backpropagation, optimizers - by building it step-by-step with full visibility into everything, and without…
2025 · github.com
- 16OI
Nov 2025 · github.com
- 17T5
2021 · tensorbase.io
- 18FS
I want to share a really dumb, but very practical project I have packaged this summer, to perform operations on strings much faster. I was using Python to work with a multi-terabyte newline-delimited file. Reading, splitting, and shuffling it was a nightmare. So, I wrapped a trivial hardware-friendly heuristic I've been using for the last few years into a CPython library. The part I enjoyed the most is implementing SIMD behavior without SIMD instructions... Using 64-bit words to work at 8-bit granularity. Unlike conventional SIMD, the code would remain the same for ~~almost~~ any hardware.…
2023 · ashvardanian.com
- 19ZA
This is a low-level opensource library I developed for my own use and decided to share, as it makes it possible to process large checkpoints of neural networks without renting high-RAM instances, on a regular PC. It replaces torch.load() with a custom function that produces a dictionary that materializes tensors on the fly. Compared to other solutions it doesn't require sharding or re-encoding checkpoints and uses them completely as-is. It is a foundation to make it possible to run inference and compress language models and other large models one layer at a time - in principle, even one…
2023 · github.com
- 20RA
We built RapidFire AI, an open-source Python tool to speed up LLM fine-tuning and post-training with a powerful level of control not found in most tools: Stop, resume, clone-modify and warm-start configs on the fly—so you can branch experiments while they’re running instead of starting from scratch or running one after another. - Works within your OSS stack: PyTorch, HuggingFace TRL/PEFT), MLflow. - Hyperparallel search: launch as many configs as you want together, even on a single GPU - Dynamic real-time control: stop laggards, resume them later to revisit, branch promising configs in…
Sep 2025 · github.com
- 21IM
My cofounder and I run a startup in oncology, where we handle cancer genomics data. It occurred to me that, thanks to a recent complexity theory result, there's a clever way to run bioinformatics algorithms using far less RAM. I built this Rust engine for running whole-genome workloads in under 100MB of RAM. Runtime is a little longer as a result - O(TlogT) instead of O(T). But it should enable whole-genome analytics on consumer-grade hardware.
Nov 2025 · github.com
- 22LT
I wanted to share a project I've been working on for the past few weeks: llgtrt. It's a Rust implementation of a HTTP REST server for hosting Large Language Models using llguidance library for constrained output with NVIDIA TensorRT-LLM. The server is compatible with the OpenAI REST API and supports structured JSON schema enforcement as well as full context-free grammars (via Guidance). It's similar in spirit to the Python-based TensorRT-LLM OpenAI server example but written entirely in Rust and built with constraints in mind. No Triton Inference Server involved. This also serves as a demo…
2024 · github.com
- 23MM
Hi HN! We (Thomas and Stéphan, hello!) recently released Model2Vec, a Python library for distilling any sentence transformer into a small set of static embeddings. This makes inference with such a model up to 500x faster, and reduces model size by a factor of 15 (7.5M params or 15/30MB on disk, depending on whether you use float16 or float32). This allows you to embed 50-100k documents per second on a cpu on a macbook. This reduction of course comes at a cost: distilled models are worse than their parent models. Even so, they are actually a lot better than large sets of conventional…
2024 · github.com
- 24FH
Hi, This is Dan and Genevieve from Burstable AI. We've iterated and made a 45 degree pivot, taking what we learned from developing burst (https://news.ycombinator.com/item?id=28191459) to introduce a cloud service that provides access to a GPU-enabled machine using Jupyterlab to provide notebooks, shell access, and a code/text editor. GPU access is measured and the first 50 hours are free. This is *not* a platform to do crypto mining or run weeks of model training for free. We are focused on the R & D phase of modern AI/ML, where developers/scientists are…
2022 · cloudburst.host
Ranked by how close each launch is in meaning, then by votes. Refine with a description →