Alternatives
Products that do what Taliesin – bit-exact KV-cache restore, 21x faster, cross-GPU verified does
- 1IV
The video demo runs a 7b Model on a normal gaming GPU. I think it already works quite well (accounting for the limited hardware power). :)
2024 · github.com
- 28F
Hi HN! I'm just sharing a project I've been working on during the LLM Efficiency Challenge - you can now finetune Llama with QLoRA 5x faster than Huggingface's original implementation on your own local GPU. Some highlights: 1. Manual autograd engine - hand derived backprop steps. 2. QLoRA / LoRA 80% faster, 50% less memory. 3. All kernels written in OpenAI's Triton language. 4. 0% loss in accuracy - no approximation methods - all exact. 5. No change of hardware necessary. Supports NVIDIA GPUs since 2018+. CUDA 7.5+. 6. Flash Attention support via Xformers. 7. Supports 4bit and 16bit…
2023 · github.com
- 3AG
This is a vector index I built that supports insertion and k-nearest neighbors (k-NN) querying, optimized for GPUs. It operates entirely in CUDA and can process queries on half a billion vectors in under 200 milliseconds. The codebase is structured as a standalone library with an HTTP API for remote access. It’s intended for high-performance search tasks—think similarity search, AI model retrieval, or reinforcement learning replay buffers. The codebase is located at https://github.com/rodlaf/BinaryGPUIndex.
2025 · rlafuente.com
- 4RA
This started out as a personal effort to learn more about machine learning. It's currently a CLI app where you give it a JSON file specifying your network architecture and hyperparameters and point it to your training data, then invoke it again in 'eval' mode with some data it's not seen before and it will try to classify each sample. I don't see many other people using Vulkan for GPGPU, and there may be many good reasons for that, but I wanted to try something a bit different. I've made every attempt to make the code very clean and readable and I've written up the math in…
2024 · github.com
- 5MA
Superficially, it may look similar to Nvidia's Warp or OpenAI's Triton, but instead of transpiling a subset of Python to the target language, it implements dynamic codegen in pure Python. These slides discuss the existing GPU programming solutions and make a case for Metashade's approach: https://docs.google.com/presentation/d/e/2PACX-1vQtYIwXIkMnV...
2024 · github.com
- 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
- 8CA
I built Chonkie because I was tired of rewriting chunking code for RAG applications. Existing libraries were either too bloated (80MB+) or too basic, with no middle ground. Core features: - 21MB default install vs 80-171MB alternatives - 33x faster token chunking than popular alternatives - Supports multiple chunking strategies: token, word, sentence, and semantic - Works with all major tokenizers (transformers, tokenizers, tiktoken) - Zero external dependencies for basic functionality Technical optimizations: - Uses tiktoken with multi-threading for faster tokenization - Implements…
2024 · github.com
- 9CR
Jun 2026 · github.com
- 10FG
Sep 2025 · github.com
- 11CF
Oct 2025 · github.com
- 12AS
2020 · nosferalatu.com
- 13CA
I am the author of the platform, happy to reply to any question you might have! It scales up really nicely thanks to a year of research and development of the hashtable implemented in cachegrand, on the hardware used for benchmarking, an AMD EPYC 7502P, it was able to reach up to 5mln GET QPS and 4.5mln SET QPS, with batching up to 60mln GET QPS and up to 26MLN SET QPS. cachegrand is fast, it's fully Open Source, it's under a BSD 3-clause license - it can be used easily as standalone platform or incorporated in other ones without any licensing issue - and we are working to expand the Redis…
2022 · github.com
- 14
We recently used DeepSeek V4 Flash as a teacher for finance tasks with GPT-OSS-120B. Distillation works well on this problem. At a constrained 8k token budget, our self-distilled 120B scores 83.61% on FinanceReasoning, above Kimi K3 (81.93%) and Inkling (65.13%). We released the 20B open weights. With V4 as the teacher though, we realized it would be timely to measure if the censorship characteristic of it transferred to the distilled version of the base model. tl;dr it didn't, the teacher answered politically sensitive questions 7 SDs differently than expected, but the distilled model's…
Jul 2026 · ctgt.ai
- 15
- 16UA
The standard GPU utilization metric reported by nvidia-smi, nvtop, Weights & Biases, Amazon CloudWatch, Google Cloud Monitoring, and Azure Monitor is highly misleading. It reports the fraction of time that any kernel is running on the GPU, which means a GPU can report 100% utilization even if only a small portion of its compute capacity is actually being used. In practice, we've seen workloads with ~1–10% real compute throughput while dashboards show 100%. This becomes a problem when teams rely on that metric for capacity planning or optimization decisions, it can make underutilized systems…
Apr 2026 · systalyze.com
- 17KY
2025 · github.com
- 18CL
Hey HN, we’re the developers of OpenLake, an open source storage engine for offloading LLM KV caches from GPU memory into a shared tier of RAM and NVMe. We built OpenLake because KV caches are outgrowing GPU memory. A single 256K token conversation on Gemma 4 31B produces approximately 43GB of KV state, more than half the memory of an 80GB H100. The problem becomes even harder across a cluster: a prefix cached on one GPU host is unavailable when the next request lands on a different GPU, forcing the new GPU to repeat work the fleet has already completed. Once the KV cache is offloaded,…
Jul 2026 · github.com
- 19PO
Our company Vertex.AI has been working on this for a while but this is the first public release. We're starting with using PlaidML to bring OpenCL support to Keras and more frameworks, platforms, etc are coming. Yes, this means you can use use your AMD GPU for deep learning dev. Sorry, no Mac or Windows support yet although the brave can try building from source (it should work). http://vertex.ai/blog/announcing-plaidml https://github.com/plaidml/plaidml
2017
- 20KC
May 2026 · pythongiant.github.io
- 21ZA
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
- 22
- 23FC
2024 · determinate.systems
- 24LF
We're excited to announce that we've open-sourced LeanRL, a lightweight PyTorch reinforcement learning library that provides recipes for fast RL training using torch.compile and CUDA graphs. By leveraging these tools, we've achieved significant speed-ups compared to the original CleanRL implementations - up to 6x faster! Reinforcement learning is notoriously CPU-bound due to the high frequency of small CPU operations. PyTorch's powerful compiler can help alleviate these issues, but comes with its own costs. LeanRL addresses this challenge by providing simple recipes to accelerate your…
2024 · github.com
Ranked by how close each launch is in meaning, then by votes. Refine with a description →