Why Low-Precision Transformer Training Fails: An Analysis on Flash Attention
☆68Apr 7, 2026Updated last month
Alternatives and similar repositories for why-low-precision-training-fails
Users that are interested in why-low-precision-training-fails are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Official implementation for the paper "Controlled Sparsity via Constrained Optimization"☆12Aug 10, 2022Updated 3 years ago
- ☆49May 20, 2025Updated last year
- ☆15Jul 25, 2024Updated last year
- ☆34Sep 19, 2025Updated 8 months ago
- ☆17May 14, 2020Updated 6 years ago
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- codes and plots for "Active-Dormant Attention Heads: Mechanistically Demystifying Extreme-Token Phenomena in LLMs"☆11Dec 30, 2024Updated last year
- Read, modify and write DICOS files with python code☆13Nov 24, 2025Updated 6 months ago
- Supporting code for "LLMs for your iPhone: Whole-Tensor 4 Bit Quantization"☆11Mar 31, 2024Updated 2 years ago
- Code for the paper "Interpreting and Improving Diffusion Models from an Optimization Perspective", appearing in ICML 2024☆15Sep 30, 2024Updated last year
- ☆13Nov 27, 2025Updated 6 months ago
- ☆17Apr 30, 2025Updated last year
- Code and dataset for the EMNLP 2024 paper: GoldCoin: Grounding Large Language Models in Privacy Laws via Contextual Integrity Theory☆49Sep 26, 2024Updated last year
- JsonTuning: Towards Generalizable, Robust, and Controllable Instruction Tuning☆10Nov 3, 2024Updated last year
- ☆11Apr 5, 2023Updated 3 years ago
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- ☆16Apr 26, 2023Updated 3 years ago
- [CVPR 2025 Highlight] FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation☆28Jun 16, 2025Updated 11 months ago
- SMART introduces a novel test-time framework where Small Language Models (SLMs) reason step-by-step, and Large Language Models (LLMs) pro…☆12Jul 9, 2025Updated 10 months ago
- [CVPR 2025] QuartDepth☆18Mar 24, 2025Updated last year
- AiTer Optimized Model☆93Updated this week
- ☆15Apr 6, 2026Updated last month
- Train small sequence models in your browser with WebGPU.☆34Dec 3, 2025Updated 5 months ago
- A fast, simple, multi-threaded string interning library.☆18Jul 11, 2025Updated 10 months ago
- A compiler of Decaf(an object-oriented compiler)☆12Sep 26, 2017Updated 8 years ago
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- The ICDAR2015-TextSR dataset. This dataset was originally presented for the ICDAR2015 Competition on Text Image Super-Resolution. The rel…☆15Oct 8, 2019Updated 6 years ago
- Source code of our TNNLS paper "Boosting Convolutional Neural Networks with Middle Spectrum Grouped Convolution"☆12Apr 14, 2023Updated 3 years ago
- ☆16Dec 9, 2023Updated 2 years ago
- Lightweight arXiv literature digest skill for OpenClaw — Zotero-driven interest profiling, 3-dimensional candidate ranking, abstract-firs…☆44Mar 24, 2026Updated 2 months ago
- ☆14Jul 14, 2025Updated 10 months ago
- [ICML 2025] Official PyTorch implementation of "NegMerge: Sign-Consensual Weight Merging for Machine Unlearning"☆16Nov 25, 2025Updated 6 months ago
- ☆15Apr 11, 2024Updated 2 years ago
- LLM Inference with Microscaling Format☆34Nov 12, 2024Updated last year
- Try to export the ONNX QDQ model that conforms to the AXERA NPU quantization specification. Currently, only w8a8 is supported.☆11Sep 10, 2024Updated last year
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Information and artifacts for "LoRA Learns Less and Forgets Less" (TMLR, 2024)☆21Sep 27, 2024Updated last year
- ☆10Apr 24, 2024Updated 2 years ago
- FlexAttention w/ FlashAttention3 Support☆27Oct 5, 2024Updated last year
- CUDA keyring packaging for Debian☆14Apr 14, 2023Updated 3 years ago
- Alias mutliple derives as one.☆11Nov 30, 2024Updated last year
- ☆17Mar 10, 2025Updated last year
- The implementation for FREE-Merging: Fourier Transform for Model Merging with Lightweight Experts (ICCV25)☆15Jun 26, 2025Updated 11 months ago