amazon-science / mxfp4-llmLinks
Official implementation for Training LLMs with MXFP4
☆116Updated 8 months ago
Alternatives and similar repositories for mxfp4-llm
Users that are interested in mxfp4-llm are comparing it to the libraries listed below
Sorting:
- ☆157Updated 10 months ago
- ☆114Updated this week
- Boosting 4-bit inference kernels with 2:4 Sparsity☆90Updated last year
- The evaluation framework for training-free sparse attention in LLMs☆108Updated 2 months ago
- Work in progress.☆76Updated last month
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆135Updated last year
- ☆162Updated 6 months ago
- Flash-Muon: An Efficient Implementation of Muon Optimizer☆225Updated 6 months ago
- ☆133Updated 7 months ago
- Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of…☆146Updated last year
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆131Updated last year
- Fast and memory-efficient exact attention☆75Updated 10 months ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆175Updated last year
- Explore training for quantized models☆25Updated 5 months ago
- PB-LLM: Partially Binarized Large Language Models☆157Updated 2 years ago
- KV cache compression for high-throughput LLM inference☆148Updated 11 months ago
- Repo for "LoLCATs: On Low-Rank Linearizing of Large Language Models"☆249Updated 11 months ago
- Code for data-aware compression of DeepSeek models☆66Updated 3 weeks ago
- ShiftAddLLM: Accelerating Pretrained LLMs via Post-Training Multiplication-Less Reparameterization☆111Updated last year
- Code for studying the super weight in LLM☆121Updated last year
- QuIP quantization☆61Updated last year
- [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization☆396Updated last year
- ☆54Updated last year
- Load compute kernels from the Hub☆357Updated 3 weeks ago
- HALO: Hadamard-Assisted Low-Precision Optimization and Training method for finetuning LLMs. 🚀 The official implementation of https://arx…☆29Updated 10 months ago
- ☆127Updated last year
- Triton-based implementation of Sparse Mixture of Experts.☆259Updated 3 months ago
- Training-free Post-training Efficient Sub-quadratic Complexity Attention. Implemented with OpenAI Triton.☆148Updated 2 months ago
- 🚀 Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flash…☆277Updated last month
- [ICLR'25] Fast Inference of MoE Models with CPU-GPU Orchestration☆253Updated last year