torchtune: Easily fine-tune LLMs using PyTorch
by Team PyTorch We’re pleased to announce the alpha release of torchtune, a PyTorch-native library for easily fine-tuning large language models. Staying true to PyTorch’s design principles, torchtune provides composable and modular building blocks along with easy-to-extend training recipes to fine-tune popular LLMs on a variety of consumer-grade and professional GPUs. torchtune supports the full f
Quanto: a PyTorch quantization backend for Optimum
Quantization is a technique to reduce the computational and memory costs of evaluating Deep Learning Models by representing their weights and activations with low-precision data types like 8-bit integer (int8) instead of the usual 32-bit floating point (float32). Reducing the number of bits means the resulting model requires less memory storage, which is crucial for deploying Large Language Models
Pytorchã«ã‚ˆã‚‹LLMã®é«˜é€ŸåŒ–
アドベントカレンダー「ã»ã¼æ¨ªæµœã®æ°‘ã€ã®11日目ã®è¨˜äº‹ã§ã™ã€‚ 今年㯠LLM ã®é«˜é€ŸåŒ–実装ã«ã¤ã„ã¦æ›¸ã„ã¦ã„ã¾ã™ã€‚ç§ã¯LLMã®å°‚門家ã§ã¯ãªã„ã§ã™ãŒå‰ã€…ã‹ã‚‰èˆˆå‘³ãŒã‚ã£ãŸã®ã§å°‘ã—勉強ã—ã¦ã¿ã¾ã—ãŸã€‚ ã“ã®è¨˜äº‹ã‚’èªã‚“ã§ã‚ã‹ã‚‹ã“㨠LLMãŒæ–‡ç« を生æˆã™ã‚‹ä»•çµ„ã¿ torch.compile ã«ã‚ˆã£ã¦ LLM ã¯ã©ã®ã‚ˆã†ã«é«˜é€ŸåŒ–ã•ã‚Œã‚‹ã®ã‹ï¼Ÿ Speculative Decoding ã¨ã¯ï¼Ÿ 背景 å°‘ã—å‰ã« Accelerating Generative AI with Pytorch II: GPT, Fast ã¨ã„ã†ç´ 晴らã—ã„ブãƒã‚°è¨˜äº‹ã‚’見ã‹ã‘ã¾ã—ãŸã€‚ã“ã®è¨˜äº‹ã¯ Pytorch ãƒãƒ¼ãƒ ã‹ã‚‰å‡ºã•ã‚ŒãŸã‚‚ã®ã§ã€ç´ ã® Pytorch ã®ã¿ã‚’用ã„㦠LLM ã®æŽ¨è«–ã‚’ 10 å€é«˜é€ŸåŒ–ã§ãã‚‹ã¨ã„ã†ã‚‚ã®ã§ã—ãŸã€‚一体ã©ã®ã‚ˆã†ã« 10 å€ã‚‚ã®é«˜é€ŸåŒ–を実ç¾ã—ã¦ã„ã‚‹ã®ã‹æ°—ã«ãªã£ãŸã®ã§ã€å€‹äººçš„ãªå‹‰å¼·ã‚‚å…¼ãã¦ã“ã®è¨˜äº‹ã‚’書ã„ã¦ã„ã¾ã™ã€‚
Reproducibility — PyTorch 2.4 documentation
Reproducibility¶ Completely reproducible results are not guaranteed across PyTorch releases, individual commits, or different platforms. Furthermore, results may not be reproducible between CPU and GPU executions, even when using identical seeds. However, there are some steps you can take to limit the number of sources of nondeterministic behavior for a specific platform, device, and PyTorch relea
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
Pytorch Conference
GitHub - Lightning-AI/LitServe: Lightning-fast serving engine for AI models. Flexible. Easy. Enterprise-scale.
GitHub - pytorch/torchchat: Run PyTorch LLMs locally on servers, desktop and mobile
Reducing Model Checkpointing Times by Over 10x with PyTorch Distributed Asynchronous Checkpointing
Ecosystem compatibility with numpy 2.0 · Issue #26191 · numpy/numpy
GitHub - pytorch/torcheval: A library that contains a rich collection of performant PyTorch model metrics, a simple interface to create new metrics, a toolkit to facilitate metric computation in distributed training and tools for PyTorch model evaluations
ExecuTorch Alpha: Taking LLMs and AI to the Edge with Our Community and Partners
Torch compile does not work on python 3.12 · Issue #120233 · pytorch/pytorch
pytorch/torch/distributed/checkpoint/examples/async_checkpointing_example.py at release/2.3 · pytorch/pytorch
Release PyTorch 2.3: User-Defined Triton Kernels in torch.compile, Tensor Parallelism in Distributed · pytorch/pytorch
pytorch-distributed-collective-communication/main.py at main · FrancescoSaverioZuppichini/pytorch-distributed-collective-communication
Google Colab
DeepSpeedExamples/training/imagenet/main.py at master · microsoft/DeepSpeedExamples
PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation
GitHub - pytorch/tensordict: TensorDict is a pytorch dedicated tensor container.
TensorDict
{{#tags}}- {{label}}
{{/tags}}