Search

Byte / code
Byte / code @byte_labox ·
時系列データの予測にTransformerを使うと、単純なRNNより精度が上がる PyTorchで時系列予測モデルを組む際、LSTMの代わりにTransformerブロック（nn.TransformerEncoder）を採用。例えば… dev.to/beck_moulton/t… #ByteLab #PyTorch #Transformer #機械学習 #時系列予測
Taming the Glucose Spike: Predicting Postprandial Peaks with Transformers and PyTorch

Living with a Continuous Glucose Monitor (CGM) is like having a dashboard for your metabolism. But...

From dev.to
3
odysseyzjh
odysseyzjh @zjhzxw ·
PyTorch 2.11带来了分布式训练优化、硬件专属算子支持相关更新。3月31日周二上午10点，Andrey Talman与Nikita Shulga将开启直播，带来版本更新详解和问答，内容涉及分布式训练可微分集合、FlexAttention相关进展。 #PyTorch #深度学习
26
odysseyzjh
odysseyzjh @zjhzxw ·
AI训练过程中的NCCL watchdog超时错误，是最容易被误解、也最难调试的常见问题之一。此前PyTorch曾推出Flight Recorder工具简化这类错误的调试，但它的输出内容一直很难解读。 #PyTorch #AI训练
5
Red Hat Community
Red Hat Community @redhatopen ·
How does a Python tensor op become a lightning-fast C++ kernel? Christopher Leonard traces the complete PyTorch call stack. Get ready for #PyTorch Conference Europe with a deep dive! red.ht/47sp9vR
PyTorch Call Stack Deep Dive: Tracing Tensor Operations from Python to C++ Kernels

Explore the PyTorch call stack, detailing how high-level Python tensor operations are converted, dispatched, and executed by highly optimized C++ kernels for peak performance.

From next.redhat.com
238
实用干货铺
实用干货铺 @DiannRudy87765 ·
🔥ML面试刷题神器！TorchCode开源来袭 PyTorch版LeetCode，39道顶尖AI公司真实面试题 从零实现ReLU/softmax/Transformer/GPT-2自动评分+梯度校验 即时反馈+提示+参考解答 无GPU/注册/云端依赖 MIT开源免费，HuggingFace可直接体验 github.com/duoan/TorchCode9t #ML面试 #PyTorch #开源工具 #AI编zc
4
87
نادي كتاب الذكاء الاصطناعي (بإدارة نبساي)
نادي كتاب الذكاء الاصطناعي (بإدارة نبساي) @aideas_club ·
📘 التعلم العميق باستخدام PyTorch ابدأ رحلتك في الذكاء الاصطناعي ببناء نماذج حقيقية 🤖 من الشبكات العصبية إلى Transformers والتطبيقات العملية باستخدام PyTorch. 🚀 مناسب لمبرمجي Python #AI #PyTorch #DeepLearning #Aideas_club #نبساي #تمكين_AI #نادي_كتاب_الذكاء_الاصطناB61
6
Zane Chen
Zane Chen @chenzeling4 ·
LeetCode, but for PyTorch. 🔥 40 exercises to master softmax, attention, GPT-2 from scratch. Instant auto-grading in Jupyter. No GPU needed. The exact skills Meta, DeepMind, and OpenAI test for in interviews. ⭐ 2.7K #MachineLearning #PyTorchS
1
19
GPU Rental
GPU Rental @GpuRental24741 ·
Fine-tuning a model with 32GB VRAM vs 8GB: 8GB: constant OOM errors, gradient checkpointing, batch size 1 32GB: full precision, large batch, done in one run The RTX 5090 just removes the ceiling. gpu-rental.io — from $7/hr #MachineLearning #PyTorch #RTX5090
32
Brian
Brian @Brjen ·
Replying to @Brjen
The fix: export MALLOC_MMAP_THRESHOLD_=65536 export MALLOC_TRIM_THRESHOLD_=65536 Full writeup + benchmark: github.com/brjen/pytorch-… #PyTorch #MachineLearning #MLOps #Linux #AI #LLM #StableDiffusion #OOM
GitHub - brjen/pytorch-memory-fix: Two environment variables that fix PyTorch/glibc memory creep on...

Two environment variables that fix PyTorch/glibc memory creep on Linux. Zero code changes. Zero performance cost. - brjen/pytorch-memory-fix

From github.com
43