244 skills · sorted by relevance
Use when writing or changing PyTorch code — model, Dataset/DataLoader, training and evaluation loops, determinism, mixed precision, torch.compile, checkpoints with safetensors, Hugging Face models, ONNX export, and tests that run on CPU
The English summary is being prepared. Showing the original description.
PyTorch深度学习模式与最佳实践，用于构建稳健、高效且可复现的训练流程、模型架构和数据加载。
The English summary is being prepared. Showing the original description.
Use when diagnosing NPU + PyTorch or LLM Serving Host CPU affinity, NUMA locality, cgroup/cpuset constraints, CPU range conflicts, PyTorch/runtime threading, DataLoader, tokenizer, scheduler, vLLM-Ascend, SGLang, TTFT, TPOT, tokens/s, QPS, or multi-rank/multi-worker CPU binding issues.
The English summary is being prepared. Showing the original description.
Fornece orientação para treinamento RL agentico nativo do PyTorch usando torchforge, biblioteca da Meta que separa infraestrutura de algoritmos. Use quando você quer abstrações limpas de RL, fácil experimentação de algoritmos, ou treinamento escalável com Monarch e TorchTitan.
The English summary is being prepared. Showing the original description.
Otimiza atenção em transformers com Flash Attention para ganho de 2-4x em velocidade e redução de 10-20x em memória. Use ao treinar/executar transformers com sequências longas (>512 tokens), ao encontrar problemas de memória GPU com atenção, ou quando precisa de inferência mais rápida. Suporta SDPA nativo do PyTorch, biblioteca flash-attn, H100 FP8 e sliding window attention.
The English summary is being prepared. Showing the original description.
Orquestração de treinamento distribuído em clusters. Escala PyTorch/TensorFlow/HuggingFace do laptop para milhares de nós. Ajuste de hiperparâmetros integrado com Ray Tune, tolerância a falhas, escalabilidade elástica. Use ao treinar modelos massivos em múltiplas máquinas ou executar varreduras distribuídas de hiperparâmetros.
The English summary is being prepared. Showing the original description.
Processamento escalável de dados para workloads de ML. Execução em streaming em CPU/GPU, suporta Parquet/CSV/JSON/imagens. Integra-se com Ray Train, PyTorch, TensorFlow. Escala de uma única máquina para centenas de nós. Use para inferência em lote, pré-processamento de dados, carregamento de dados multi-modal ou pipelines distribuídos de ETL.
The English summary is being prepared. Showing the original description.
API de treinamento distribuído mais simples. 4 linhas para adicionar suporte distribuído a qualquer script PyTorch. API unificada para DeepSpeed/FSDP/Megatron/DDP. Posicionamento automático de device, precisão mista (FP16/BF16/FP8). Config interativo, comando de launch único. Padrão do ecossistema HuggingFace.
The English summary is being prepared. Showing the original description.
PyTorch deep learning development with transformers, diffusion models, and GPU optimization.
The English summary is being prepared. Showing the original description.
Expert guidance for deep learning, transformers, diffusion models, and LLM development with PyTorch, Transformers, Diffusers, and Gradio.
The English summary is being prepared. Showing the original description.
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.
The English summary is being prepared. Showing the original description.
Patterns for loading PDE simulation datasets (PDEBench, PhiFlow, JAX-CFD) from HDF5 files. Handles layout detection (single tensor vs separate variables), spatial/temporal downsampling, multi-variable systems, HuggingFace and DaRUS data sources, and efficient PyTorch DataLoader creation. Use when preparing PDE data for neural operator training.
The English summary is being prepared. Showing the original description.
Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.
The English summary is being prepared. Showing the original description.
Expert guidance for Fully Sharded Data Parallel training with PyTorch FSDP - parameter sharding, mixed precision, CPU offloading, FSDP2
The English summary is being prepared. Showing the original description.
High-level PyTorch framework with Trainer class, automatic distributed training (DDP/FSDP/DeepSpeed), callbacks system, and minimal boilerplate. Scales from laptop to supercomputer with same code. Use when you want clean training loops with built-in best practices.
The English summary is being prepared. Showing the original description.
AI基础设施实战——从GPU集群调度到模型推理部署的全栈实操指南。覆盖：昇腾/GPU统一调度、PyTorch分布式训练、vLLM推理部署、模型量化实战、LLMOps、GPU故障诊断、HuggingFace生态、MLOps流水线。触发词：AI基础设施、GPU、昇腾、训练、推理、vLLM、PyTorch分布式、模型部署、LLMOps、MLOps、集群调度、HuggingFace
The English summary is being prepared. Showing the original description.
Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.
The English summary is being prepared. Showing the original description.
Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.
The English summary is being prepared. Showing the original description.
Add machine learning to an existing application through a structured workflow. The skill guides problem definition, data checks, separation from business logic, and training and evaluation of a simple initial model.
Use SAM to separate objects from images by marking points or drawing boxes, without training a model for each task. Covers automatic region generation, refining selections, preparing annotation data, and exporting models for deployment.
Set up GPU servers on Lambda Labs for machine learning training and inference. Follow workflows for SSH access, persistent storage, Python API operations, and distributed training that spreads work across multiple GPUs or servers.
Train language models on pairs of preferred and rejected answers without a separate reference model. The skill guides training setup, model-specific configurations, parameter tuning, and troubleshooting for unstable training or memory limits.
Train and run Uni-Mol Tools models to predict molecular properties from CSV data. Organize experiments into projects, compare model performance, rank results, and manage storage through interactive commands or machine-readable JSON output.
Build and evaluate single-agent reinforcement learning experiments, where an agent learns actions from rewards. Use Stable Baselines3 to choose algorithms, create environments, monitor training, and save models with the normalization data needed for consistent evaluation.