CMS Machine Learning Documentation

SWAN + AWS

Initializing search

cms-ml/documentation

CMS Machine Learning Documentation

cms-ml/documentation

Home
Newsletters
Innovation
Innovation
- ML Journal Club
- ML Hackathons
Resources
Resources
- Cloud Resources
- Dataset Resources
- FPGA Resource
- GPU Resources
  GPU Resources
Guides
Guides
- Software environments
  Software environments
  - LCG environments
  - Using containers
- Optimization
  Optimization
- General Advice
  General Advice
  - Introduction
  - Before training
    Before training
    
    Domains
    
    Features
    
    Inputs
    
    Model
    
    Metrics & Losses
  - During training
    During training
    
    Overfitting
    
    Cross-validation
    
    Optimisation problems
  - After training
- Inference
  Inference
  - Direct inference
    Direct inference
    
    TensorFlow 2
    
    TensorFlow AOT
    
    PyTorch
    
    PyTorch Geometric
    
    ONNX
    
    XGBoost
    
    hls4ml
    
    conifer
  - Inference as a service
    Inference as a service
    
    Sonic/Triton
    
    TFaaS
  - Non-standard workflows
    Non-standard workflows
    
    Standalone framework
    
    SWAN + AWS
  - Integration checklist
  - Performance
  - Successful integrations
    Successful integrations
    
    ParticleNet
- Training
  Training
  - Decorrelation
  - Training as a Service
    Training as a Service
    
    MLaaS4HEP
  - Autoencoders

SWAN + AWS

Todo.

Ideas: best practices cost model instance priving need to log out monitoring madatory

November 17, 2024

Copyright © 2020-2023 CMS Machine Learning Group

Made with Material for MkDocs