Skip to content

This is a community-supported tool. If you'd like to contribute, check out our GitHub repository. Your contributions are welcome!

meds-torch

ddp.yaml

Initializing search

meds-torch

Home
Inductive Experiments
Terminology
Input Encoding Methods
API Reference
API Reference
- data
  data
  - components
    
    components
    
    multiwindow_pytorch_dataset
    
    pytorch_dataset
    
    random_windows_pytorch_dataset
  - datamodule
- eval
- finetune
- generate_trajectories
- input_encoder
  input_encoder
- latest_dir
- models
  models
  - base_model
  - components
    
    components
    
    lstm
    
    transformer_decoder
    
    transformer_encoder
    
    utils
  - ebcl_model
  - eic_forecasting
  - ocp_model
  - supervised_model
  - triplet_forecasting
  - utils
  - value_forecasting
- predict
- schemas
  schemas
  - generate_analysis_schema
  - predict_schema
- train
- tune
- utils
  utils
Config Reference
Config Reference
- callbacks
  callbacks
- data
  data
- debug
  debug
- eval.yaml
- experiment
  experiment
  - example.yaml
  - mimiciv.yaml
- extras
  extras
  - default.yaml
- finetune.yaml
- generate_trajectories.yaml
- hparams_search
  hparams_search
- hydra
  hydra
  - default.yaml
- latest_dir.yaml
- logger
  logger
- model
  model
  - backbone
    backbone
    
    default.yaml
    
    eic_lstm.yaml
    
    eic_transformer_decoder.yaml
    
    eic_transformer_decoder_alibi.yaml
    
    eic_transformer_encoder.yaml
    
    eic_transformer_encoder_attn_avg.yaml
    
    textcode_transformer_encoder.yaml
    
    triplet_lstm.yaml
    
    triplet_transformer_decoder.yaml
    
    triplet_transformer_encoder.yaml
    
    triplet_transformer_encoder_attn_avg.yaml
  - ebcl.yaml
  - eic_forecasting.yaml
  - input_encoder
    input_encoder
    
    eic_encoder.yaml
    
    textcode_encoder.yaml
    
    triplet_encoder.yaml
    
    triplet_prompt_encoder.yaml
  - ocp.yaml
  - optimizer
    optimizer
    
    adam.yaml
  - scheduler
    scheduler
    
    cosine_annealing.yaml
    
    reduce_lr_on_plateau.yaml
  - supervised.yaml
  - trajectory_labeler
    trajectory_labeler
    
    aces_schema_labeler.yaml
  - triplet_forecasting.yaml
  - value_forecasting.yaml
- paths
  paths
  - default.yaml
- predict.yaml
- train.yaml
- trainer
  trainer
  - cpu.yaml
  - ddp.yaml
  - ddp_sim.yaml
  - default.yaml
  - gpu.yaml
  - mps.yaml
  - ray.yaml
Issues

ddp

defaults:
  - default

strategy: ddp_find_unused_parameters_true

accelerator: gpu
devices: 4
num_nodes: 1
sync_batchnorm: True

February 6, 2025

Made with Material for MkDocs