Skip to content

feat: Training pipeline — YAML config, seeds, checkpoint resume (Closes #16) - #176

Open
laurentketterle-hub wants to merge 2 commits into
mergeos-bounties:masterfrom
laurentketterle-hub:feat/training-pipeline-16
Open

feat: Training pipeline — YAML config, seeds, checkpoint resume (Closes #16)#176
laurentketterle-hub wants to merge 2 commits into
mergeos-bounties:masterfrom
laurentketterle-hub:feat/training-pipeline-16

Conversation

@laurentketterle-hub

Copy link
Copy Markdown

Implements training pipeline for PoseGuide #16 (100 MRG).

Files added:

  • config/training_config.yaml — Complete training config (model, data, optimizer, scheduler, seeds)
  • poseguide/train.py — TrainingPipeline class with seed management and checkpoint resume

Features:

  • YAML-driven configuration for reproducible runs
  • Random seed management (Python, NumPy, PyTorch, CUDA deterministic)
  • Automatic checkpoint resume detection
  • Device auto-detection (CUDA / MPS / CPU)
  • Configurable data splits, augmentations, and learning rate scheduling
python poseguide/train.py -c config/training_config.yaml -d ./data

Implements reproducible training pipeline for PoseGuide#16.
- YAML config with all hyperparameters
- Seed management for reproducibility
- Checkpoint resume support
- Device auto-detection (CUDA/MPS/CPU)

Signed-off-by: laurentketterle-hub <laurentketterle-hub@users.noreply.github.com>
Implements reproducible training pipeline for PoseGuide#16.
- YAML config with all hyperparameters
- Seed management for reproducibility
- Checkpoint resume support
- Device auto-detection (CUDA/MPS/CPU)

Signed-off-by: laurentketterle-hub <laurentketterle-hub@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant