tools
Modules
Utilities for initializing child models from parent models via bypassed training. |
|
Utilities for loading and initializing PyTorch model checkpoints (AnyModel / HF layouts). |
|
Utilities for loading and saving Hugging Face-format checkpoints ( |
|
Utilities for hydra config initialization. |
|
Knowledge distillation loss functions. |
|
Provides utilities for distributed loading, saving, and manipulation of large language model checkpoints across multiple GPUs/processes. |
|
Provides a function to validate a model. |
|
|
Validates puzzle solutions by applying layer replacements and evaluating model performance. |
Utility functions for validating models and extracting hidden states and similarity metrics. |
Shared tools: logging, hydra config, checkpoint utilities, and validation helpers.