In #6490 we enabled non-blocking torch device transfers throughout the model manager's memory management code. When using this torch feature, torch attempts to wait until the tensor transfer has completed before allowing any access to the tensor. Theoretically, that should make this a safe feature to use.
This provides a small performance improvement but causes race conditions in some situations. Specific platforms/systems are affected, and complicated data dependencies can make this unsafe.
- Intermittent black images on MPS devices - reported on discord and #6545, fixed with special handling in #6549.
- Intermittent OOMs and black images on a P4000 GPU on Windows - reported in #6613, fixed in this commit.
On my system, I haven't experience any issues with generation, but targeted testing of non-blocking ops did expose a race condition when moving tensors from CUDA to CPU.
One workaround is to use torch streams with manual sync points. Our application logic is complicated enough that this would be a lot of work and feels ripe for edge cases and missed spots.
Much safer is to fully revert non-locking - which is what this change does.
We can get black outputs when moving tensors from CPU to MPS. It appears MPS to CPU is fine. See:
- https://github.com/pytorch/pytorch/issues/107455
- https://discuss.pytorch.org/t/should-we-set-non-blocking-to-true/38234/28
Changes:
- Add properties for each device on `TorchDevice` as a convenience.
- Add `get_non_blocking` static method on `TorchDevice`. This utility takes a torch device and returns the flag to be used for non_blocking when moving a tensor to the device provided.
- Update model patching and caching APIs to use this new utility.
Fixes: #6545
- Fix type errors
- Enable SilenceWarnings to be used as both a context manager and a decorator
- Remove duplicate implementation
- Check the initial verbosity on __enter__() rather than __init__()
Pending:
- Move model install calls into model manager and create passthrus in invocation_context.
- Consider splitting load_model_from_url() into a call to get the path and a call to load the path.
* introduce new abstraction layer for GPU devices
* add unit test for device abstraction
* fix ruff
* convert TorchDeviceSelect into a stateless class
* move logic to select context-specific execution device into context API
* add mock hardware environments to pytest
* remove dangling mocker fixture
* fix unit test for running on non-CUDA systems
* remove unimplemented get_execution_device() call
* remove autocast precision
* Multiple changes:
1. Remove TorchDeviceSelect.get_execution_device(), as well as calls to
context.models.get_execution_device().
2. Rename TorchDeviceSelect to TorchDevice
3. Added back the legacy public API defined in `invocation_api`, including
choose_precision().
4. Added a config file migration script to accommodate removal of precision=autocast.
* add deprecation warnings to choose_torch_device() and choose_precision()
* fix test crash
* remove app_config argument from choose_torch_device() and choose_torch_dtype()
---------
Co-authored-by: Lincoln Stein <lstein@gmail.com>
- Allow user-defined precision on MPS.
- Use more explicit logic to handle all possible cases.
- Add comments.
- Remove the app_config args (they were effectively unused, just get the config using the singleton getter util)
- Replace AnyModelLoader with ModelLoaderRegistry
- Fix type check errors in multiple files
- Remove apparently unneeded `get_model_config_enum()` method from model manager
- Remove last vestiges of old model manager
- Updated tests and documentation
resolve conflict with seamless.py
- Rename old "model_management" directory to "model_management_OLD" in order to catch
dangling references to original model manager.
- Caught and fixed most dangling references (still checking)
- Rename lora, textual_inversion and model_patcher modules
- Introduce a RawModel base class to simplfy the Union returned by the
model loaders.
- Tidy up the model manager 2-related tests. Add useful fixtures, and
a finalizer to the queue and installer fixtures that will stop the
services and release threads.
- Implement new model loader and modify invocations and embeddings
- Finish implementation loaders for all models currently supported by
InvokeAI.
- Move lora, textual_inversion, and model patching support into
backend/embeddings.
- Restore support for model cache statistics collection (a little ugly,
needs work).
- Fixed up invocations that load and patch models.
- Move seamless and silencewarnings utils into better location
- Cache stat collection enabled.
- Implemented ONNX loading.
- Add ability to specify the repo version variant in installer CLI.
- If caller asks for a repo version that doesn't exist, will fall back
to empty version rather than raising an error.
We have two different classes named `ModelInfo` which might need to be used by API consumers. We need to export both but have to deal with this naming collision.
The `ModelInfo` I've renamed here is the one that is returned when a model is loaded. It's the object least likely to be used by API consumers.
* feat: allow bfloat16 to be configurable in invoke.yaml
* fix: `torch_dtype()` util
- Use `choose_precision` to get the precision string
- Do not reference deprecated `config.full_precision` flat (why does this still exist?), if a user had this enabled it would override their actual precision setting and potentially cause a lot of confusion.
---------
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
* remove MacOS Sonoma check in devices.py
As of pytorch 2.1.0, float16 works with our MPS fixes on Sonoma, so the check is no longer needed.
* remove unused platform import
