- Removed GraspCriticNetworkConfig class and integrated its parameters into SACConfig.
- Added num_discrete_actions parameter to SACConfig for better action handling.
- Updated SACPolicy to conditionally create grasp critic networks based on num_discrete_actions.
- Enhanced grasp critic forward pass to handle discrete actions and compute losses accordingly.
- Integrated the grasp critic gradient update to the training loop in learner_server
- Added Adam optimizer and configured grasp critic learning rate in configuration_sac
- Added target critics networks update after the critics gradient step
- Enhanced type annotations for variables in the `SACPolicy` class to improve code clarity.
- Updated method calls to use keyword arguments for better readability.
- Streamlined the extraction of batch components, ensuring consistent typing across the class methods.
- Updated the `forward` method in `SACPolicy` to handle loss computation for actor, critic, and temperature models.
- Replaced direct calls to `compute_loss_*` methods with a unified `forward` method in `learner_server`.
- Enhanced batch processing by consolidating input parameters into a single dictionary for better readability and maintainability.
- Removed redundant code and improved documentation for clarity.
Moved HilSerl env config to configs/env/configs.py
fixes in actor_server and modeling_sac and configuration_sac
added the possibility of ignoring missing keys in env_cfg in get_features_from_env_config function
- Reduced frame rate in `ManiskillEnvConfig` from 400 to 200.
- Enhanced `SACConfig` with new dataclasses for actor, learner, and network configurations.
- Improved input and output feature management in `SACConfig`.
- Refactored `actor_server` and `learner_server` to access configuration properties directly.
- Updated training pipeline to validate configurations and handle dataset repo IDs more robustly.
- Deleted the `find_and_copy_params` function and the `Ensemble` class, as they were deemed unnecessary.
- Cleaned up imports by removing `from_modules` from `tensordict` to enhance code clarity.
- Simplified the assertion in the `Policy` class for better readability.
- Updated SACPolicy to create critic heads using a list comprehension for better readability.
- Simplified the saving and loading of models using `save_model` and `load_model` functions from the safetensors library.
- Introduced `initialize_offline_replay_buffer` function in the learner server to streamline offline dataset handling and replay buffer initialization.
- Enhanced logging for dataset loading processes to improve traceability during training.
- Introduced `grad_clip_norm` parameter in SAC configuration for gradient clipping
- Updated SACPolicy to store temperature as an instance variable for consistent usage
- Modified loss calculations in SACPolicy to utilize the instance temperature
- Enhanced MLP and CriticHead to support a customizable final activation function
- Implemented gradient clipping in the learner server during training steps for both actor and critic
- Added tracking for gradient norms in training information
- Implement `_save_pretrained` method to handle TensorDict state saving
- Add `_from_pretrained` class method for loading SAC policy from files
- Create utility function `find_and_copy_params` to handle parameter copying
- Removed `@torch.no_grad` decorator from Unnormalize forward method
- Added TODO comment for optimizing next action prediction in SAC policy
- Minor formatting adjustment in NaN assertion for log standard deviation
Co-authored-by: Yoel Chornton <yoel.chornton@gmail.com>
- Introduced Ensemble and CriticHead classes for more efficient critic network handling
- Added support for multiple camera inputs in observation encoder
- Optimized image encoding by batching image processing
- Updated configuration for ManiSkill environment with reduced image size and action scaling
- Compiled critic networks for improved performance
- Simplified normalization and ensemble handling in critic networks
Co-authored-by: michel-aractingi <michel.aractingi@gmail.com>
- Fixed big issue in the normalization of the actions in the `forward` function of the critic -- remove the `torch.no_grad` decorator in `normalize.py` in the normalization function
- Fixed performance issue to boost the optimization frequency by setting the storage device to be the same as the device of learning.
Co-authored-by: Adil Zouitine <adilzouitinegm@gmail.com>
- added `torch.compile` to the actor and learner servers.
- organized imports in `train_sac.py`
- optimized the parameters push by not sending the frozen pre-trained encoder.
Co-authored-by: Adil Zouitine <adilzouitinegm@gmail.com>
Added functions for converting the replay buffer from and to LeRobotDataset. When we want to save the replay buffer, we convert it first to LeRobotDataset format and save it locally and vice-versa.
Co-authored-by: Adil Zouitine <adilzouitinegm@gmail.com>
- Optimized critic design that improves the performance of the learner loop by a factor of 2
- Cleaned the code and fixed style issues
- Completed the config with actor_learner_config field that contains host-ip and port elemnts that are necessary for the actor-learner servers.
Co-authored-by: Adil Zouitine <adilzouitinegm@gmail.com>
- Updated SACConfig to replace standard deviation parameterization with log_std_min and log_std_max for better control over action distributions.
- Modified SACPolicy to streamline action selection and log probability calculations, enhancing stochastic behavior.
- Removed deprecated TanhMultivariateNormalDiag class to simplify the codebase and improve maintainability.
These changes aim to enhance the robustness and performance of the SAC implementation during training and inference.
- Updated standard deviation parameterization in SACConfig to 'softplus' with defined min and max values for improved stability.
- Modified action sampling in SACPolicy to use reparameterized sampling, ensuring better gradient flow and log probability calculations.
- Cleaned up log probability calculations in TanhMultivariateNormalDiag for clarity and efficiency.
- Increased evaluation frequency in YAML configuration to 50000 for more efficient training cycles.
These changes aim to enhance the robustness and performance of the SAC implementation during training and inference.
- Updated action selection to use distribution sampling and log probabilities for better stochastic behavior.
- Enhanced standard deviation clamping to prevent extreme values, ensuring stability in policy outputs.
- Cleaned up code by removing unnecessary comments and improving readability.
These changes aim to refine the SAC implementation, enhancing its robustness and performance during training and inference.
AdilZouitine
pre-commit-ci[bot]
s1lent4gnt
Michel Aractingi
Yoel
Eugene Mironov
KeWang1017