lebrice/sequoia issues and pull requests

#98 - Add HPO Sweep functionality to other Methods than BaslineMethod

Issue - State: closed - Opened by lebrice almost 4 years ago
Labels: enhancement

#97 - Add LwF demo

Pull Request - State: closed - Opened by liuyudut almost 4 years ago - 1 comment

#96 - Add a CI/CD Pipeline

Issue - State: open - Opened by lebrice almost 4 years ago
Labels: help wanted, tests

#95 - "Fix" Class/Task-Incremental Settings, Add DomainIncrementalSetting

Pull Request - State: closed - Opened by lebrice almost 4 years ago - 4 comments

#94 - Add a Multi-Task Supervised Learning Setting

Issue - State: closed - Opened by lebrice almost 4 years ago - 1 comment
Labels: enhancement, New Setting

#93 - Update README.md

Pull Request - State: closed - Opened by optimass almost 4 years ago

#92 - consider switching `master` branch to `main`

Issue - State: open - Opened by fgolemo almost 4 years ago
Labels: low priority

#91 - fixed setup

Pull Request - State: closed - Opened by fgolemo almost 4 years ago - 1 comment

#90 - Add link for Mailing list in README.md

Pull Request - State: closed - Opened by optimass about 4 years ago

#89 - Fix small bugs

Pull Request - State: closed - Opened by lebrice about 4 years ago

#88 - Add "Multi-Task RL" and partial fix for BaselineModel in that context

Pull Request - State: closed - Opened by lebrice about 4 years ago

#87 - Add Submitit integration for launching lots of SLURM jobs!

Issue - State: open - Opened by lebrice about 4 years ago - 3 comments
Labels: enhancement, low priority

#86 - More General CL Framework

Issue - State: open - Opened by prlz77 about 4 years ago
Labels: enhancement

#85 - List of existing Methods should precede Instructions on adding a new Method in README

Issue - State: open - Opened by lebrice about 4 years ago
Labels: documentation, good first issue

#84 - Cleanup / Fix Failing tests

Pull Request - State: closed - Opened by lebrice about 4 years ago

#83 - Add Replay to BaselineMethod in RL Settings

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement

#82 - Move documentation, improve docs of hparam_tuning.py

Pull Request - State: closed - Opened by lebrice about 4 years ago

#81 - IID RL should sample new tasks after each episode

Issue - State: closed - Opened by lebrice about 4 years ago
Labels: bug, enhancement, design problem

#80 - Rebased "reg_methods" - Add EwcMethod and EWC Auxiliary Task

Pull Request - State: closed - Opened by lebrice about 4 years ago - 1 comment

#79 - Add `hparam_sweep` method on the BaselineMethod class

Pull Request - State: closed - Opened by lebrice about 4 years ago

#78 - Reorganize PNN Demo files (see desc.)

Pull Request - State: closed - Opened by lebrice about 4 years ago

#77 - Fix Encoder having no gradients, simplified backward pass of output head in RL

Pull Request - State: closed - Opened by lebrice about 4 years ago

#76 - Add synbols from continuum

Pull Request - State: closed - Opened by prlz77 about 4 years ago

#75 - Add Progressive Neural Networks (PNN) examples for SL and RL

Pull Request - State: closed - Opened by JuliousHurtado about 4 years ago

#74 - Add MonsterKong environment support

Pull Request - State: closed - Opened by lebrice about 4 years ago

#73 - Added better / simpler demo for creating a Method for RL + SL

Pull Request - State: closed - Opened by lebrice about 4 years ago

#72 - Fix #62, use a FCNet encoder in Cartpole state

Pull Request - State: closed - Opened by lebrice about 4 years ago - 1 comment

#71 - Encoder doesn't get gradients

Issue - State: closed - Opened by oleksost about 4 years ago
Labels: bug

#70 - Test the initial continual-RL track on MonsterKong

Issue - State: closed - Opened by optimass about 4 years ago - 2 comments

#69 - Rename "fit" --> "train" and "get_actions" -> "predict"

Issue - State: open - Opened by lebrice about 4 years ago
Labels: low priority

#68 - bugfixes for mac

Pull Request - State: closed - Opened by fgolemo about 4 years ago

#67 - "ClassIncrementalSetting" isn't really Class-Incremental, more like Domain-Incremental

Issue - State: closed - Opened by lebrice about 4 years ago - 7 comments
Labels: design problem, New Setting

#66 - Synbols

Pull Request - State: closed - Opened by prlz77 about 4 years ago - 3 comments

#65 - Hat demo

Pull Request - State: closed - Opened by JuliousHurtado about 4 years ago

#64 - Design / Add a mechanism to facilitate Hyper-Parameter Optimization of Methods

Issue - State: closed - Opened by lebrice about 4 years ago
Labels: enhancement, low priority

#63 - Add Forward Model Auxiliary task for Model-Based RL

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement, low priority

#62 - BaselineModel doesn't have any shared weights (no encoder) in cartpole-state

Issue - State: closed - Opened by lebrice about 4 years ago - 1 comment
Labels: bug

#61 - Misc changes: Reorganize the examples, Add render for passive envs, etc.

Pull Request - State: closed - Opened by lebrice about 4 years ago

#60 - Design a mechanism for auto-generating citations

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement, design problem, low priority

#59 - monitoring: Disk Usage

Issue - State: closed - Opened by optimass about 4 years ago
Labels: enhancement

#58 - monitoring: GPU usage

Issue - State: closed - Opened by optimass about 4 years ago
Labels: enhancement

#57 - monitoring: RAM usage

Issue - State: closed - Opened by optimass about 4 years ago
Labels: enhancement

#56 - monitoring: add total training/test time

Issue - State: closed - Opened by optimass about 4 years ago
Labels: enhancement

#55 - monitoring: Add average performance over time

Issue - State: closed - Opened by optimass about 4 years ago - 2 comments
Labels: enhancement

#54 - Add a table listing the differences in assumptions between Settings

Issue - State: closed - Opened by lebrice about 4 years ago - 1 comment
Labels: documentation

#53 - [WIP] Add Uml Class diagrams to the repo

Pull Request - State: closed - Opened by lebrice about 4 years ago - 1 comment
Labels: documentation

#52 - Add stack(), split(), torch() methods to Batch

Pull Request - State: closed - Opened by lebrice about 4 years ago

#51 - Add Class diagram for the BaselineMethod

Issue - State: closed - Opened by lebrice about 4 years ago

#50 - "proper" logging to wandb

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement

#49 - Add interchangeable Task-inference modules

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement

#48 - [WIP] Baseline Method works better in RL, partial refactoring of common/transforms

Pull Request - State: closed - Opened by lebrice about 4 years ago - 1 comment

#47 - Improved command-line API

Issue - State: open - Opened by lebrice about 4 years ago - 4 comments
Labels: help wanted, design problem

#46 - Refactor the training/evaluation loops to produce entire transfer matrix

Issue - State: closed - Opened by lebrice about 4 years ago - 1 comment
Labels: enhancement

#45 - Reg methods

Pull Request - State: closed - Opened by oleksost about 4 years ago

#44 - Add integration tests that run each example and verify the results

Issue - State: open - Opened by lebrice about 4 years ago
Labels: tests

#43 - Add more types of supervised datasets apart from classification, (e.g. Regression, Segmentation, etc)

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement, help wanted, New Dataset

#42 - possible bug in ewc_in_rl

Issue - State: open - Opened by oleksost about 4 years ago
Labels: bug

#41 - MonsterKong Environment

Issue - State: closed - Opened by mattriemer about 4 years ago
Labels: enhancement, New Dataset, New Setting

#40 - add support for Synbols

Issue - State: closed - Opened by optimass about 4 years ago
Labels: enhancement, New Dataset

#39 - Installing other repos (e.g. nngeometry) by git url as dependencies through `pip install -e .`

Issue - State: closed - Opened by lebrice about 4 years ago - 1 comment
Labels: bug

#38 - Add 'sequoia' as entry-point command for the package,

Pull Request - State: closed - Opened by lebrice about 4 years ago

#37 - xvbf fix

Pull Request - State: closed - Opened by prlz77 about 4 years ago - 2 comments

#36 - Add example of how to add regularization-type loss to SB3 methods

Pull Request - State: closed - Opened by lebrice about 4 years ago

#35 - Rename repo to Sequoia, reorganize files/folders

Pull Request - State: closed - Opened by lebrice about 4 years ago

#34 - Refactoring + BaselineMethod works in batched RL Settings

Pull Request - State: closed - Opened by lebrice about 4 years ago

#33 - Add CN-DPM as a method for task-free Supervised CL Settings

Issue - State: open - Opened by lebrice about 4 years ago
Labels: New Method

#32 - Meta/Few-shot Learning Setting(s)

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement, design problem, low priority, New Setting

#31 - Add Meta-World as a source of datasets/tasks for Continual RL settings

Issue - State: open - Opened by lebrice about 4 years ago
Labels: New Dataset

#30 - Add GDumb Method for Supervised CL

Issue - State: closed - Opened by lebrice about 4 years ago - 1 comment
Labels: help wanted, New Method

#29 - Minor tweaks (typing, formatting, etc.)

Pull Request - State: closed - Opened by lebrice about 4 years ago

#28 - EWC regularization method

Pull Request - State: closed - Opened by oleksost about 4 years ago

#27 - Add "proper" seeding: Make Settings entirely deterministic given a random seed

Issue - State: open - Opened by lebrice about 4 years ago
Labels: enhancement

#26 - RL Rework, more methods from Stable-Baselines3

Pull Request - State: closed - Opened by lebrice about 4 years ago

#25 - Experience replay

Pull Request - State: closed - Opened by pclucas14 about 4 years ago

#24 - Fix import bugs, make the demo scripts easier to read and to run

Pull Request - State: closed - Opened by lebrice about 4 years ago

#23 - Validate the RL "baseline" Methods

Issue - State: closed - Opened by lebrice over 4 years ago - 1 comment

#22 - Make the BaselineMethod truly applicable on RL Settings

Issue - State: closed - Opened by lebrice over 4 years ago - 1 comment

#21 - Add the Continual RL Branch to the Tree

Pull Request - State: closed - Opened by lebrice over 4 years ago

#20 - Add support for larger batch sizes in RL

Pull Request - State: closed - Opened by lebrice over 4 years ago

#19 - Big rework / Simplified API

Pull Request - State: closed - Opened by lebrice over 4 years ago

#18 - Add Smooth Task Boundaries in Supervised CL

Issue - State: open - Opened by lebrice over 4 years ago
Labels: New Setting

#17 - Add Support for Semi-Supervision

Issue - State: open - Opened by lebrice over 4 years ago - 1 comment
Labels: design problem

#16 - Make it possible to use existing models from pl_bolts as Methods and datamodules as IIDSettings in the tree with minimal overhead.

Issue - State: open - Opened by lebrice over 4 years ago
Labels: help wanted, design problem

GitHub / lebrice/sequoia issues and pull requests