jolibrain/wheatley issues and pull requests

#108 - WIP: conflicts as attention

Pull Request - State: open - Opened by fantes 10 months ago

#107 - feat: add pdf conv

Pull Request - State: closed - Opened by fantes 10 months ago

#106 - feat: padding in the agent actions, removing the hard-coded fixed max number of nodes

Pull Request - State: open - Opened by pierrot-lc 10 months ago

#105 - feat: structured reward

Pull Request - State: closed - Opened by fantes 10 months ago

#104 - feat: allow to store rollouts on disk for gym envs

Pull Request - State: closed - Opened by royale 11 months ago

#103 - feat: allow to choose critic loss in L1, L2, smoothL1

Pull Request - State: closed - Opened by fantes 11 months ago

#102 - feat(psp): Two hot value encoding + symlog internal prediction

Pull Request - State: closed - Opened by fantes 11 months ago

#101 - feat(results): cplex results

Pull Request - State: closed - Opened by pierrot-lc 11 months ago

#100 - feat(visdom): rename PPO to Wheatley in charts

Pull Request - State: closed - Opened by royale 11 months ago

#99 - feat(psp): deterministic train/test split

Pull Request - State: closed - Opened by royale 11 months ago

#98 - doc: fix PSP train command

Pull Request - State: closed - Opened by royale 11 months ago

#97 - feat(psp/graphgym): allow to store observation of rollouts on disk

Pull Request - State: closed - Opened by fantes 11 months ago

#96 - fix(psp): make resource as node work

Pull Request - State: closed - Opened by fantes 11 months ago

#95 - Dispatching rules: handle missing tasks and floating-points durations

Pull Request - State: closed - Opened by pierrot-lc 12 months ago

#94 - fix: return based scaling in jssp/eval.py

Pull Request - State: closed - Opened by pierrot-lc 12 months ago

#93 - fix: skip advantage normalization if last minibatch size is 1

Pull Request - State: closed - Opened by royale 12 months ago

#92 - Observation horizon, subgraphs as obs and fast forward

Pull Request - State: closed - Opened by fantes 12 months ago

#91 - Pb with wheatley on Mac: device not recognized

Issue - State: closed - Opened by galleon 12 months ago - 4 comments

#90 - docs: psp training command and removed old USAGE.md

Pull Request - State: closed - Opened by pierrot-lc 12 months ago
Labels: documentation, ci:skip-tests

#89 - Different arrival times [feature request]

Issue - State: open - Opened by Maros112358 12 months ago - 7 comments

#88 - feat(instances): adding results for each instances

Pull Request - State: closed - Opened by pierrot-lc 12 months ago

#87 - Adding mps:0 as device for Metal

Pull Request - State: closed - Opened by galleon 12 months ago - 1 comment

#86 - feat(psp): ag1 loader

Pull Request - State: closed - Opened by fantes 12 months ago

#86 - feat(psp): ag1 loader

Pull Request - State: closed - Opened by fantes 12 months ago

#85 - feat: allow inverse learn pool

Pull Request - State: closed - Opened by fantes 12 months ago

#85 - feat: allow inverse learn pool

Pull Request - State: closed - Opened by fantes 12 months ago

#84 - chore(psp): cleanup torch deadlock workaround

Pull Request - State: closed - Opened by fantes 12 months ago

#83 - chore(psp): internally use rcpsp class instead of dict

Pull Request - State: closed - Opened by fantes 12 months ago

#82 - tests(psp): non-regression test

Pull Request - State: closed - Opened by pierrot-lc 12 months ago

#81 - chore(ml): GAE is off by default (lambda = 1.0)

Pull Request - State: closed - Opened by beniz 12 months ago
Labels: rl:algorithm, jssp, rcpsp

#80 - feat(psp): add option for (sym)log reward normalization

Pull Request - State: closed - Opened by fantes about 1 year ago

#79 - feat: return-based-scaling (2105.05347)

Pull Request - State: closed - Opened by fantes about 1 year ago
Labels: enhancement

#78 - chore(psp): add graphgym test

Pull Request - State: closed - Opened by fantes about 1 year ago
Labels: ci:tests

#77 - chore(psp): add graphgym test

Pull Request - State: closed - Opened by fantes about 1 year ago
Labels: ci:tests

#76 - Provide an easy way to use a trained model for inference

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#75 - Get a solution

Issue - State: closed - Opened by Maros112358 about 1 year ago - 5 comments

#74 - Generate JSSP validation instances, loading those instances at evaluation

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#73 - feat(rcpsp): homemade gym-like env

Pull Request - State: closed - Opened by fantes about 1 year ago

#72 - fix(psp): allow null durations in stochastic psp 2 (fix the fix)

Pull Request - State: closed - Opened by fantes about 1 year ago

#71 - fix(psp): allow null durations in stochastic psp

Pull Request - State: closed - Opened by fantes about 1 year ago
Labels: rcpsp

#70 - fix(psp): null durations in stochastic psp

Pull Request - State: closed - Opened by beniz about 1 year ago
Labels: rcpsp

#69 - fix(psp): fix reward for multi-sinks

Pull Request - State: closed - Opened by fantes about 1 year ago

#68 - Save ortools solutions on disk when evaluating on the same fixed instances

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago
Labels: data:loaders

#67 - feat(tests): RC-PSP argparse tests

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago - 1 comment
Labels: ci:tests

#66 - feat(tests): separate performance tests from other tests

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#65 - fix(psp): allow unconnected sinks + small fixes

Pull Request - State: closed - Opened by fantes about 1 year ago

#64 - chore: Jenkins CI

Pull Request - State: closed - Opened by beniz about 1 year ago
Labels: ci:tests

#63 - Unittest scripts

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago
Labels: ci:tests

#62 - Docs: an updated version

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago
Labels: documentation

#61 - refactor: separate psp from jssp

Pull Request - State: closed - Opened by fantes about 1 year ago

#60 - fix(psp): add seed info

Pull Request - State: closed - Opened by fantes about 1 year ago

#59 - fix(rcpsp): add wd missing param

Pull Request - State: closed - Opened by fantes about 1 year ago

#58 - Reinitialize agent heads (value and actor) before starting PPO

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#57 - feat(psp): add seperated rwpe net

Pull Request - State: closed - Opened by fantes about 1 year ago

#56 - Scatter plot: OR-Tools vs PPO

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#55 - feat(jssp): remove machine id in edge type as a default

Pull Request - State: closed - Opened by fantes about 1 year ago

#54 - Problem resets and generation

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#53 - fix(ppo): make iter_size work

Pull Request - State: closed - Opened by fantes about 1 year ago

#52 - Pretrain : Visdom plots and evaluation metrics

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#51 - Handling multiple ortools strategies

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#50 - Reload previous training curves on resume

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago

#49 - feat(psp): allow multiple sources and sinks

Pull Request - State: closed - Opened by fantes about 1 year ago

#48 - Benchmark a trained model on all instances

Pull Request - State: closed - Opened by pierrot-lc about 1 year ago