v1.1.0

Security

• Raise the locked versions of dependencies with open Dependabot advisories:
  Pillow 13.0.0, certifi 2026.10.23, fsspec 2026.9.0, GitPython 3.2.1,
  nbconvert 7.17.2, filelock 4.2.4, virtualenv 21.14.13, setuptools 84.2.1,
  opentelemetry-api 1.45.1 and Werkzeug 3.1.9. constraint-dependencies
  now keeps each of them at or above its patched version.

Changed

• simulusion train MODEL runs registered tasks on the shared Trainer.
  Commands written for the older per-family entrypoints (a --config without
  task:, or flat overrides such as total_epochs=100) still reach them, as
  does --legacy.
• The per-family training entrypoints run on the shared Trainer.
  DreamerAgent.train(), IRISTrainer.train(), DiamondAgent.train(),
  Planet.train() / train_planet.main, GenieTrainer.train(),
  DiT.fit() / train_dit, JEPAAgent.train(), train_convae,
  train_mdn_rnn and train_controller keep their signatures, return values
  and the files they always wrote, through one helper
  (simulusion.training.legacy.fit_legacy). Their own loops are gone. They
  now also write resumable checkpoints and run.json / run_config.yaml /
  train_metrics.jsonl into the same folder. Small differences (checkpoint
  step names, IRIS/DIAMOND not evaluating after epoch 0, MDN-RNN latents
  sampled per batch, the controller scored in one process) are listed in the
  trainer guide. W&B sweeps of JEPAAgent.train still use train_jepa.
• The Trainer keeps the best evaluation's checkpoint (best_metric,
  best_mode), lets a task end training early (TrainingTask.should_stop),
  and writes the last checkpoint after the final evaluation.
• Episode folders read World Models rollout files (per-step observations /
  actions / rewards / terminals) directly.

Fixed

• ReduceLROnPlateau.state_dict() included the live optimizer, so any
  checkpoint holding it (train_convae's and train_mdn_rnn's best.tar)
  could not be loaded with weights_only=True, as the repo loads everything.
• The World Models controller's real-environment rollouts unpacked
  GymImageEnv.reset() as a Gymnasium (obs, info) pair; it returns the
  observation alone, so every rollout failed.
• A resumed DIAMOND run kept a replay buffer only as large as it was at the
  checkpoint: save_checkpoint recorded the trimmed length as the capacity.
  Its replay and observation side files are now found next to the
  checkpoint even after the directory moves.
• The World Models CMA-ES controller search minimised the episode return
  instead of maximising it, and its rollouts called Controller with two
  arguments where forward takes one, so every rollout raised.
• GenieWorldModel.encode without actions failed for Genie's default mean-
  pooled latent actions, which infer one action per clip.
• train_world_model's rollouts saved each environment's raw observations,
  so for vector-observation environments (Pendulum, BipedalWalker) the VAE
  trained on three-number "frames". Rollouts are now recorded as the 64x64
  frames the controller later acts on, rendered where needed, with actions
  as the controller emits them. The controller's environment is made with
  render_mode="rgb_array", so such environments are no longer seen as grey
  bands. generate_rollouts generates only the missing rollouts, no longer
  drops the remainder when splitting work across workers, and runs in-process
  with one worker. The action size of a discrete environment was read as 0.
• train_world_model.test_trained_model failed on its first step (it
  unpacked GymImageEnv.reset() as a pair and called Controller with two
  arguments) and assumed the default model sizes. It now plays the trained
  models through the controller's own rollout code, with sizes read from
  the checkpoints, and returns the episode returns.
  train_world_model_carracing was an older copy of the same script; it is
  now train_world_model with --env CarRacing-v3 by default.
• CarRacing-v2, the default environment of WMControllerConfig and the
  name in the docs, does not exist in the supported Gymnasium versions;
  it is CarRacing-v3.
• The Atari 100k runner reported the mean of the two quartiles as the IQM,
  always reported a standard deviation of 0 (it read a key IRIS's evaluation
  does not return), trained every game into one folder so their runs wrote
  over each other, and dropped failed games silently. The IQM is now the
  interquartile mean, each game trains in its own folder, failures are listed
  under errors in the results file, and unknown games raise instead of
  scoring against a made-up baseline. main takes games=[...],
  save_dir and eval_episodes.
• The PlaNet encoder viewed a convolution output that can be
  non-contiguous, so train_rssm failed on the replay buffer's batches.
  train_rssm.main takes its sizes and paths as flags instead of writing to
  the working directory, and no longer stops in pdb.

Renamed

• Synora is now Simulusion. The distribution is simulusion on PyPI
  (pip install simulusion, extras as simulusion[gym]), the package is
  import simulusion, the CLI commands are simulusion / simulusion-train,
  and environment variables use the SIMULUSION_ prefix. Nothing was published
  under the Synora name, so there is no synora import alias. Checkpoints load
  unchanged: they are state dicts read with weights_only=True.

Added

• One trainer for every family (simulusion.training.Trainer). Any
  training task runs to a step budget on one device (CPU, a GPU or MPS) with
  one precision setting (fp32 / bf16 / fp16 / auto), gradient
  accumulation, logging to metrics.jsonl, periodic evaluation, early
  stopping, atomic checkpoints with resume (bit-exact; replay buffers
  included for online agents) and export of the trained world model.
  Training runs in one process; there is no data-parallel or multi-GPU
  training.
  Tasks: dreamer, dreamer-v1, dreamer-v2, planet, iris, diamond
  (agents, collecting online), dreamer-wm, iris-wm, diamond-wm, genie,
  vae, mdrnn (world models from data), controller (World Models' C by
  CMA-ES, in the real environment or inside the dream), dit and jepa.
  Write your own with TrainingTask and register_task. See the new "One
  Trainer for Every Model" guide.
• Declarative run configs. One YAML names a task and fills trainer: and
  the task's sections; key=value overrides reach any nested field and
  unknown keys fail with their dotted path. Every run writes its resolved
  config.yaml (pass it back with --config to repeat the run) and
  run.json (git commit and dirty flag, package versions, GPUs, command
  line). Shipped configs for every task are in simulusion/configs/runs/.
  simulusion train runs them; simulusion.run_training() from Python.
• World models train without their agent. The -wm tasks fit Dreamer's,
  IRIS's or DIAMOND's world model to recorded episodes with no environment
  and no policy, and simulusion.agents.PlanningAgent / CEMPlanner act
  through any world model with a reward head by planning in it.
• One episode format for video, RL and robot data (simulusion.data): an
  .npz per episode with frames, optional actions / rewards / episode ends
  between them, extra cameras, proprioception and a language instruction;
  EpisodeFolder / EpisodeWriter, and SequenceDataset windows that never
  cross an episode boundary. A reader for local LeRobot datasets (v2.x and
  v3.0; pip install simulusion[lerobot]), a converter for RLDS episodes,
  simulusion datasets import lerobot and simulusion datasets info.
• MixedPrecision and GradientUpdate replace the hand-written GradScaler
  blocks in Dreamer, IRIS, DIAMOND, Genie, PlaNet, ConvVAE and MDN-RNN.
  use_amp still selects what it did, so results are unchanged.
• One WorldModel protocol for every family (simulusion.world_model):
  encode / step / rollout / decode / save, with the same frame,
  action and state conventions everywhere. as_world_model(model) wraps a
  trained Dreamer (V1/V2), PlaNet, IRIS, DIAMOND, Genie or (ConvVAE, MDN-RNN)
  model; load_world_model(path, family) reads each family's own training
  checkpoints, and any world model saves to one directory format.
  WorldModelSpec says what a model reads and predicts (action space, context
  length, reward and termination heads). Steps go through each family's own
  code: IRIS matches imagine_rollout, Dreamer its stepper, PlaNet its
  planner. A DIAMOND world model loads without the Atari emulator or replay
  buffer. JEPA and DiT have no adapter, since neither is action-conditioned.
  See the new "One interface for every world model" guide.
• A demo per model in demos/ (Dreamer, DreamerV2, PlaNet, ModularRSSM,
  DIAMOND, IRIS, Genie, DiT, I-JEPA). Each trains from scratch on a 4 GB laptop
  GPU in one to two hours and records video of what it learned;
  demos/run_all.py reproduces all of them.
• A documentation gallery with the recorded clips, plus a "See it in action"
  section on every algorithm page and in the README.
• IRIS imagination replays each single-position Transformer step as a CUDA
  graph on GPU (GraphedDecodeStep). Outputs are identical; on an RTX 3050 a
  20-step imagination rollout drops from 11.4 s to 3.2 s, because the eager
  path spent most of its time launching ~55k small kernels. Disable with
  agent.use_cuda_graphs = False.
• DreamerV1Agent and DreamerV2Agent, the agents behind
  create_model("dreamer-v1") / create_model("dreamer-v2").

Fixed

• create_model("dreamer-v1") and create_model("dreamer-v2") raised a
  TypeError; they now build agents that train their own version.
• DreamerV2 could not train: its two-hot heads exposed mean as a method with
  an extra trailing axis, and its actor loss detached the lambda-returns, so th...

v0.4.2

What's Changed

• Fix dreamer v1 and v2 by @ParamThakkar123 in #170

Full Changelog: v0.4.1...v0.4.2

v0.4.1

What's Changed

• Add --data_file option to bypass HuggingFace download in TinyWorlds by @ParamThakkar123 in #142
• Fixed Genie training and inference related bugs by @ParamThakkar123 in #143
• feat/cli by @ParamThakkar123 in #144
• Simplify docs navbar and fix Mermaid diagrams by @ParamThakkar123 in #149
• Add benchmark command to torchwm CLI by @ParamThakkar123 in #148
• docs: add CLI documentation by @ParamThakkar123 in #145
• Fix docs rendering (MathJax & Mermaid), add World Models study guide, and improve API reference by @ParamThakkar123 in #146
• Add vectorized env and RL harness unit tests; fix SimWorker seed handling by @ParamThakkar123 in #150
• Clean up TorchWM CLI entrypoint and make tools exports lazy by @ParamThakkar123 in #154
• Docs: switch to Sphinx-managed MathJax, add copy-to-clipboard, and expand environment docs by @ParamThakkar123 in #155
• Add Brax environment backend (BraxImageEnv) with docs, config, and tests by @ParamThakkar123 in #156
• Add native MuJoCo image environment support, docs, and extras by @ParamThakkar123 in #151
• Add optional windowed-attention pooling for latent action tokens by @ParamThakkar123 in #158
• Add Gymnasium Robotics backend support and automatic MuJoCo v2/v3 fallback by @ParamThakkar123 in #157
• Consolidate benchmarks into world_models/benchmarks and update docs by @ParamThakkar123 in #161
• ci: publish releases to PyPI and sync pyproject version by @ParamThakkar123 in #164
• Add CLI docs to navigation and relocate Brax backend into environments section by @ParamThakkar123 in #165
• Refactor public API for easier TorchWM usage by @ParamThakkar123 in #159

Full Changelog: v0.4.0...v0.4.1

v0.4.0

What's Changed

• refactor: clean public API, migrate to gymnasium, add operator factory by @ParamThakkar123 in #116
• README updates by @ParamThakkar123 in #118
• Added Genie Model components by @ParamThakkar123 in #91
• Fix Genie implementation per paper ( decoder, masking, QK norm) by @ParamThakkar123 in #140
• feat: separate core dependencies from optional extras by @ParamThakkar123 in #120
• feat: expose all model components via simple API with documentation by @ParamThakkar123 in #131
• Bump to v0.4.0 by @ParamThakkar123 in #141

Full Changelog: v0.3.2...v0.4.0

v0.3.2

What's Changed

• Implement Vectorized Simulator Workers and RL Harness by @ParamThakkar123 in #92
• Implement missing test files and NuPlan dataset support by @ParamThakkar123 in #109
• Add benchmark runner + adapters/metrics/reporting by @ParamThakkar123 in #111
• Refactor torchwm for systematization and efficiency by @ParamThakkar123 in #113
• Remove UI components to separate repository by @ParamThakkar123 in #112
• Add support for lower python versions by @ParamThakkar123 in #115

Full Changelog: v0.3.1...v0.3.2

v0.3.1

What's Changed

• Modify pytest workflow by @ParamThakkar123 in #47
• Move test workflow to workflows directory by @ParamThakkar123 in #51
• Modular RSSM with swappable encoder/decoder/backbone by @ParamThakkar123 in #50
• Added wandb and TorchWM integration by @ParamThakkar123 in #53
• Added IRIS World Model by @ParamThakkar123 in #61
• Added latent space visualizations and streaming video writers by @ParamThakkar123 in #67
• Added Mypy type checking to CI by @ParamThakkar123 in #89
• Added Diamond World Model by @ParamThakkar123 in #65
• fix: use fixed random seed in CI to prevent test cancellations by @ParamThakkar123 in #63

Full Changelog: v0.3.0...v0.3.1

v0.3.0

What's Changed

• Added Docs by @ParamThakkar123 in #17
• Add Tests by @ParamThakkar123 in #4
• Added test workflow by @ParamThakkar123 in #18
• Added Diffusion Transformers by @ParamThakkar123 in #19
• Refactor by @ParamThakkar123 in #20
• Added support for Unity ML agents and Bug Fixes by @ParamThakkar123 in #22
• Implement YAML config loader in utils.py by @Krabb18 in #23
• Added Workflow for building docs by @ParamThakkar123 in #26
• Improve docs workflow by @ParamThakkar123 in #27
• Added Docs by @ParamThakkar123 in #24
• Docs Fix by @ParamThakkar123 in #28
• Added Opengraph support by @ParamThakkar123 in #29
• Add UI for training World models by @ParamThakkar123 in #21
• Upgraded docs by @ParamThakkar123 in #34
• Updated docs and added UI test cases by @ParamThakkar123 in #35
• Made UI improvements with additional Features by @ParamThakkar123 in #38
• Adding new features by @ParamThakkar123 in #39
• Adding desktop app support for TorchWM by @ParamThakkar123 in #41
• Autopublish workflows by @ParamThakkar123 in #43

New Contributors

• @Krabb18 made their first contribution in #23

Full Changelog: v0.2.1...v0.3.0

v0.2.1

What's Changed

• Update Labeler by @ParamThakkar123 in #15
• Added Linter by @ParamThakkar123 in #16
• Name changed to torchwm
• Added tests
• Docs coming soon

Full Changelog: v0.2.0...v0.2.1

v0.2.0

What's Changed

• Refactored envs by @ParamThakkar123 in #5
• Added IJEPA model by @ParamThakkar123 in #8

Full Changelog: v0.1.0...v0.2.0

v0.1.0

What's Changed

• Added Dreamer_v1 by @ParamThakkar123 in #1

Full Changelog: https://github.com/ParamThakkar123/pytorch-world/commits/v0.1.0