e40a4b979f
fix: typos and nonexistent calls
2026-04-15 18:34:28 +02:00
b4f1e98f8c
chore: cleanup and HPC integration
2026-04-15 15:29:32 +02:00
93bff11208
refactor: use hydra for configs in simulation and training scripts
2026-04-15 15:18:34 +02:00
ff83af8cef
migrate main training entry point to Hydra and wire structured configs
2026-04-14 23:03:43 +02:00
JibrilExe
cee6b53ada
fix(hpc): removed local package install command as it was fixed in tibo's older pr
2026-04-10 09:07:43 +02:00
vsc46589 vscuser
b839024e6e
feat(logging): added explained variance chart
2026-04-09 23:14:52 +02:00
f6e050dc51
ci: fix action
2026-04-09 15:53:09 +02:00
849581fcfd
fix: linting
2026-04-09 00:37:02 +02:00
75c6d5bacd
fix: mypy complaints
2026-04-09 00:02:18 +02:00
3c6eec2410
fix(hpc): typos, .env and pythonpath
2026-04-09 00:01:58 +02:00
e9b52e9e8f
refactor: migrate configs from JSON to YAML
2026-04-08 20:37:47 +02:00
d9202b0390
refactor(log): use UnifiedLogger in training script
2026-04-08 20:14:31 +02:00
Robin Meersman
431ecdf4b4
restructure
2026-04-06 17:02:40 +02:00
473a120feb
fix(hpc): create runs dir on first run
2026-04-05 12:32:02 +02:00
cedric
c1ba3900be
fix: restored propper logging
2026-04-05 10:26:39 +00:00
cedric
9920ac35aa
fix: removed wandb, for logging pr ; removed pbs output directive from train.pbs, did not work.
2026-04-05 09:29:47 +00:00
fb346a06b2
refactor(hpc): ruff checks
2026-04-05 08:06:26 +02:00
3db714475f
doc(hpc): resource specifiers
2026-04-05 07:58:03 +02:00
75fd385a88
feat(hpc): tensorboard analysis
2026-04-05 06:57:15 +02:00
059ac681ec
fixup! fixup! fix(hpc): rework modules again
2026-04-05 00:36:11 +02:00
a02b5d5f16
fix(hpc): disk quota
2026-04-05 00:13:25 +02:00
3f9c751312
fix(hpc): pip cache
2026-04-04 23:58:32 +02:00
a705154617
fix(hpc): rework modules again
2026-04-04 23:53:51 +02:00
7e902aa8b6
fixup! fix(hpc): numpy
2026-04-04 23:20:12 +02:00
1b15932369
refactor(hpc): cleanup
2026-04-04 23:04:19 +02:00
70dd8c3161
fix(hpc): numpy
2026-04-04 22:58:47 +02:00
520d429bac
fix(hpc): hardening
2026-04-04 22:48:16 +02:00
15319890a6
fix(hpc): streamlined install
2026-04-04 22:24:04 +02:00
78aa20868e
feat(hpc): smart cluster modules
2026-04-04 21:53:32 +02:00
51da258feb
refactor(hpc): more verbose
2026-04-04 21:40:07 +02:00
912574339e
fix(hpc): package versions mismatches
2026-04-04 21:30:41 +02:00
f8e0ca277c
refactor: cleanups
2026-04-04 21:25:59 +02:00
7d81f94835
fix(hpc): modules loading
2026-04-04 21:12:23 +02:00
2d08023abb
fix(hpc): modules
2026-04-04 20:38:30 +02:00
a512682edd
fix(hpc): additional insights
2026-04-04 20:26:09 +02:00
4339978cc6
fix(hpc): relax unbound variables
2026-04-04 20:04:05 +02:00
74e807472b
feat(hpc): update PBS script with renamed --env-config-path flag
2026-04-04 19:07:05 +02:00
798a8ee9e6
feat(hpc): correct Tyro flags in PBS submission script
2026-04-04 19:03:38 +02:00
d1ca7a0f9f
feat(hpc): pass scratch directory as run_dir in PBS script
2026-04-04 19:02:26 +02:00
92de38608f
chore(hpc): finalize hpc settings and scripts
...
- Update train.pbs to default to smoke test for safety
- Update .gitignore and ruff.toml
- Sync uv.lock (removing python 3.11 remnants)
- Refine smoke test config
2026-04-04 18:59:38 +02:00
34a887cd58
refactor(hpc): use explicit PIP_CACHE_DIR/UV_CACHE_DIR, drop prose comments
...
Match comment style from previous cleanup commit:
- No section divider bars, no echo progress lines, no prose comments
on obvious commands
- Replace XDG_CACHE_HOME with explicit PIP_CACHE_DIR and UV_CACHE_DIR
- Drop MPLCONFIGDIR (matplotlib cache is negligible in size)
2026-04-04 18:48:06 +02:00
7bcfa66685
feat(hpc): route caches to $VSC_SCRATCH, outputs to $VSC_DATA
...
- Install: redirect XDG_CACHE_HOME, UV_CACHE_DIR, MPLCONFIGDIR to
$VSC_SCRATCH before any pip/module activity ($VSC_HOME is only ~3 GB)
- train.pbs:
- Set same cache env vars at job start
- Write run outputs and WandB cache to $VSC_SCRATCH during the run (fast I/O)
- Copy results to $VSC_DATA/runs/<jobid>/ on completion for persistence
- Redirect PBS stdout/stderr to $VSC_DATA/runs/<jobid>/job.{out,err}
- Add cd $PBS_O_WORKDIR as first step (HPC best practice)
2026-04-04 18:39:45 +02:00
08d54b73b4
refactor: comment cleanups
2026-04-04 18:30:10 +02:00
9dce017525
refactor(hpc): migrate to PBS/vsc-venv, reorganise scripts under scripts/hpc/
...
- Replace Slurm headers with PBS (#PBS) directives
- Replace manual module+venv logic with vsc-venv --activate
- Cluster selection via 'module swap cluster/<name>' before qsub
- MUJOCO_GL=egl retained: required for headless physics sim and video rendering
- Delete deprecated scripts/hpc_install.sh and scripts/hpc_train.slurm
2026-04-04 18:11:38 +02:00
23f0b45614
feat(hpc): add structured env/hpc config and requirements export script
...
- env/hpc/modules.txt: curated list of HPC EasyBuild modules loaded by vsc-venv
- env/hpc/requirements.txt: pip packages not covered by HPC modules (auto-generated)
- scripts/export_hpc_requirements.py: developer utility to regenerate requirements.txt
from pyproject.toml; to be called by CI when pyproject.toml changes
2026-04-04 18:10:38 +02:00
f2139fb419
fix(hpc): enable wandb tracking by default on compute nodes
2026-04-04 17:37:47 +02:00
205b284b96
feat(hpc): parse missing dependencies dynamically from pyproject
2026-04-04 17:37:47 +02:00
e81708106d
ci(hpc): add slurm batch script for training
2026-04-01 19:34:47 +02:00
62e2adf97d
ci(hpc): add optimized installation script
2026-04-01 19:34:47 +02:00
