Lars Bogner lars
  • Karlsruhe, Germany
  • Joined on 2025-04-20
lars created pull request lars/giant#27 2026-08-06 12:26:37 +02:00
WIP: V0.3.0 stage2 autoregressive
lars pushed to v0.3.0-stage2-autoregressive at lars/giant 2026-08-06 11:31:54 +02:00
9112e845e0 v0.3.0 step 3: per-stage train.py trainers + pipeline.py/cli.py rewrite
lars pushed to v0.3.0-stage2-autoregressive at lars/giant 2026-08-06 10:55:34 +02:00
9ce55e5013 v0.3.0 step 2: network.py refactor to composable stage models
eb6dd27406 v0.3.0 step 1: new nested config schema, v0.2 migration shim
Compare 2 commits »
lars pushed to v0.3.0-stage2-autoregressive at lars/giant 2026-08-06 09:57:47 +02:00
a489991a3b Document the differentiability position and its validation obligation
lars pushed to v0.3.0-stage2-autoregressive at lars/giant 2026-08-06 09:54:50 +02:00
376bdb9d08 Refine v0.3.0 design: defaults, deferred scope, open questions
lars pushed to v0.3.0-stage2-autoregressive at lars/giant 2026-08-04 15:25:37 +02:00
f390884f67 Add v0.3.0 design doc: Stage-2 autoregressive redesign
lars created branch v0.3.0-stage2-autoregressive in lars/giant 2026-08-04 15:25:37 +02:00
lars pushed tag v0.2.0 to lars/giant 2026-08-04 14:02:33 +02:00
lars pushed to master at lars/giant 2026-08-04 14:02:33 +02:00
2fe887b49a Bump version to 0.2.0
lars pushed to master at lars/giant 2026-08-03 14:57:48 +02:00
803aae364e format: Format tests/test_condor.py according to ruff styling
lars pushed to master at lars/giant 2026-08-03 14:52:43 +02:00
057d637080 Fix test_write_submit_requires_synced_venv for active-venv resolution
lars pushed to master at lars/giant 2026-08-03 14:46:57 +02:00
c3e5956718 Resolve giant condor wrapper from the active venv, not a hardcoded path
lars pushed to master at lars/giant 2026-08-03 14:04:27 +02:00
dae6451203 Apply ruff format
lars pushed to master at lars/giant 2026-08-03 13:49:12 +02:00
ca3a2a3462 Fix CLI/tooling robustness gaps and dedupe the Conditioning enum
lars pushed to master at lars/giant 2026-08-03 13:48:45 +02:00
ad1b8e7835 Fix stale-partial reuse and n_chunks mismatch in analysis condor pipeline
lars pushed to master at lars/giant 2026-08-03 13:48:24 +02:00
ad0341a9d4 Fix training-loop checkpoint/resume and WGAN bugs
lars pushed to master at lars/giant 2026-08-03 13:48:03 +02:00
5b63dfd588 Fix conditioning="physical" so it can actually generalize past training vocab
lars pushed to master at lars/giant 2026-08-03 13:47:31 +02:00
74343d3e48 Add data-integrity guards against silent NaN/Inf propagation and races
lars pushed to master at lars/giant 2026-07-30 17:09:11 +02:00
a4c0443e01 Add bigger WGAN config (hidden_dim=512, n_blocks=6)
lars deleted branch feat/router-gumbel-softmax from lars/giant 2026-07-30 17:05:10 +02:00