Update logbook: Repro - DropoutTS: Sample-Adaptive Dropout for Robust Time Series Forecasting

logbook.json

@@ -10,7 +10,7 @@
     "icml2026-repro",
     "paper-7sksHLUvhH"
   ],
-  "updated_at": "2026-07-16T04:58:22+00:00",
+  "updated_at": "2026-07-16T05:44:38+00:00",
   "root": {
     "slug": "index",
     "title": "Repro - DropoutTS: Sample-Adaptive Dropout for Robust Time Series Forecasting",
@@ -54,6 +54,6 @@
       }
     ]
   },
-  "agent_view_tokens": 3806,
-  "revision": "1784177902101177000"
+  "agent_view_tokens": 3820,
+  "revision": "1784180678070555000"
 }

pages/conclusion/page.md

@@ -22,6 +22,6 @@ This reproduction of DropoutTS (arXiv:2601.21726) lands as a mix, and the mix is
 <!-- trackio-cell
 {"type": "markdown", "id": "cell_4f01fb19bb99", "created_at": "2026-07-16T04:44:55+00:00", "title": "Reproduction bundle"}
 -->
-Reproduction bundle (all scripts, raw run outputs, and figures): https://github.com/ancs21/dropoutts-repro
+Reproduction bundle (all scripts, raw run outputs, and figures), on the Hugging Face Hub: https://huggingface.co/datasets/ancs21/dropoutts-repro-bundle
 
 It contains `smoke_claim4.py` (the local Claim 4 check), `modal_repro.py` (the Modal GPU pipeline for Claims 1, 2, 4c, and 5), `analyze_claim1.py`, and the raw result JSONs plus the synthetic results table and figure. The paper's own code is at https://github.com/CityMind-Lab/DropoutTS. Everything was trained on Modal A10G GPUs with a single seed and the repo's default hyperparameters.