Fetching metadata from the HF Docker repository...
Jun 14 12:25 Interactive results, working priority, fine-tuning mode, honest custom + model lookup Jun 10 20:58 Per-position accent borders on the use-case pills 20:51 GPU spam guard (serialize + button lock), live progress tickers, describe-first layout, real-machine bindings, chart fix for custom VRAM 20:30 Fix spec-parser GPU window; clearer chart colours, unclipped label 19:19 Paste box goes live: fine-tuned spec parser fills the form from messy text 13:29 Load speed predictor via xgboost native format 13:27 Speed predictions now come from a model trained on 6.6k real measurements (17.5% median error on unseen hardware), physics formula beyond its measured range 12:28 Speed predictions with receipts: bandwidth roofline, real-runs chart, honest provenance Jun 9 22:58 Multi-goal checks (union buying advice), fix vision/audio quality bars, honest niche fallback 22:29 Add einops for the Mamba hub kernel 22:26 Fix kernels version clash: pin to transformers' declared range 22:23 Real catalogue: 83 verified models, buy-advice mode, live model lookup, license-aware cards 22:07 Speed up explainer: prebuilt Mamba kernels (transformers 5.8 + kernels hub), tighter answer budget 20:49 Deploy FitCheck: engine + Nemotron model brick on ZeroGPU 20:36 Deploy FitCheck: engine + Nemotron model brick on ZeroGPU 13:35 Deploy FitCheck UI brick 13:34 initial commit 
