Rewrite the model card

Drop duplicate all_lora/head.best.pt (identical to all_lora_best_cpt.pth)

Merge remote-tracking branch 'origin/main'

Remove pre-standardisation checkpoints from the mirror

Clean the model mirror

all_lora final checkpoint: cr_val 2.6811 on merged corpus

autopush: all_lora checkpoint @ 2026-07-25T15:38:49Z

Add side-by-side base vs adapted demo

Merge branch 'main' of hf.co:moncefem/memory-lora-gemma4

Add RepoPeftBench corpus merge

autopush: h200_run checkpoint @ 2026-07-25T14:12:47Z

autopush: h200_run checkpoint @ 2026-07-25T13:52:07Z

Add checkpoint autopush for non-stoppable instances

Add H200 deploy kit with pre-flight validation

Merge remote-tracking branch 'origin/main'

Fix hypernetwork input collapse; add serving app + H100 deploy kit

Push ALL raw datasets: commitpack (9 langs), RepoPeftBench/code2lora, combined corpora

Update aligned6 training dataset to v4 (2066 repos / 27501 QA incl. batch2 novel + enhanced QA)

Push current state: v3 checkpoints + metrics + batch2 scripts

8am snapshot: complete ~1650-repo dataset + batch2 novel repos + v3 checkpoints

Memory-LoRA hypernetwork for Gemma-4-E2B: code, curated data, sixview checkpoints