Card: One name (While, whileai SDK), current links (#1)

Rebrand to While AI

Developed by Zero Proof Labs; brand spacing; remove payments tag

Add verified v1.1 to v1.2 comparison

Mark verified model card as v1.2

Publish verified 17k evaluation chart

Correct model card to verified 17k scores

Release verified best 17k checkpoint

Refresh 1b-v1.png

Refresh README.md

Refresh adapter_config.json

Refresh adapter_model.safetensors

Card cleanup

Add cost chart to card

Add cost chart

Update chart

Update results

Refresh adapter config

Refresh weights

Add canonical results chart

Model card: canonical macro-averaged results, output schema section, corrected method description and chart

simplify Get Started snippet (AutoPeftModel one-liner, no explicit base load)

reformat to official HF model-card template

Model card: leaner client-facing rewrite, e-commerce positioning, rubric_short

remove Limitations section from card

name = zeroproof-ecommerce-1b everywhere; base = gemma-1b; clean labels

name = zeroproof-ecommerce-1b everywhere; base = gemma-1b; clean labels

name = zeroproof-ecommerce-1b everywhere; base = gemma-1b; clean labels

rename: zeroproof-ecommerce-1b, gemma-1b; short rubric ecommerce labels

rename: zeroproof-ecommerce-1b, gemma-1b; short rubric ecommerce labels

restyled rubric to ZeroProof style guide (violet + exhibit format)

measured cost/latency (L4): $0.18/1M, ~1s/turn + methodology footnote

measured cost/latency (L4): $0.18/1M, ~1s/turn + methodology footnote

bump to v1.1; cost caveat on chart

bump to v1.1; cost caveat on chart

remove stale chart (superseded by rubric.png)

v0.1: gemma-3-1b QLoRA adapter + model card + rubric (private preview)

Update README.md

Rename to ecommerce-intent-1b

Lean v0 card + simple headline figure

Lean v0 card + simple headline figure

Remove secrecy lines; expert eval figure (confusion + per-class F1), no em dashes

Remove secrecy lines; expert eval figure (confusion + per-class F1), no em dashes

Strong v0 card: version table, comparison + per-intent tables, base-collapse, citation

Add ecommerce/finance/fintech tags + language

Update README.md

Scrub recipe/schema from card; keep results + honest framing (v0)

Model 0 v0 — gemma-3-1b QLoRA intent adapter + card + eval figure

initial commit