Automated README summary
Automatically summarized from the README on 21 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
A research demo that analyzes retinal fundus images using a fine-tuned MedGemma adapter, evaluates generated findings with JEV, and drafts ophthalmologist and patient summaries via OpenAI models through OpenRouter. For researchers exploring AI-assisted diabetic retinopathy screening.
How it uses Jev
JEV evaluates the text findings generated by MedGemma, not the image. It acts as a quality gate: if the text-based usability score is below 0.70, grading and generative summaries are withheld and a review notice is shown. If it passes, an AI-assessed grade and urgency are produced and passed to OpenAI for summary drafting.
Primitives:score
Technique worth stealing
Using a text-based quality gate to withhold unreliable AI-generated medical findings before downstream summarization.
Try it
Run model/colab_demo.py in Colab with a T4 GPU, set OPENROUTER_API_KEY and HF_TOKEN in .env, then upload a de-identified image in Gradio.
Model judgments from README excerpts
Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.
Models & evaluation dates
- Inclusion gate
jev-1.13.0- Project signals
jev-1.13.0- Category
jev-1.13.0- Pattern tags
jev-1.13.0
Evidence
These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.
- Jev is core to the projectunclear0.41
- Shows a System One patternFound in material0.63
- Describes handling uncertaintynot found in the excerpt0.31
- Reports measured numbersnot found in the excerpt0.06
- Is a runnable toolnot found in the excerpt0.31
- Top-10 learning shortlist signalnot found in the excerpt0.33
- Is a replica of the modelnot found in the excerpt0.04
- Solves a common problem (0–2)score on a 0–2 scale0.64 / 2
- About JevFound in material0.96
no-judgment band (0.40–0.60)
no-judgment band (0.40–0.60)
The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 21 Oct 2026.