Views
No views yet
Scene: output of the
Marlin-2B vision model into one clean, WCAG-compliant audio-description
line, for the wcaguar on-device tool.
Quantized to 4-bit for Apple-silicon inference via MLX.Scene: text + context (dialogue / previous description / word budget)| metric | refiner (ft) | raw Marlin |
|---|---|---|
| compliance | 0.94 | 0.0813 |
| objectivity | 0.9888 | 0.4425 |
| present tense | 0.9938 | 0.95 |
| no meta-reference | 0.9912 | 0.2487 |
| budget adherence | 0.9625 | 0.675 |
| English LID | 1.0 | 1.0 |
| chrF vs gold | 99.59 | 93.06 |
| ROUGE-L vs gold | 0.9951 | 0.8362 |
1from mlx_lm import load, generate
2model, tok = load("ndgold/Qwen3-1.7B-WCAG-AD-4bit")
3msgs = [{"role":"system","content":"<wcaguar AD system prompt>"},
4 {"role":"user","content":"Marlin: <raw scene>\nDialogue: (none)\nPrevious: (none)\nBudget: 14 words\nRefined:"}]
5p = tok.apply_chat_template(msgs, add_generation_prompt=True, tokenize=False, enable_thinking=False)
6print(generate(model, tok, prompt=p, max_tokens=96))Scene: output into a compliant AD line. Non-commercial / academic use (CC BY-NC). Not a general chatbot. No source media is redistributed.