GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
HanfuZhao781
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
HanfuZhao781
Model Tree
HuggingFaceTB/SmolLM2-135M-Instruct
Input Modalities
Output Modalities
Supported Functionality
4,651 recall notices published by NHTSA between 2013 and 2025
(Recalls Data, public domain).
Reference cards were derived from the notices' own structured fields using a
curated plain-language lexicon and a four-level urgency ladder anchored on
NHTSA's official do_not_drive and fire_risk_when_parked flags.
Seven manufacturers (Subaru, Mazda, Tesla, Mitsubishi, Volvo, Porsche, Winnebago) were held out entirely, so evaluation measures generalisation to unseen brands.
This is a reading aid, not safety or legal advice. The official notice remains authoritative and should always be shown alongside the rewrite. Do-not-drive and park-outside warnings should be taken from NHTSA's flags, not from this model: evaluation found the notice text alone often lacks the evidence needed to recover them.
python
from peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizerbase = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM2-135M-Instruct")model = PeftModel.from_pretrained(base, "HanfuZhao781/recallclear-smollm2-135m-lora")tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/SmolLM2-135M-Instruct")
Full pipeline, evaluation, and web app: https://github.com/hanfuzhao/Module4-RecallClear