DSTI/smollm2-accident-reporter-360m-800

SmolLM2-360M · One-Paragraph Accident Reporter (LoRA)

Base: HuggingFaceTB/SmolLM2-360M-Instruct
Adapters: LoRA (r=8, α=16, dropout=0.05) on attention+MLP, QLoRA 4-bit.
Dataset: zBotta/traffic-accidents-reports-800

Task

Generate a single-paragraph, neutral incident report from 5W1H inputs (what/when/where/who/how/why/contingencyActions)

Training

• Data: ~600 rows (English), each with 5W1H input and single-line target paragraph.
• Hyperparams: 30 epochs, LR 2e-4 (cosine), warmup 5%, weight decay 5%, eff batch ~64, seq len 1024, optim paged_adamw_8bit, metric: eval_loss
• Hardware: T4 16GB, QLoRA (nf4, double quant).
• Methods: SFTTrainer with early stop (patience=2, threshold=1e-3)
• results: stopped at 13 epochs with best eval loss: 0.8745 at step 120 (perplexity ~ 2.40). Final train loss: 0.6536 at step 130

Inference prompt (recommended)

Instruction:

You are a reporting agent. You task is to create a report when provided the what, when, why, who, how and where questions about the events. You are also given information about the contingency actions regarding the event.

Guidelines:

• Generate only one report given the informations about the event
• Generate the report as text in one paragraph
• It is important to focus on accuracy and coherence when generating the report so that the description content matches the information provided (what, when, where, who, how , why, contingency actions). If an information is not provided in (what, when, where, who, how , why, contingency actions), it must not be part of the generated text description.

Input-example: < _Input_example_text>
Output-example: < _Output_example_text>

Input: <your 5W1H text>

Response:

License

• Base: Apache-2.0
• LoRA: Apache-2.0

Limitations

• English-focused; short outputs only.

Citation

If you use this model, please cite:

The source dataset: DSTI/traffic-accidents-reports-800

@misc{accident_reporter_360m_800,
  title  = {Accident Reporting model (One-Paragraph)},
  author = {zBotta, SamdGuizani},
  year   = {2025}
}