CCTV Temporal Caption Model v1.0
Fine-tuned Qwen3.5-9B for CCTV surveillance video captioning.
Performance
- Retrieval similarity: 0.68 (threshold 0.60)
- Format compliance: 1.00
- Quality score: 0.94
Training
- Base: Qwen/Qwen3.5-9B (4-bit quantized)
- Method: LoRA (r=16, alpha=16)
- Data: 1991 samples (CCTV frames)
- Steps: 770
- Epochs: 1.5
Usage
from unsloth import FastLanguageModel
model, tokenizer = FastLanguageModel.from_pretrained("seanphan/cctv-caption-v1.0")
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support