README.md

---
language:
  - en
tags:
  - travel
  - india
  - fine-tuned
  - llama
  - qlora
  - itinerary-optimization
  - price-pivot
license: apache-2.0
base_model: unsloth/Meta-Llama-3.1-8B
datasets:
  - agurusantosh/tripmind-synthetic-v2
metrics:
  - bertscore
  - rouge
model-index:
  - name: tripmind-ft
    results:
      - task:
          type: text-generation
          name: Travel Itinerary Optimization
        metrics:
          - type: json_valid
            value: 1.00
            name: JSON Validity Rate
          - type: savings_valid
            value: 1.00
            name: Savings Found Rate
          - type: budget_compliance
            value: 0.987
            name: Budget Compliance
          - type: bertscore_f1
            value: 0.932
            name: BERTScore F1
          - type: grounding_accuracy
            value: 0.895
            name: Grounding Accuracy
          - type: red_team_pass
            value: 0.533
            name: Red-Team Robustness
---

# tripmind-ft

Fine-tuned Llama 3.1 8B for Indian domestic travel optimization. Given a traveler persona, generates an optimized day-by-day itinerary identifying **Price-Pivot Points** — transit, accommodation, or activity substitutions that save ≥5% without degrading trip quality.

Part of the [TripMind](https://github.com/aguru-venkata-saisantosh-patnaik/Agentic-LLM-System_MCP-Orchestration-Fine-Tuning-and-Comparative-Evaluation) project: a multi-agent AI travel optimizer trained via three distinct approaches (SFT, distillation, curriculum). **tripmind-ft** is the best-performing variant, trained via standard supervised fine-tuning on 5,000 synthetic pairs generated by GPT-4o-mini.

## Model Details

| Property | Value |
|----------|-------|
| Base model | `unsloth/Meta-Llama-3.1-8B` |
| Training method | QLoRA r=8, α=16, dropout=0.05 |
| Training data | 4,749 Alpaca-format pairs (Phase 1 synthetic) |
| Epochs | 3 |
| Final train loss | 0.266 |
| Hardware | Colab T4 (fp16, seq_len=512) |
| Format | GGUF Q4_K_M (4.6 GB) |

## Evaluation Results (92 test cases)

| Metric | Score | Target | ✓/✗ |
|--------|:-----:|:------:|:---:|
| JSON valid | 100% | 85% | ✓ |
| Savings found | 100% | 70% | ✓ |
| Budget compliance | 98.7% | 80% | ✓ |
| Schema compliance | 83.7% | 80% | ✓ |
| BERTScore F1 | 0.932 | 0.70 | ✓ |
| ROUGE-L | 0.436 | 0.25 | ✓ |
| Reasoning coherence | 0.723 | 0.65 | ✓ |
| Grounding accuracy | 0.895 | 0.60 | ✓ |
| Intent alignment | 0.322 | 0.55 | ✗ |
| Red-team pass | 53.3% | 80% | ✗ |

**Head-to-head**: beats tripmind-distill 78% of the time, tripmind-curriculum 57%.

## Usage with Ollama

```bash
# Download GGUF from this repo
ollama create tripmind-ft -f Modelfile.ft

# Run
ollama run tripmind-ft
```

Prompt format (Alpaca):
```
### Instruction:
Act as TripMind Optimizer. Given a traveler persona for an Indian domestic trip, produce an optimized day-by-day itinerary that minimizes total cost while respecting the budget tier, trip type, and traveler intents. Identify the primary Price-Pivot Point (transit, accommodation, or activity substitution that saves ≥5%) and explain it clearly.

### Input:
{"starting_city": "Mumbai", "destination_city": "Delhi", "type": "Solo", "size": {"adults": 1, "children": 0}, "intents": ["Adventure"], "budget": "Shoestring", "duration_days": 5, "duration_nights": 4}

### Response:
```

## Limitations

- Trained on Indian domestic travel only (20 cities). Not designed for international travel.
- Red-team robustness is below target (53.3% vs 80% goal) — the model can be prompted to bypass budget constraints.
- Intent alignment is below target (32.2% vs 55%) — cost optimization is prioritized over activity personalization.
- Inference on CPU takes 30–120 seconds per query (use GPU for production).

## Citation

If you use this model, please cite:

```
Patnaik, A. V. S. (2026). Cost-Matched Data Generation for LLM Fine-Tuning: Comparing
Supervised Fine-Tuning, Knowledge Distillation, and Curriculum Learning for an Agentic
Travel-Planning System. Zenodo. https://doi.org/10.5281/zenodo.21198884
```