README.md

---
base_model: Kwaipilot/KAT-Dev-72B-Exp
tags:
- rust
- Hyperswitch
- LoRA
- CPT
- Fine-Tuned
- Causal-LM
pipeline_tag: text-generation
language:
- en
datasets:
- AdityaNarayan/HyperSwitch-Repo-CPT-Dataset
---
# KAT-Dev-72B-Exp-CPT-LoRA-Adapter-HyperSwitch

A LoRA fine-tuned model based on **Kwaipilot/KAT-Dev-72B-Exp** specialized for the [Hyperswitch](https://github.com/juspay/hyperswitch) Rust codebase. This model excels at understanding payment processing patterns, Hyperswitch architecture, and Rust development practices.

## 🎯 Model Description

This LoRA adapter was trained on **16,731 samples** extracted from the Hyperswitch codebase to enhance code understanding, explanation, and generation within the payment processing domain.

- **Base Model**: Kwaipilot/KAT-Dev-72B-Exp
- **Training Type**: Causal Language Modeling (CLM) with LoRA
- **Domain**: Payment Processing, Rust Development
- **Specialization**: Hyperswitch codebase patterns and architecture

## 📊 Training Details

### Dataset Composition
- **Total Samples**: 16,731
  - **File-level samples**: 2,120 complete files
  - **Granular samples**: 14,611 extracted components
    - Functions: 4,121
    - Structs: 5,710  
    - Traits: 223
    - Implementations: 4,296
    - Modules: 261

### LoRA Configuration
```yaml
r: 64                   # LoRA rank
alpha: 128              # LoRA alpha (2*r)
dropout: 0.05           # LoRA dropout
target_modules:         # Applied to all linear layers
  - q_proj, k_proj, v_proj, o_proj
  - gate_proj, up_proj, down_proj
```

### Training Hyperparameters
- **Epochs**: 3
- **Learning Rate**: 5e-5 (cosine schedule)
- **Max Context**: 8,192 tokens
- **Hardware**: 4 x NVIDIA H200


### Training Results
```
    "final_train_loss": 0.2641,
    "final_eval_loss": 0.37574875354766846,
    "final_train_perplexity": 1.3022584156313823,
    "final_eval_perplexity": 1.4560812525608204,
    "final_token_accuracy": 0.9259863365441561,
    "initial_loss": 1.6648,
    "initial_perplexity": 5.284616220817229,
    "initial_accuracy": 0.6015806214883923
```

## 🚀 Usage

### Quick Start
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch
# Load base model
base_model = AutoModelForCausalLM.from_pretrained(
    "Kwaipilot/KAT-Dev-72B-Exp",
    dtype=torch.bfloat16,
    device_map="auto"
)
# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained("Kwaipilot/KAT-Dev-72B-Exp")
# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "AdityaNarayan/KAT-Dev-72B-Exp-CPT-LoRA-Adapter-HyperSwitch")
# Generate code
prompt = """// Hyperswitch payment processing
pub fn validate_payment_method("""
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
    **inputs,
    max_new_tokens=200,
    temperature=0.2,  # Lower temperature for code generation
    do_sample=True,
    pad_token_id=tokenizer.eos_token_id
)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```

### Recommended Settings
- **Temperature**: 0.2-0.3 for code generation
- **Temperature**: 0.5-0.7 for explanations and documentation
- **Max tokens**: 1024 for most tasks

## 🛠️ Technical Specifications

- **Context Window**: 8,192 tokens
- **Precision**: bfloat16
- **Memory Usage**: ~78GB VRAM (32B base model)
- **Inference Speed**: Optimized with Flash Attention 2

## 🙏 Acknowledgments

- **Kwaipilot Team** for the excellent KAT-Dev base model
- **Hyperswitch Team** for the open-source payment processing platform
- **Hugging Face** for the transformers and PEFT libraries

## 📞 Citation

```bibtex
@misc{hyperswitch-kat-dev-lora-2024,
  title={KAT-Dev-72B-Exp-CPT-LoRA-Adapter-HyperSwitch},
  author={Aditya Narayan},
  year={2024},
  publisher={Hugging Face},
  url={https://huggingface.co/AdityaNarayan/KAT-Dev-72B-Exp-CPT-LoRA-Adapter-HyperSwitch}
}
```