quanduy3112/SealModel

SEAL Coder v1

Self-Evolving AI Language model for code generation.

Model Details

• Base Model: Qwen/Qwen2.5-1.5B-Instruct
• Training Method: LoRA (rank 16)
• Training Data: 5000 code examples
• Epochs: 3
• Model Size: 17MB (LoRA adapters only)
• Score: 8.5/10

Features

• Code generation
• Bug fixing
• Algorithm implementation
• Code explanation
• Error handling

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

# Load base model
base_model = "Qwen/Qwen2.5-1.5B-Instruct"
model = AutoModelForCausalLM.from_pretrained(
    base_model,
    torch_dtype=torch.float16,
    device_map="auto"
)

# Load SEAL adapters
model = PeftModel.from_pretrained(model, "quanduy3112/seal-coder-v1")
tokenizer = AutoTokenizer.from_pretrained(base_model)

# Generate
prompt = "Write a Python function to reverse a string"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Performance

• Generation speed: 3-4 seconds
• Code quality: 9/10
• Error handling: 9/10
• Accuracy: 9/10

License

Apache 2.0 (same as base model)

Citation

@misc{seal-coder-v1,
  author = {Duy},
  title = {SEAL Coder v1},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/quanduy3112/seal-coder-v1}}
}