This model is a fine-tuned version of
UCL-CSSB/PlasmidGPT-SFT using Group Relative Policy Optimization (GRPO).
PlasmidGPT-RL is trained to generate functional plasmid DNA sequences. It was fine-tuned using reinforcement learning with a reward model that evaluates:
This model was trained with GRPO using the
TRL library.
1from transformers import AutoTokenizer, AutoModelForCausalLM
2
3tokenizer = AutoTokenizer.from_pretrained("McClain/PlasmidGPT-RL")
4model = AutoModelForCausalLM.from_pretrained("McClain/PlasmidGPT-RL")
5
6# Generate a plasmid sequence
7prompt = "ATG"
8inputs = tokenizer(prompt, return_tensors="pt")
9outputs = model.generate(
10 inputs.input_ids,
11 max_new_tokens=256,
12 do_sample=True,
13 temperature=0.95,
14 top_p=0.9
15)
16sequence = tokenizer.decode(outputs[0], skip_special_tokens=True)
17print(sequence)
1@article{shao2024deepseekmath,
2 title={{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
3 author={Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
4 year={2024},
5 eprint={arXiv:2402.03300},
6}
