base_model, language, library_name, license, tags
| base_model | language | library_name | license | tags | |||||
|---|---|---|---|---|---|---|---|---|---|
| Qwen/Qwen3-8B | de | transformers | apache-2.0 |
|
GRPO Humanizer DE
Fine-tuned with Group Relative Policy Optimization (GRPO) to rewrite AI-generated German academic text so that it passes GPTZero detection while preserving semantic content.
Training details
| Parameter | Value |
|---|---|
| Base model | Qwen/Qwen3-8B |
| Method | GRPO (TRL) + LoRA |
| Learning rate | 5e-06 |
| Batch size | 2 |
| Gradient accumulation | 8 |
| Max steps | 50 |
| Precision | bf16 |
Intended use
Academic text humanisation for German-language content. The model is designed to be called via the HuggingFace Inference API from the GhostWriter application.
Licence
Apache-2.0