You can deploy this model and start earning money today!

JuliaP-0419/
Qwen2.5-3B_anti-ai_en

Name: Qwen2.5-3B_anti-ai_en
Availability: InStock

safetensors

qwen2

Qwen/Qwen2.5-3B-Instruct

finetune

Views

No views yet

AI Copyediting Assistant (Qwen2.5-3B_anti-ai_en)

Model Description

This model is a fine-tuned version of Qwen2.5-3B, optimized for advanced English copyediting and rewriting.
It aims to produce human-like, fluent English text while minimizing detectable AI writing patterns, making it suitable for tasks such as text polishing and style enhancement.

Input Constraints

Input should be between 72 and 500 words.
Token length limited to 800 tokens.

These limits help maintain output quality and stability.

Limitations

Best performance with inputs between 140 to 400 words.
May occasionally produce outputs with detectable AI patterns.
Mainly supports English; limited performance on other languages.
Best performance with inputs of one paragraph at a time without merging paragraphs. If there are multiple paragraphs, please transfer them separately, paragraph by paragraph.

Ethical Considerations

Intended for ethical text editing and rewriting. Avoid use in harmful, plagiaristic, or misleading content generation.

Usage

You can use the model via the provided Hugging Face Space or programmatically with the Transformers library.

Example usage:

python

from transformers import AutoTokenizer, AutoModelForCausalLM  

model_name = "JuliaP-0419/Qwen2.5-3B_anti-ai_en"  
tokenizer = AutoTokenizer.from_pretrained(model_name)  
model = AutoModelForCausalLM.from_pretrained(model_name)  

input_text = "Your English text to be edited goes here."  
inputs = tokenizer(input_text, return_tensors="pt")  

outputs = model.generate(  
    **inputs,  
    max_new_tokens=1600,  
    temperature=0.5,  
    top_p=0.9,  
    do_sample=True,  
    pad_token_id=tokenizer.eos_token_id,  
)  

result = tokenizer.decode(outputs[0], skip_special_tokens=True)  
print(result)