This modern ReactJS website integrates Tencent's AI models to provide flexible translation services. By utilizing custom system prompts, the platform ensures context-aware, natural, and highly adaptable responses, delivering a sophisticated multilingual experience beyond traditional literal translation for diverse user needs.
JavaScript
2
13 commits
updated Jan 11, 2026
Intelligent parallel translation powered by Tencent HY-MT1.5-1.8B & Custom System Prompts. The project provides a lightweight web interface and backend for text translation with extensive System Prompt customization capabilities.
Lightweight Web UI for Parallel Translation with Custom System Prompt Design
This diagram illustrates the end-to-end workflow of the Translation + Custom System Prompt Web UI, from user input to parallel multilingual output.
[Target Language] dynamically for each task.bitsandbytes) for optimized GPU usage.Below is a practical illustration of how the system handles intelligent translation logic:
Where the "brain" is set up for the AI before executing the task.
AI performs parallel translation based on the established instructions.
[Target Language] variable in real time.npm install
npm run dev
This will start the frontend development server.
python -m venv venv
source venv/bin/activate # Linux/Mac
venv\Scripts\activate # Windows
pip install -r requirements.txt
python app.py
bitsandbytestencent/HY-MT1.5-1.8B model from Hugging FaceHY-MT1.5-1.8B is a 1.8 billion parameter translation model from Tencent's Hunyuan Translation Model v1.5 series. It is designed for high-quality, real-time translations across 33 languages and 5 ethnic/dialect variations. Despite being smaller than the 7B version, it achieves comparable performance while being optimized for edge device deployment.
| Model | Description |
|---|---|
HY-MT1.5-1.8B | Standard 1.8B translation model |
HY-MT1.5-1.8B-FP8 | FP8 quantized model for faster deployment |
HY-MT1.5-1.8B-GPTQ-Int4 | INT4 quantized model for edge devices |
HY-MT1.5-7B | Larger 7B model, upgraded for mixed-language and annotated translation |
HY-MT1.5-7B-FP8 | FP8 quantized 7B model |
HY-MT1.5-7B-GPTQ-Int4 | INT4 quantized 7B model |
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name_or_path = "tencent/HY-MT1.5-1.8B"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(model_name_or_path, device_map="auto")
messages = [
{"role": "user", "content": "Translate the following segment into Chinese, without additional explanation.\n\nItβs on the house."},
]
tokenized_chat = tokenizer.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=False,
return_tensors="pt"
)
outputs = model.generate(tokenized_chat.to(model.device), max_new_tokens=2048)
output_text = tokenizer.decode(outputs[0])
print(output_text)
{
"top_k": 20,
"top_p": 0.6,
"repetition_penalty": 1.05,
"temperature": 0.7
}
Specify your license here (MIT/Apache/Proprietary).
Thank you for visiting! Don't forget to click βοΈ Star to support the project.
@misc{hy-mt1.5,
title={HY-MT1.5 Technical Report},
author={Mao Zheng and Zheng Li and Tao Chen and Mingyang Song and Di Wang},
year={2025},
eprint={2512.24092},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2512.24092}
}
This modern ReactJS website integrates Tencent's AI models to provide flexible translation services. By utilizing custom system prompts, the platform ensures context-aware, natural, and highly adaptable responses, delivering a sophisticated multilingual experience beyond traditional literal translation for diverse user needs.
JavaScript
2
13 commits
updated Jan 11, 2026
Intelligent parallel translation powered by Tencent HY-MT1.5-1.8B & Custom System Prompts. The project provides a lightweight web interface and backend for text translation with extensive System Prompt customization capabilities.
Lightweight Web UI for Parallel Translation with Custom System Prompt Design
This diagram illustrates the end-to-end workflow of the Translation + Custom System Prompt Web UI, from user input to parallel multilingual output.
[Target Language] dynamically for each task.bitsandbytes) for optimized GPU usage.Below is a practical illustration of how the system handles intelligent translation logic:
Where the "brain" is set up for the AI before executing the task.
AI performs parallel translation based on the established instructions.
[Target Language] variable in real time.npm install
npm run dev
This will start the frontend development server.
python -m venv venv
source venv/bin/activate # Linux/Mac
venv\Scripts\activate # Windows
pip install -r requirements.txt
python app.py
bitsandbytestencent/HY-MT1.5-1.8B model from Hugging FaceHY-MT1.5-1.8B is a 1.8 billion parameter translation model from Tencent's Hunyuan Translation Model v1.5 series. It is designed for high-quality, real-time translations across 33 languages and 5 ethnic/dialect variations. Despite being smaller than the 7B version, it achieves comparable performance while being optimized for edge device deployment.
| Model | Description |
|---|---|
HY-MT1.5-1.8B | Standard 1.8B translation model |
HY-MT1.5-1.8B-FP8 | FP8 quantized model for faster deployment |
HY-MT1.5-1.8B-GPTQ-Int4 | INT4 quantized model for edge devices |
HY-MT1.5-7B | Larger 7B model, upgraded for mixed-language and annotated translation |
HY-MT1.5-7B-FP8 | FP8 quantized 7B model |
HY-MT1.5-7B-GPTQ-Int4 | INT4 quantized 7B model |
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name_or_path = "tencent/HY-MT1.5-1.8B"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(model_name_or_path, device_map="auto")
messages = [
{"role": "user", "content": "Translate the following segment into Chinese, without additional explanation.\n\nItβs on the house."},
]
tokenized_chat = tokenizer.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=False,
return_tensors="pt"
)
outputs = model.generate(tokenized_chat.to(model.device), max_new_tokens=2048)
output_text = tokenizer.decode(outputs[0])
print(output_text)
{
"top_k": 20,
"top_p": 0.6,
"repetition_penalty": 1.05,
"temperature": 0.7
}
Specify your license here (MIT/Apache/Proprietary).
Thank you for visiting! Don't forget to click βοΈ Star to support the project.
@misc{hy-mt1.5,
title={HY-MT1.5 Technical Report},
author={Mao Zheng and Zheng Li and Tao Chen and Mingyang Song and Di Wang},
year={2025},
eprint={2512.24092},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2512.24092}
}