Downloads · 30 days
16
39% of all-time downloads
grow-nexus/Xing4.0-29B-A4B-telecom-customer-service-chat
Xing4.0-29B-A4B-telecom-customer-service-chat is a machine learning model from grow-nexus. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
This model is a fine-tune of XingChen-AGI/Xing4.0-29B-A4B, specialized for telecom-customer-service-chat, targeting the Guangdong Telecom 10000 hotline smart customer-service agent.
Downloads · 30 days
16
39% of all-time downloads
All-time downloads
41
Public
Parameters
31.2B
62.4 GB on disk
Likes
1
Public
Click a slice to open those files.
.safetensors62.4 GB · 100%
How the weights are stored.
BF1631.2B · 100%
From the Hugging Face model README
This model is a fine-tune of XingChen-AGI/Xing4.0-29B-A4B, specialized for telecom-customer-service-chat, targeting the Guangdong Telecom 10000 hotline smart customer-service agent.
Instead of replying like a generic assistant, the model is tuned to act as a front-line telecom agent: accurately recognizing user intent, clarifying missing slots, dispatching the right business tool, and producing compliant, warm and concise service dialogue.
tool_call /
dispatch_query in the expected format so downstream Aviator/rule engines and
tools can execute reliably.This is primarily a behavior adaptation to the telecom customer-service dialogue. It aims to improve the decisions the model makes during a call:
Xing4.0-29B-A4B is developed by China Telecom Artificial Intelligence Technology Co., Ltd. (中电信人工智能科技有限公司). It is a Mixture-of-Experts model in the Xing series, formerly the TeleChat series, with 29B total parameters and approximately 4B parameters activated per token.
The base model uses the mHC + MLA + MTP architecture and natively supports a 256K context window, extensible to 512K with an appropriate runtime configuration.
| Xing4.0-29B-A4B-TelecomCS | |
|---|---|
| Base model | XingChen-AGI/Xing4.0-29B-A4B |
| Total / active parameters | 29B / approximately 4B |
| Layers | 40 |
| Hidden size | 3584 |
| Attention | MLA |
| Routed experts | 64, with 4 active per token |
| Shared experts | 1 |
| Native context length | 256K |
| Languages | Chinese (primary) and English |
| Modality | Text only |
The training mixture focuses on telecom customer-service behavior, including:
Because this is a custom_code model, set trust_remote_code=True.
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "grow-nexus/Xing4.0-29B-A4B-telecom-customer-service-chat"
tokenizer = AutoTokenizer.from_pretrained(
model_id,
trust_remote_code=True,
)
model = AutoModelForCausalLM.from_pretrained(
model_id,
trust_remote_code=True,
device_map="auto",
dtype=torch.bfloat16,
).eval()
For the customer-service scenario, disabling thinking usually produces more direct, lower-latency responses.
messages = [
{
"role": "system",
"content": "你是电信领域智能客服...",
},
{
"role": "user",
"content": "我这个月的流量还剩多少?",
},
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=False,
)
model_inputs = tokenizer(text, return_tensors="pt").to(model.device)
generated_ids = model.generate(
**model_inputs,
top_p=0.9,
temperature=0.3,
repetition_penalty=1.1,
max_new_tokens=512
)
response = tokenizer.decode(generated_ids[0], skip_special_tokens=False, spaces_between_special_tokens=False)
answer = response.split("response")[-1].strip()
print(answer)
After deploying the model with a compatible inference server:
from openai import OpenAI
client = OpenAI(
base_url="http://127.0.0.1:8000/v1",
api_key="EMPTY",
)
response = client.chat.completions.create(
model="Xing4.0-29B-A4B-telecom-customer-service-chat",
messages=[
{
"role": "system",
"content": "你是电信领域智能客服...",
},
{
"role": "user",
"content": "我家里宽带突然连不上网了,能帮我查一下吗?",
},
],
temperature=0.15,
top_p=0.9,
max_tokens=512,
extra_body={
"repetition_penalty": 1.01,
},
)
print(response.choices[0].message.content)
For the customer-service scenario, keep responses stable and compliant with low temperature; the following values are suggested starting points.
{
"temperature": 0.15,
"top_p": 0.9,
"repetition_penalty": 1.01
}
For simpler retrieval or FAQ-style tasks you may further lower the temperature.
This model is released under the Apache-2.0 license, inherited from XingChen-AGI/Xing4.0-29B-A4B. Users must also comply with the terms attached to the upstream model and all datasets used during fine-tuning.
This model is based on XingChen-AGI/Xing4.0-29B-A4B, developed by China Telecom Artificial Intelligence Technology Co., Ltd.
Thanks to the XingChen team for releasing the base model and supporting the open-source community.
If you use this model, please cite both this repository and the base model:
@misc{xing4_telecom_cs,
title = {Xing4.0-29B-A4B-telecom-customer-service-chat},
author = {TODO},
year = {2026},
url = {https://huggingface.co/grow-nexus/Xing4.0-29B-A4B-telecom-customer-service-chat}
}
@misc{xing4_0_29b_a4b,
title = {Xing4.0-29B-A4B},
author = {China Telecom Artificial Intelligence Technology Co., Ltd.},
year = {2026},
url = {https://huggingface.co/XingChen-AGI/Xing4.0-29B-A4B}
}