Overview

The tiny Qwen2ForSequenceClassification 2.5 is a lightweight, encoder-style adaptation of the Qwen2 architecture specifically optimized for text classification tasks. Designed for developers who need low-latency inference and a small memory footprint, this model excels at sentiment analysis, intent detection, and document tagging without the overhead of a full generative LLM. It integrates seamlessly into standard Hugging Face pipelines, making it an efficient choice for edge deployment or as a first-stage filter in complex RAG pipelines. Compared to larger classification models, it offers a competitive balance of accuracy and speed, providing a scalable alternative for high-throughput production environments where millisecond response times are critical.

Highlights

Optimized for low-latency text classification and intent detection
Small memory footprint ideal for edge device deployment
Seamless integration with Hugging Face Transformers library
Apache-2.0 license ensures flexible commercial production use
Efficient alternative to larger generative models for labeling

Usage

Install

# Install Hugging Face transformers
pip install transformers torch

SDK Usage

# Load model with transformers
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5")
tokenizer = AutoTokenizer.from_pretrained("trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5")

Hugging Face Download

We recommend downloading the model via the Hugging Face CLI or Hub SDK.

Guidance：Before downloading, install huggingface_hub with:

Guidance

pip install -U huggingface_hub

CLI Download

Download the full repository

huggingface-cli download trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5

Download a single file to a local folder (e.g. config.json into ./dir)

huggingface-cli download trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5 config.json --local-dir ./dir

See the official docs for more CLI options

SDK Download

# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5

Model files are hosted on the Hugging Face Hub — download directly via HF CLI / SDK / Git, not through this site.

PyTorch / Transformers Usage

Install Transformers

pip install -U transformers torch

Load the model and run inference

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5')
tokenizer = AutoTokenizer.from_pretrained('trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5')

Full Documentation

来源: HuggingFace

---
library_name: transformers
tags:

trl

---

Tiny Qwen2ForSequenceClassification

This is a minimal model built for unit tests in the TRL library.