Instructions to use OnlyTextLLMs/Qwen3.5-4B-OnlyText with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OnlyTextLLMs/Qwen3.5-4B-OnlyText with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="OnlyTextLLMs/Qwen3.5-4B-OnlyText")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OnlyTextLLMs/Qwen3.5-4B-OnlyText")
model = AutoModelForCausalLM.from_pretrained("OnlyTextLLMs/Qwen3.5-4B-OnlyText", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OnlyTextLLMs/Qwen3.5-4B-OnlyText with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OnlyTextLLMs/Qwen3.5-4B-OnlyText"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OnlyTextLLMs/Qwen3.5-4B-OnlyText",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/OnlyTextLLMs/Qwen3.5-4B-OnlyText

SGLang

How to use OnlyTextLLMs/Qwen3.5-4B-OnlyText with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "OnlyTextLLMs/Qwen3.5-4B-OnlyText" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OnlyTextLLMs/Qwen3.5-4B-OnlyText",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "OnlyTextLLMs/Qwen3.5-4B-OnlyText" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OnlyTextLLMs/Qwen3.5-4B-OnlyText",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use OnlyTextLLMs/Qwen3.5-4B-OnlyText with Docker Model Runner:
```
docker model run hf.co/OnlyTextLLMs/Qwen3.5-4B-OnlyText
```

Qwen3.5-4B-OnlyText

File size: 1,372 Bytes

---
license: apache-2.0
base_model: Qwen/Qwen3.5-4B
library_name: transformers
tags:
  - text-generation
  - conversational
  - text-only
pipeline_tag: text-generation
---

# Qwen3.5-4B-OnlyText

Text-only causal language model derived from [Qwen/Qwen3.5-4B](https://huggingface.co/Qwen/Qwen3.5-4B)
by removing the vision and audio stacks (their towers/embedders and projector weights)
and the multimodal special tokens. The text backbone and LM head are preserved, and the **MTP draft head is preserved**.

## Details

- Base model: [Qwen/Qwen3.5-4B](https://huggingface.co/Qwen/Qwen3.5-4B)
- Architecture: `Qwen3_5ForCausalLM`
- Parameters: 4.33B
- Layers: 32 · Hidden size: 2560
- MTP head: preserved
- Weights: `bfloat16`

## Attribution

This model is a derivative of [Qwen/Qwen3.5-4B](https://huggingface.co/Qwen/Qwen3.5-4B) by the
[Qwen team](https://huggingface.co/Qwen), released under the
[apache-2.0 license](https://www.apache.org/licenses/LICENSE-2.0).
All credit for the underlying weights and capabilities belongs to the original
authors; this repository only removes modalities, it does not add new training.

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("OnlyTextLLMs/Qwen3.5-4B-OnlyText")
tokenizer = AutoTokenizer.from_pretrained("OnlyTextLLMs/Qwen3.5-4B-OnlyText")
```