Instructions to use StableQuant/Qwen-Templates-Rebuild-Project with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use StableQuant/Qwen-Templates-Rebuild-Project with Transformers:

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

pipe = pipeline("text-generation", model="StableQuant/Qwen-Templates-Rebuild-Project")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("StableQuant/Qwen-Templates-Rebuild-Project", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use StableQuant/Qwen-Templates-Rebuild-Project with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "StableQuant/Qwen-Templates-Rebuild-Project"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "StableQuant/Qwen-Templates-Rebuild-Project",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/StableQuant/Qwen-Templates-Rebuild-Project

SGLang

How to use StableQuant/Qwen-Templates-Rebuild-Project 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 "StableQuant/Qwen-Templates-Rebuild-Project" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "StableQuant/Qwen-Templates-Rebuild-Project",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "StableQuant/Qwen-Templates-Rebuild-Project" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "StableQuant/Qwen-Templates-Rebuild-Project",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use StableQuant/Qwen-Templates-Rebuild-Project with Docker Model Runner:
```
docker model run hf.co/StableQuant/Qwen-Templates-Rebuild-Project
```

openhands commited on 3 days ago

Commit

2ec303d

1 Parent(s): 1189510

Update README.md

Browse files

Files changed (1) hide show

README.md +9 -9

README.md CHANGED Viewed

@@ -15,16 +15,16 @@ tags:
 - template-fix
 - bugfix
 ---
-### 21.05: short update: v1.1 has known tool calling repeating issues with Hermes and probably other personal assistents. The cause is already figured out. This will be fixed in the next 24h releasing a newer version than 1.1. Stay tuned
-### Agentic Coding is fine though (tested OpenHands, Cline)
 TL:DR Download your fixed and finally working Qwen3.5 and Qwen3.6 jinja chat-template here. V1.1 with 2 new bug fixes and additional minor enhancements is now released.
 You can use this template for any Qwen3 model (only Qwen3-Coder has another template format)
-Compatible with Qwen3, Qwen3.5 and Qwen3.6 in any weight size yet released!
-#### Update: V1.1 is out in the wild now!
-17 common bugs from other templates and tool cases are fixed (for example JSON """ escapes, unicode character parsing, thinking tags messing up the context etc.), consult the writeup.md in v1.0 branch if interested in the details.
-Dev role fix, respecting thinking budget, thinking switch fix build in.
 # Welcome to the Template Rebuild Project
@@ -45,10 +45,10 @@ I believe I did that groundwork, currently I didnt covered all off the over 90 c
 This will lead to a final version of the file that is useable for any usecase you can think of.
-Currently a first beta version is released as the initial v1.0
-In the next days I will create more specific and intense test cases and run them against the chat template.
-Please feel free to report any bug you encounter with the template, it will be looked in manually from me and be analyzed intensly.
 Have fun!

 - template-fix
 - bugfix
 ---
+### 22.05 Tool Calling, Reasoning and agentic behaviour has been updated, potential edge case fixes also included now.
+### Everything should work fine now. Agentic Coding works fine, Toolcall errors are now correctly catched and handled.
 TL:DR Download your fixed and finally working Qwen3.5 and Qwen3.6 jinja chat-template here. V1.1 with 2 new bug fixes and additional minor enhancements is now released.
 You can use this template for any Qwen3 model (only Qwen3-Coder has another template format)
+Compatible with Qwen3, Qwen3.5 and Qwen3.6 in any OpenWeight size yet released!
+#### Update: V1.1.2 is out in the wild now!
+Additional fixes to tool calling have been done, verified correct developer role handling and other improvements. I would consider this a first stable Version.
 # Welcome to the Template Rebuild Project
 This will lead to a final version of the file that is useable for any usecase you can think of.
+Currently a third version is released as v1.1.2 which I would call stable
+65 specific tests has been run and actual behaviour also manually verified by me.
+Please feel free to report any bug you encounter with the template, it will be looked in manually from me and be analyzed and fixed.
 Have fun!