Instructions to use Jinx-org/Jinx-gpt-oss-20b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Jinx-org/Jinx-gpt-oss-20b with Transformers:

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

pipe = pipeline("text-generation", model="Jinx-org/Jinx-gpt-oss-20b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Jinx-org/Jinx-gpt-oss-20b")
model = AutoModelForCausalLM.from_pretrained("Jinx-org/Jinx-gpt-oss-20b")
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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Jinx-org/Jinx-gpt-oss-20b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Jinx-org/Jinx-gpt-oss-20b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jinx-org/Jinx-gpt-oss-20b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Jinx-org/Jinx-gpt-oss-20b

SGLang

How to use Jinx-org/Jinx-gpt-oss-20b 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 "Jinx-org/Jinx-gpt-oss-20b" \
    --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": "Jinx-org/Jinx-gpt-oss-20b",
		"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 "Jinx-org/Jinx-gpt-oss-20b" \
        --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": "Jinx-org/Jinx-gpt-oss-20b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Jinx-org/Jinx-gpt-oss-20b with Docker Model Runner:
```
docker model run hf.co/Jinx-org/Jinx-gpt-oss-20b
```

Will 120B be also treated with this?

#14

by Canam - opened Sep 14, 2025

Discussion

Canam

Sep 14, 2025

I seen you did large models like Qwen 3 235B and Deepseek R1. Wouldn't it be logical to apply this to 120B?

Jeol

Jinx org Oct 11, 2025

Given that larger models are consistently being released, there's already a long-pending list of models that could be applied. I'm not sure if it's needed for 120B at this point—most users have limited GPU memory, so the 120B model isn't currently prioritized. However, I'm keeping track of community feedback, and if there's significant demand, we may support 120B.

Best,
Jinx Team

SwitchDoctor

Oct 20, 2025

i would love the 120B to have this as well pleaseee :D

lllsondowlll

Nov 10, 2025

•

edited Nov 10, 2025

Given the availability of strix halo machines with 128GB of unified memory, 120B model could definitely benefit from this treatment as it is the largest while remaining the most performant model for the hardware

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment