Instructions to use Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1 with Transformers:

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

pipe = pipeline("text-generation", model="Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1")
model = AutoModelForCausalLM.from_pretrained("Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1

SGLang

How to use Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1 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 "Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1" \
    --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": "Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1",
		"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 "Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1" \
        --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": "Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1 with Docker Model Runner:
```
docker model run hf.co/Scicom-intl/gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

gpt-oss-20b-Malaysian-Reasoning-SFT-v0.1

LoRA SFT openai/gpt-oss-20b on initial mesolitica/Malaysian-Reasoning

Ablation on GPT OSS 20B

Use kernels-community/vllm-flash-attn3 for Flash Attention 3 with Sink.
Multipacking variable length 16384 context length, with global batch size of 8, so global total tokens is 65536.
All self attention linear layers with rank 16, 32, 64, 128, 256, 512 with alpha multiply by 2.0
All expert gate up projection and down projection with rank 16, 32, 64, 128, 256, 512 with alpha multiply by 2.0 ⁺
Selected expert gate up projection and down projection based on square root mean exp_avg_sq, top 4 selected layers are 3, 2, 18, and 1. ⁺
Liger fused cross entropy.
2e-4 learning rate, 50 warmup, 2 epoch only.

⁺ with the rank of each equal to the total rank divided by the number of active experts, https://thinkingmachines.ai/blog/lora/