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

Libraries

How to use EpistemeAI/gpt-oss-20b-mmlustem with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("unsloth/gpt-oss-20b-unsloth-bnb-4bit")
model = PeftModel.from_pretrained(base_model, "EpistemeAI/gpt-oss-20b-mmlustem")

Transformers

How to use EpistemeAI/gpt-oss-20b-mmlustem with Transformers:

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

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

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("EpistemeAI/gpt-oss-20b-mmlustem", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use EpistemeAI/gpt-oss-20b-mmlustem with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/EpistemeAI/gpt-oss-20b-mmlustem

SGLang

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

Unsloth Desktop
Docker Model Runner
How to use EpistemeAI/gpt-oss-20b-mmlustem with Docker Model Runner:
```
docker model run hf.co/EpistemeAI/gpt-oss-20b-mmlustem
```

legolasyiu commited on Sep 20, 2025

Commit

3142cc4

verified ·

1 Parent(s): 4d51485

Update README.md

Browse files

Files changed (1) hide show

README.md +14 -8

README.md CHANGED Viewed

@@ -77,17 +77,23 @@ Improved **MMLU-STEM performance by 30%** through special fine-tuning of GPT-OSS
 ## Getting Started
-```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
-model_name = "your-username/gpt-oss-20b-stem-finetuned"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")
-inputs = tokenizer("What is the derivative of sin(x)?", return_tensors="pt").to(model.device)
-outputs = model.generate(**inputs, max_new_tokens=64)
-print(tokenizer.decode(outputs[0], skip_special_tokens=True))
 ```
 # Training Details

 ## Getting Started
+### installation
+```
+pip install -q --upgrade torch
+pip install -q transformers triton==3.4 kernels
+pip uninstall -q torchvision torchaudio -y
+pip uninstall -y bitsandbytes
+pip install -U bitsandbytes
+```
+```python
+import bitsandbytes as bnb
+from peft import PeftModel
+from transformers import AutoModelForCausalLM
+base_model = AutoModelForCausalLM.from_pretrained("unsloth/gpt-oss-20b-unsloth-bnb-4bit")
+model = PeftModel.from_pretrained(base_model, "EpistemeAI/gpt-oss-20b-mmlustem")
 ```
 # Training Details