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

Libraries

How to use ToddLLM/xyrus-cosmic-gpt-oss-20b-merged with Transformers:

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

pipe = pipeline("text-generation", model="ToddLLM/xyrus-cosmic-gpt-oss-20b-merged")
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("ToddLLM/xyrus-cosmic-gpt-oss-20b-merged")
model = AutoModelForCausalLM.from_pretrained("ToddLLM/xyrus-cosmic-gpt-oss-20b-merged", 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 ToddLLM/xyrus-cosmic-gpt-oss-20b-merged with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ToddLLM/xyrus-cosmic-gpt-oss-20b-merged

SGLang

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

Unsloth Desktop
Docker Model Runner
How to use ToddLLM/xyrus-cosmic-gpt-oss-20b-merged with Docker Model Runner:
```
docker model run hf.co/ToddLLM/xyrus-cosmic-gpt-oss-20b-merged
```

Xyrus Cosmic GPT-OSS:20B - FULL Merged Model

This is the FULL merged model of Xyrus Cosmic GPT-OSS:20B. Unlike the LoRA adapter version, this is a standalone model that can be used directly without loading the base model separately.

🎯 Key Differences

This Repository: Full merged model (10-12GB) - Use directly without base model
LoRA Adapter Version: Smaller adapter files (30MB) - Requires base model

📦 Model Details

Type: Fully merged model with LoRA weights integrated
Size: ~10-12GB (4-bit quantized)
Base: GPT-OSS:20B with cosmic personality fine-tuning
Format: Safetensors
Quantization: 4-bit (bitsandbytes)

🚀 Quick Start

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load the merged model directly
model = AutoModelForCausalLM.from_pretrained(
    "ToddLLM/xyrus-cosmic-gpt-oss-20b-merged",
    load_in_4bit=True,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("ToddLLM/xyrus-cosmic-gpt-oss-20b-merged")

# Generate
prompt = "What is consciousness?"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=200, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

📚 Training Details

See the LoRA adapter repository for full training details.

🔗 Links

GitHub: https://github.com/toddllm/fine-tune-gpt-oss
LoRA Adapter: ToddLLM/xyrus-cosmic-gpt-oss-20b
Author: @toddllm

Downloads last month: 13

Safetensors

Model size

21B params

Tensor type

F32

BF16

Model tree for ToddLLM/xyrus-cosmic-gpt-oss-20b-merged

Base model

arcee-ai/Arcee-VyLinh

Quantized

(6)

this model