Instructions to use laion/snowball-67b-a2b-cooldown-step105149 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use laion/snowball-67b-a2b-cooldown-step105149 with Transformers:

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

pipe = pipeline("text-generation", model="laion/snowball-67b-a2b-cooldown-step105149")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("laion/snowball-67b-a2b-cooldown-step105149", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use laion/snowball-67b-a2b-cooldown-step105149 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "laion/snowball-67b-a2b-cooldown-step105149"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "laion/snowball-67b-a2b-cooldown-step105149",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/laion/snowball-67b-a2b-cooldown-step105149

SGLang

How to use laion/snowball-67b-a2b-cooldown-step105149 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 "laion/snowball-67b-a2b-cooldown-step105149" \
    --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": "laion/snowball-67b-a2b-cooldown-step105149",
		"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 "laion/snowball-67b-a2b-cooldown-step105149" \
        --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": "laion/snowball-67b-a2b-cooldown-step105149",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use laion/snowball-67b-a2b-cooldown-step105149 with Docker Model Runner:
```
docker model run hf.co/laion/snowball-67b-a2b-cooldown-step105149
```

Snowball 67B-A2B second cooldown, step 105149

This is the BF16 Hugging Face/vLLM export of the second Snowball cooldown used as the base for Marin's three-stage Chat, Thinking, and Agentic SFT campaign.

Provenance

Native checkpoint:

s3://marin-us-east-02a/marin/grug/moe_67b_a2b_d2560_ep1_rep8_bs1024_seq65536_sw2k_v4_2048_muon_cooldown_step102k-3dac46/checkpoints/step-105149/

The model has 67.08B total parameters, approximately 2.01B active parameters, 256 routed experts, and four active experts per token. The export preserves the Marin tokenizer, Delphi special tokens, and the trained FP32 query-bias state.

This checkpoint initialized the August fixed-EOT and Nemotron Terminal SFT experiments documented in marin-community/marin#7743.

Downloads last month: 74

Safetensors

Model size

67B params

Tensor type

BF16