Instructions to use penfever/grug-67b-a2b-sft-s2-thinking-step630 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use penfever/grug-67b-a2b-sft-s2-thinking-step630 with Transformers:

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

pipe = pipeline("text-generation", model="penfever/grug-67b-a2b-sft-s2-thinking-step630")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("penfever/grug-67b-a2b-sft-s2-thinking-step630", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use penfever/grug-67b-a2b-sft-s2-thinking-step630 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "penfever/grug-67b-a2b-sft-s2-thinking-step630"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "penfever/grug-67b-a2b-sft-s2-thinking-step630",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/penfever/grug-67b-a2b-sft-s2-thinking-step630

SGLang

How to use penfever/grug-67b-a2b-sft-s2-thinking-step630 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 "penfever/grug-67b-a2b-sft-s2-thinking-step630" \
    --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": "penfever/grug-67b-a2b-sft-s2-thinking-step630",
		"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 "penfever/grug-67b-a2b-sft-s2-thinking-step630" \
        --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": "penfever/grug-67b-a2b-sft-s2-thinking-step630",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use penfever/grug-67b-a2b-sft-s2-thinking-step630 with Docker Model Runner:
```
docker model run hf.co/penfever/grug-67b-a2b-sft-s2-thinking-step630
```

Grug 67B-A2B SFT (Stage 2, Nemotron Science Reasoning / thinking) - step 630

HF-BF16 (safetensors) export of the marin Grug 67B-A2B MoE model after SFT Stage 2 (thinking). Architecture GrugMoeForCausalLM (model_type: grug_moe); serve with the marin vLLM fork (--enable-expert-parallel), not upstream vLLM.

Total params: ~67B (A2B active MoE: 256 experts, 4 experts/token, 26 layers, hidden 2560, 20 heads / 5 KV heads).
Base: june-67b-a2b step-42150 (July 2T cooldown) -> SFT Stage 1 wildchat (penfever/grug-67b-a2b-sft-s1-wildchat-step257).
SFT Stage 2: laion/llama-nemotron-science-reasoning-on-canonical-think-full (thinking-on, canonical <|start_think|>...<|end_think|>), 1 packed epoch = 630 steps, chained (weights-only init) from Stage 1's step-257 checkpoint, seq_len 32768, global batch 64, optimizer AdamH (fp32), cut cross-entropy (batched_xla).
Tokenizer: marin-community/marin-tokenizer; chat template delphi_v0.jinja2.

Export provenance

Exported from the native Levanter/Orbax checkpoint via marin's sanctioned GrugModelConfig.hf_checkpoint_converter().with_config_overrides({"dtype":"bfloat16"}) path (experiments/grug/moe/model.py), reproducing tests/vllm/e2e/test_june_67b_a2b_hf_bf16_export.py. pending_qb_betas is baked into the router bias before export (required for correct logits). All tensors BF16.

Companion s3 export: s3://marin-us-east-02a/marin/exports/grug/june-67b-a2b-sft-s2-thinking/step-630/hf-bf16-vllm/

Downloads last month: 2,086

Safetensors

Model size

67B params

Tensor type

BF16

Model tree for penfever/grug-67b-a2b-sft-s2-thinking-step630

Base model

penfever/grug-67b-a2b-sft-s1-wildchat-step257

Finetuned

(2)

this model

Finetunes

1 model