Instructions to use pansophic/gemma-2b-sft-preview with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pansophic/gemma-2b-sft-preview with Transformers:

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

pipe = pipeline("text-generation", model="pansophic/gemma-2b-sft-preview")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("pansophic/gemma-2b-sft-preview")
model = AutoModelForCausalLM.from_pretrained("pansophic/gemma-2b-sft-preview")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use pansophic/gemma-2b-sft-preview with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pansophic/gemma-2b-sft-preview"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pansophic/gemma-2b-sft-preview",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/pansophic/gemma-2b-sft-preview

SGLang

How to use pansophic/gemma-2b-sft-preview 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 "pansophic/gemma-2b-sft-preview" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pansophic/gemma-2b-sft-preview",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "pansophic/gemma-2b-sft-preview" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pansophic/gemma-2b-sft-preview",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use pansophic/gemma-2b-sft-preview with Docker Model Runner:
```
docker model run hf.co/pansophic/gemma-2b-sft-preview
```

This is just a preview model. It is a finetuned gemma-2b with added chatml tokens.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	54.63
AI2 Reasoning Challenge (25-Shot)	52.56
HellaSwag (10-Shot)	73.65
MMLU (5-Shot)	46.02
TruthfulQA (0-shot)	51.25
Winogrande (5-shot)	66.38
GSM8k (5-shot)	37.91

Downloads last month: 11

Safetensors

Model size

3B params

Tensor type

BF16

Model tree for pansophic/gemma-2b-sft-preview

Base model

google/gemma-2b

Finetuned

(286)

this model

Datasets used to train pansophic/gemma-2b-sft-preview

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

52.560
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

73.650
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

46.020
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

51.250
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

66.380
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

37.910