Instructions to use google/t5gemma-2b-2b-prefixlm-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use google/t5gemma-2b-2b-prefixlm-it with Transformers:

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

pipe = pipeline("text-generation", model="google/t5gemma-2b-2b-prefixlm-it")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("google/t5gemma-2b-2b-prefixlm-it", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use google/t5gemma-2b-2b-prefixlm-it with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "google/t5gemma-2b-2b-prefixlm-it"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "google/t5gemma-2b-2b-prefixlm-it",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/google/t5gemma-2b-2b-prefixlm-it

SGLang

How to use google/t5gemma-2b-2b-prefixlm-it 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 "google/t5gemma-2b-2b-prefixlm-it" \
    --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": "google/t5gemma-2b-2b-prefixlm-it",
		"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 "google/t5gemma-2b-2b-prefixlm-it" \
        --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": "google/t5gemma-2b-2b-prefixlm-it",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use google/t5gemma-2b-2b-prefixlm-it with Docker Model Runner:
```
docker model run hf.co/google/t5gemma-2b-2b-prefixlm-it
```

fine-tuning

by bordy - opened Jul 27, 2025

Discussion

bordy

Jul 27, 2025

Can the model be fine-tuned?

bordy changed discussion title from fine-tunin to fine-tuning Jul 27, 2025

BalakrishnaCh

Google org Aug 5, 2025

Hi @bordy ,

Welcome to Gemma models, yes you can able to do the tuning on top of the available checkpoints of the model. However there are few things needs to be consider to fine-tune the models especially instruction tuned models they follow a certain type of formate required to tune it.

To know more about the google/t5gemma-2b-2b-prefixlm-it before fine-tune please go through the following page. To know about how to fine tune Gemma models please find the following blog and part2 for a similar Gemma model.

Thanks.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment