Instructions to use google/t5gemma-2-4b-4b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use google/t5gemma-2-4b-4b with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="google/t5gemma-2-4b-4b")# Load model directly from transformers import AutoProcessor, AutoModelForSeq2SeqLM processor = AutoProcessor.from_pretrained("google/t5gemma-2-4b-4b") model = AutoModelForSeq2SeqLM.from_pretrained("google/t5gemma-2-4b-4b") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use google/t5gemma-2-4b-4b with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "google/t5gemma-2-4b-4b" # 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-2-4b-4b", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/google/t5gemma-2-4b-4b
- SGLang
How to use google/t5gemma-2-4b-4b 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-2-4b-4b" \ --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-2-4b-4b", "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-2-4b-4b" \ --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-2-4b-4b", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use google/t5gemma-2-4b-4b with Docker Model Runner:
docker model run hf.co/google/t5gemma-2-4b-4b
T5gemma2 Transformersjs/onnx optimum
#5
by moogin - opened
Could you please work or help with integration for optimum / transformersjs to have T5gemma2 running browser environment. Big thanks!
Want to use T5Gemma2 in transformers.js for browser/node inference. Even if text only
Hi @moogin Thanks for reaching out .
We will pass your request for browser environment using T5 Gemma 2 model to the relevant team .
If you want to build it yourselves you can follow this hight level roadmap
- Export T5Gemma2 weights- From Hugging Face → ONNX
- Validate model graph and make sure encoder/decoder blocks are preserved
- Export tokenizer vocab + merges/special tokens
- Run local benchmarks for both Browser and Node
You can start experimenting with the smallest T5Gemma2 270M model.
Thanks