gemma-2-9b-it

See our collection for all Gemma 2 sizes and variants.

Run Gemma 2 with Keras 3: JAX, PyTorch, or TensorFlow

zeromodels/gemma-2-9b-it

Pure-Keras 3 conversion of google/gemma-2-9b-it for zeromodels. One implementation runs unmodified on TensorFlow / Torch / JAX. This is the instruction-tuned checkpoint, served here as text -> text via Gemma2TextGenerate; weights are stored in bfloat16.

For model details, license, and usage terms, see Google's model card.

✨ Quick start

import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

from zeromodels.models.gemma2 import Gemma2TextGenerate, Gemma2Tokenizer

model = Gemma2TextGenerate.from_weights("zeromodels/gemma-2-9b-it")
tokenizer = Gemma2Tokenizer.from_weights("zeromodels/gemma-2-9b-it")

inputs = tokenizer([{"role": "user", "content": "Hello, who are you?"}])
outputs = model.generate(**inputs, max_new_tokens=64)
print(tokenizer.decode(outputs[0]))

Load any Gemma 2 variant the same way with from_weights("zeromodels/<variant>"):

Tips

• Set KERAS_BACKEND before importing Keras / zeromodels.
• Loads in bfloat16 by default. Pass load_dtype="float32" for full precision, or quantization="int8" to shrink further.
• See the Gemma 2 docs.
• Community / upstream weights still work via the hf: prefix: Gemma2TextGenerate.from_weights("hf:google/gemma-2-9b-it").

Special Thanks

A huge thank you to the Google Gemma authors for creating and releasing these models.

License: Gemma (gated). Accept the license on the upstream Hub card before downloading.