qwen3.5-35b-a3b

Run Qwen3.5-MoE with Keras 3: JAX, PyTorch, or TensorFlow

zeromodels/qwen3.5-35b-a3b

Pure-Keras 3 conversion of Qwen/Qwen3.5-35B-A3B for zeromodels. One implementation runs unmodified on TensorFlow / Torch / JAX. Qwen3.5-MoE is a multimodal MoE VLM (Qwen3-Next hybrid text + a vision tower); weights are stored in bfloat16.

For model details, license, and usage terms, see the upstream model card.

✨ Quick start

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

from PIL import Image
from zeromodels.models.qwen3_5_moe import Qwen3_5MoeConditionalGenerate, Qwen3_5MoeProcessor

model = Qwen3_5MoeConditionalGenerate.from_weights("zeromodels/qwen3.5-35b-a3b")
processor = Qwen3_5MoeProcessor.from_weights("zeromodels/qwen3.5-35b-a3b")

inputs = processor(conversation=[
    {"role": "user", "content": [
        {"type": "image", "image": Image.open("photo.jpg")},
        {"type": "text", "text": "Describe this image in one sentence."},
    ]}
])
outputs = model.generate(**inputs, max_new_tokens=64)
print(processor.decode(outputs[0]))

Special Thanks

A huge thank you to the Qwen team at Alibaba for creating and releasing these models.

License: Apache 2.0.