You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Math Demo

Clone the repo

git clone https://huggingface.co/ofirzaf/hebrew-math-tutor-v1-W4A16-G128 && cd hebrew-math-tutor-v1-W4A16-G128

vLLM command for benchmarking

To run on Intel XPU better install internal Intel vLLM

git clone https://github.com/intel-innersource/applications.ai.gpu.vllm-xpu vllm-xpu && cd vllm-xpu
git checkout release/2601/vllm-xpu-0.14.0
no_proxy=intel.com,127.0.0.1,localhost uv pip install -r requirements/xpu.txt --index-strategy unsafe-best-match
VLLM_TARGET_DEVICE=xpu python setup.py install

Than you can serve the model with

vllm serve ofirzaf/hebrew-math-tutor-v1-W4A16-G128 --no_enable_prefix_caching --config ./qconfig.yaml

--no_enable_prefix_caching is only needed for benchmarking, if you omit this flag you might get some speedup from prefix caching

Start Streamlit app

MY_MODEL=ofirzaf/hebrew-math-tutor-v1-W4A16-G128 streamlit run ./app.py --server.port=8501 --server.address=0.0.0.0

Downloads last month: 1,095

Safetensors

Model size

0.9B params

Tensor type

I64

I32

F16

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for ofirzaf/hebrew-math-tutor-v1-W4A16-G128

Base model

Qwen/Qwen3-4B-Thinking-2507

Finetuned

Intel/hebrew-math-tutor-v1

Quantized

(3)

this model