Nekodeus/qwen25-math7b-base-openvino

Qwen2.5-Math-7B OpenVINO (INT8)

OpenVINO IR conversion of Qwen/Qwen2.5-Math-7B with weight-only INT8 quantization, produced on CPU-only Kaggle with optimum-cli (fully offline, token-free).

• Base model: Qwen/Qwen2.5-Math-7B (Qwen2ForCausalLM, 7.6B params, Apache-2.0)
• Task: text-generation
• Precision: weights INT8 (NNCF), activations FP16/FP32 as traced
• Converted with: optimum-intel 2.2.0, openvino 2026.4.1, transformers 5.16.1
• Specialization: mathematical reasoning (base model, no instruction tuning)

Usage

optimum-intel

from optimum.intel import OVModelForCausalLM
from transformers import AutoTokenizer

model = OVModelForCausalLM.from_pretrained("Nekodeus/qwen25-math7b-base-openvino")
tok = AutoTokenizer.from_pretrained("Nekodeus/qwen25-math7b-base-openvino")
inputs = tok("The derivative of x^2 is", return_tensors="pt")
print(tok.decode(model.generate(**inputs, max_new_tokens=64)[0]))

OpenVINO GenAI

ov::genai::LLMPipeline pipe("path/to/Nekodeus/qwen25-math7b-base-openvino", "CPU");

Reproduce

optimum-cli export openvino \
  --model Qwen/Qwen2.5-Math-7B \
  --task text-generation \
  --weight-format int8 \
  qwen25-math7b-base-openvino