Nekodeus/starcoder2-7b-openvino

StarCoder2-7B OpenVINO (INT8)

OpenVINO IR conversion of bigcode/starcoder2-7b with weight-only INT8 quantization, produced on CPU-only Kaggle with optimum-cli (fully offline, token-free).

• Base model: bigcode/starcoder2-7b (Starcoder2ForCausalLM, 7B params, BigCode OpenRAIL-M v1)
• 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: code completion (fill-in-the-middle, 8k context)

Usage

optimum-intel

from optimum.intel import OVModelForCausalLM
from transformers import AutoTokenizer

model = OVModelForCausalLM.from_pretrained("Nekodeus/starcoder2-7b-openvino")
tok = AutoTokenizer.from_pretrained("Nekodeus/starcoder2-7b-openvino", trust_remote_code=True)
inputs = tok("def fibonacci(n):\n    if n <= 1:\n        return", return_tensors="pt")
print(tok.decode(model.generate(**inputs, max_new_tokens=64)[0]))

OpenVINO GenAI

ov::genai::LLMPipeline pipe("path/to/Nekodeus/starcoder2-7b-openvino", "CPU");

Reproduce

optimum-cli export openvino \
  --model bigcode/starcoder2-7b \
  --task text-generation \
  --weight-format int8 \
  starcoder2-7b-openvino