Nekodeus's picture
add README.md
bc53834 verified
|
Raw History Blame Contribute Delete
1.59 kB
---
license: gemma
library_name: openvino
tags:
- openvino
- int8
- text-generation
- codegemma
- code
- llm
---
# CodeGemma-7B-it OpenVINO (INT8)
OpenVINO IR conversion of [google/codegemma-7b-it](https://huggingface.co/google/codegemma-7b-it) with weight-only INT8 quantization, produced on CPU-only Kaggle with `optimum-cli` (fully offline, token-free).
- Base model: `google/codegemma-7b-it` (GemmaForCausalLM, 8.5B params, Gemma license)
- 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`
- Instruction-tuned code model (Gemma-1B coder head on Gemma-7B base), 8k context, Python/program synthesis
## Usage
### optimum-intel
```python
from optimum.intel import OVModelForCausalLM
from transformers import AutoTokenizer
model = OVModelForCausalLM.from_pretrained("Nekodeus/codegemma-7b-it-openvino")
tok = AutoTokenizer.from_pretrained("Nekodeus/codegemma-7b-it-openvino")
inputs = tok("<start_of_turn>user\nWrite a Python function to reverse a string<end_of_turn>\n<start_of_turn>model\n", return_tensors="pt")
print(tok.decode(model.generate(**inputs, max_new_tokens=256)[0]))
```
### OpenVINO GenAI
```cpp
ov::genai::LLMPipeline pipe("path/to/Nekodeus/codegemma-7b-it-openvino", "CPU");
```
## Reproduce
```bash
optimum-cli export openvino \
--model google/codegemma-7b-it \
--task text-generation \
--weight-format int8 \
codegemma-7b-it-openvino
```
## License
Gemma Terms of Use (gated on HF; this conversion is un-gated).