---
library_name: optimum
tags:
- onnx
- text-generation
- minicpm5
base_model:
- GnLOLot/MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking
---

# MiniCPM5-1B Claude Opus Fable5 V2 Thinking — ONNX

This repository contains an ONNX Runtime export generated with Optimum from the matching Transformers checkpoint:

`GnLOLot/MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking`

The requested source checkpoint `mlx-community/MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking-OptiQ-4bit` is an MLX-native mixed-precision artifact. Optimum cannot load that MLX quantization format directly, so this export uses the matching full-precision Transformers checkpoint and is stored as FP32 ONNX with external data.

## Load with Optimum

```python
from optimum.onnxruntime import ORTModelForCausalLM
from transformers import AutoTokenizer

model_id = "Sharjeelbaig/MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking-ONNX"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = ORTModelForCausalLM.from_pretrained(model_id)
```

The model was exported with the `text-generation-with-past` task and validated by ONNX Runtime on CPU.