quantize-rs 0.8.0

Neural network quantization toolkit for ONNX models
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

# Example quantize-rs configuration file (YAML format)

# Global settings (applied to all models unless overridden)
bits: 8
per_channel: true

# Individual model configurations
models:
  - input: models/resnet18.onnx
    output: quantized/resnet18_int8.onnx
  
  - input: models/mobilenet.onnx
    output: quantized/mobilenet_int8.onnx
    per_channel: false  # Override global setting
  
  - input: models/efficientnet.onnx
    output: quantized/efficientnet_int8.onnx
    skip_existing: true

# Batch processing (optional)
batch:
  input_dir: "models/*.onnx"
  output_dir: quantized_batch/
  skip_existing: true
  continue_on_error: true