Implementation:Microsoft Onnxruntime FloatTensorType Init
Metadata
| Field | Value |
|---|---|
| Implementation Name | FloatTensorType_Init |
| Repository | Microsoft_Onnxruntime |
| Source Repository | https://github.com/microsoft/onnxruntime |
| Type | Wrapper Doc |
| Wrapper Tool | skl2onnx |
| Language | Python |
| Domain | ML_Inference, Model_Conversion |
| Last Updated | 2026-02-10 |
| Workflow | Train_Convert_Predict |
| Pair | 2 of 5 |
Overview
Wrapper documentation for the skl2onnx.common.data_types.FloatTensorType class, which declares typed tensor specifications for ONNX model input schema definition.
API Signature
skl2onnx.common.data_types.FloatTensorType([None, num_features])
Import
from skl2onnx.common.data_types import FloatTensorType
Code Reference
| Reference | Location |
|---|---|
| Usage example | docs/python/examples/plot_train_convert_predict.py:L53-55 |
I/O Contract
Inputs
| Parameter | Type | Required | Description |
|---|---|---|---|
shape |
list |
Yes | Tensor shape specification. Use None for dynamic dimensions (e.g., batch size). Use integers for fixed dimensions (e.g., feature count).
|
Outputs
| Output | Type | Description |
|---|---|---|
| return value | FloatTensorType |
A typed tensor descriptor used in the initial_type list for ONNX conversion.
|
| initial_type list | list[tuple[str, FloatTensorType]] |
A list of (name, type) tuples passed to convert_sklearn().
|
Shape Specification
| Shape | Meaning |
|---|---|
[None, 4] |
Dynamic batch size, 4 features (most common for Iris dataset). |
[1, 4] |
Fixed batch size of 1, 4 features. |
[None, N] |
Dynamic batch size, N features (general pattern). |
Usage Example
from skl2onnx.common.data_types import FloatTensorType
# Define input schema for a model with 4 features and dynamic batch size
initial_type = [("float_input", FloatTensorType([None, 4]))]
From the source at docs/python/examples/plot_train_convert_predict.py:L53-55:
from skl2onnx.common.data_types import FloatTensorType
initial_type = [("float_input", FloatTensorType([None, 4]))]
onx = convert_sklearn(clr, initial_types=initial_type)
Fixed Batch Size Variant
From docs/python/examples/plot_train_convert_predict.py:L183 (RandomForest example):
initial_type = [("float_input", FloatTensorType([1, 4]))]
Related Pages
Page Connections
Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment