Tensor
Tensors are N-dimensional data matrix with homogeneous type. Supported types are:
- Unsigned integers:
uint8,uint16,uint32,uint64 - Signed integers:
uint8,uint16,uint32,uint64 - Floating point numbers:
float16,float32,float64
Note: 1-D tensors are visualized as bar charts.
Components and APIs
Primary component: tensor
Python APIs: log_tensor,
Rust API: Tensor
Simple Example
"""Create and log a tensor.""" import rerun as rr from numpy.random import default_rng rng = default_rng(12345) # Create a 4-dimensional tensor tensor = rng.uniform(0.0, 1.0, (8, 6, 3, 5)) rr.init("rerun_example_tensors", spawn=True) # Log the tensor, assigning names to each dimension rr.log_tensor("tensor", tensor, names=("width", "height", "channel", "batch"))
1-D Tensor Example
"""Create and log a one dimensional tensor.""" import rerun as rr from numpy.random import default_rng rng = default_rng(12345) # Create a 1-dimensional tensor tensor = rng.laplace(0.0, 1.0, 100) rr.init("rerun_example_tensors", spawn=True) # Log the tensor, assigning names to each dimension rr.log_tensor("tensor", tensor)
