Print a tidytensor.

Prints a summary of a tidytensor as a nested hierarchy of tensors of lower rank.

# S3 method for tidytensor
print(
  x,
  show_dimnames = FALSE,
  max_per_level = 1,
  base_rank = NULL,
  max_rows = 6,
  max_cols = 6,
  max_depth = 3,
  signif_digits = 3,
  indent = 0,
  ...
)

Arguments

x: a tidytensor to summarize.
show_dimnames: show the dimension names, if present, or dimension indices if not in base-level prints.
max_per_level: only show this many sub-tensors per level.
base_rank: either NULL, 1, 2, or 3 - specifies whether the inner/bottom-most tensors should be represented as rank 1, 2, or 3 in a grid (NULL for autodetect based on tensor shape, see details).
max_rows: limit the base-level prints to include this many rows (also applies to 1d prints).
max_cols: limit the base-level prints to include this many columns.
max_depth: in 3d representation, limit the base-level prints to include this many entries of the last rank.
signif_digits: number of significant digits to print for numeric tensors.
indent: indent the printout by this much (used internally).
...: additional arguments to be passed to or from methods (ignored).

Details

The base_rank argument specifies whether the lowest ranks of the tensor (displayed as a grid) should be shown as rank 2 tensors, rank 3 tensors, or rank 1 tensors; the default of NULL will select 3 if the last rank is of size 3 or 1 (assuming an image and a "channels-last" convention), 2 if the 3rd-to-last rank is length 3 or 1 (assuming an image and a "channels-first" convention) or if there are only two ranks or if the last two ranks are equal (assuming an image channel of some kind), and otherwise will default to 1.

max_per_level indicates how many replicates

Examples

t <- as.tidytensor(array(1:(2 * 3 * 4 * 5), dim = c(2, 3, 4, 5)))
ranknames(t) <- c("samples", "batches", "rows", "cols")
print(t, base_rank = 2)
#> # Rank 4 tensor, shape: (2, 3, 4, 5), ranknames: samples, batches, rows, cols
#> |  # Rank 3 tensor, shape: (3, 4, 5)
#> |  |  # Rank 2 tensor, shape: (4, 5)
#> |  |       1  25  49  73   97 
#> |  |       7  31  55  79  103 
#> |  |      13  37  61  85  109 
#> |  |      19  43  67  91  115 
#> |  |  # ...
#> |  # ...

t <- as.tidytensor(array(1:(2 * 3 * 40 * 50 * 3), dim = c(2, 3, 40, 50, 3)))
ranknames(t) <- c("sample", "batch", "row", "pixel", "channel")
print(t, max_rows = 6, max_cols = 6, max_depth = 3, show_dimnames = TRUE, base_rank = 3)
#> # Rank 5 tensor, shape: (2, 3, 40, 50, 3), ranknames: sample, batch, row, pixel, channel
#> |  # Rank 4 tensor, shape: (3, 40, 50, 3)
#> |  |  # Rank 3 tensor, shape: (40, 50, 3)
#> |  |                          [,1,]                [,2,]                [,3,]                [,4,]                [,5,]                 [,6,]  ... 
#> |  |      [1,,]   [1, 12000, 24000]  [241, 12200, 24200]  [481, 12500, 24500]  [721, 12700, 24700]  [961, 13000, 25000]  [1200, 13200, 25200]  ... 
#> |  |      [2,,]   [7, 12000, 24000]  [247, 12200, 24200]  [487, 12500, 24500]  [727, 12700, 24700]  [967, 13000, 25000]  [1210, 13200, 25200]  ... 
#> |  |      [3,,]  [13, 12000, 24000]  [253, 12300, 24300]  [493, 12500, 24500]  [733, 12700, 24700]  [973, 13000, 25000]  [1210, 13200, 25200]  ... 
#> |  |      [4,,]  [19, 12000, 24000]  [259, 12300, 24300]  [499, 12500, 24500]  [739, 12700, 24700]  [979, 13000, 25000]  [1220, 13200, 25200]  ... 
#> |  |      [5,,]  [25, 12000, 24000]  [265, 12300, 24300]  [505, 12500, 24500]  [745, 12700, 24700]  [985, 13000, 25000]  [1220, 13200, 25200]  ... 
#> |  |      [6,,]  [31, 12000, 24000]  [271, 12300, 24300]  [511, 12500, 24500]  [751, 12800, 24800]  [991, 13000, 25000]  [1230, 13200, 25200]  ... 
#> |  |        ...                 ...                  ...                  ...                  ...                  ...                   ...  ... 
#> |  |  # ...
#> |  # ...

Arguments

Details

See also

Examples