search

gorgonia / tensor

package tensor provides efficient and generic n-dimensional arrays in Go that are useful for machine learning and deep learning purposes
Apache License 2.0
359 stars 49 forks source link

slice with step fix #74

Closed cpllbstr closed 4 years ago

cpllbstr commented 4 years ago

Added check for residual

chewxy commented 4 years ago

@cpllbstr

Can you also add these tests to dense_matop_test.go (line 563 onwards)

    // 3tensor with leading and trailing 1s

    {"3T1[0]", Range(Float64, 0, 9), Shape{1, 9, 1}, []Slice{ss(0)}, Shape{9, 1}, []int{1, 1}, Range(Float64, 0, 9)},
    {"3T1[nil, 0:2]", Range(Float64, 0, 9), Shape{1, 9, 1}, []Slice{nil, makeRS(0, 2)}, Shape{1, 2, 1}, []int{9, 1, 1}, Range(Float64, 0, 2)},
    {"3T1[nil, 0:5:3]", Range(Float64, 0, 9), Shape{1, 9, 1}, []Slice{nil, makeRS(0, 5, 3)}, Shape{1, 2, 1}, []int{9, 3, 1}, Range(Float64, 0, 5)},
    {"3T1[nil, 1:5:3]", Range(Float64, 0, 9), Shape{1, 9, 1}, []Slice{nil, makeRS(1, 5, 3)}, Shape{1, 2, 1}, []int{9, 3, 1}, Range(Float64, 1, 5)},
    {"3T1[nil, 1:9:3]", Range(Float64, 0, 9), Shape{1, 9, 1}, []Slice{nil, makeRS(1, 9, 3)}, Shape{1, 3, 1}, []int{9, 3, 1}, Range(Float64, 1, 9)},

    // 3tensor
    {"3T[0]", Range(Float64, 0, 36), Shape{2, 9, 2}, []Slice{ss(0)}, Shape{9, 2}, []int{2, 1}, Range(Float64, 0, 18)},
    {"3T[1]", Range(Float64, 0, 36), Shape{2, 9, 2}, []Slice{ss(1)}, Shape{9, 2}, []int{2, 1}, Range(Float64, 18, 36)},
    {"3T[1, 2]", Range(Float64, 0, 36), Shape{2, 9, 2}, []Slice{ss(1), ss(2)}, Shape{2}, []int{1}, Range(Float64, 22, 24)},
    {"3T[1, 2:4]", Range(Float64, 0, 36), Shape{2, 9, 2}, []Slice{ss(1), makeRS(2, 4)}, Shape{2, 2}, []int{2, 1}, Range(Float64, 22, 26)},
    {"3T[1, 2:8:2]", Range(Float64, 0, 36), Shape{2, 9, 2}, []Slice{ss(1), makeRS(2, 8, 2)}, Shape{3, 2}, []int{4, 1}, Range(Float64, 22, 34)},
    {"3T[1, 2:8:3]", Range(Float64, 0, 36), Shape{2, 9, 2}, []Slice{ss(1), makeRS(2, 8, 3)}, Shape{2, 2}, []int{6, 1}, Range(Float64, 22, 34)},
    {"3T[1, 2:9:2]", Range(Float64, 0, 126), Shape{2, 9, 7}, []Slice{ss(1), makeRS(2, 9, 2)}, Shape{4, 7}, []int{14, 1}, Range(Float64, 77, 126)},
    {"3T[1, 2:9:2, 1]", Range(Float64, 0, 126), Shape{2, 9, 7}, []Slice{ss(1), makeRS(2, 9, 2), ss(1)}, Shape{4}, []int{14}, Range(Float64, 78, 121)}, // should this be a colvec?
    {"3T[1, 2:9:2, 1:4:2]", Range(Float64, 0, 126), Shape{2, 9, 7}, []Slice{ss(1), makeRS(2, 9, 2), makeRS(1, 4, 2)}, Shape{4, 2}, []int{14, 2}, Range(Float64, 78, 123)},
}
cpllbstr commented 4 years ago

@chewxy PR updated

chewxy commented 4 years ago

sweet. Will merge once CI passes

coveralls commented 4 years ago

Coverage Status

Coverage increased (+0.005%) to 73.179% when pulling 55e116ce9bd237c2edb72d657a173e9b321b1005 on cpllbstr:master into f83015a5d83bbf6fb9437fb8ed1ce7b060e157ef on gorgonia:master.