| func.func @attention1x3x4() { |
| %init = tensor.empty() : tensor<1x3x4xf32> |
| %query = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2]]]> : tensor<1x3x4xf32> |
| |
| %key = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2]]]> : tensor<1x3x4xf32> |
| %value = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2]]]> : tensor<1x3x4xf32> |
| %scale = arith.constant 0.5 : f32 |
| %1 = iree_linalg_ext.attention {indexing_maps = [affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d4)>, |
| affine_map<(d0, d1, d2, d3, d4) -> ()>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d4)>]} |
| ins(%query, %key, %value, %scale : tensor<1x3x4xf32>, |
| tensor<1x3x4xf32>, tensor<1x3x4xf32>, f32) outs(%init : tensor<1x3x4xf32>) { |
| ^bb0(%arg0: f32): |
| iree_linalg_ext.yield %arg0 : f32 |
| } -> tensor<1x3x4xf32> |
| check.expect_almost_eq_const( |
| %1, |
| dense<[[[0.5530, 0.6530, 0.7530, 0.8530], |
| [0.6328, 0.7328, 0.8328, 0.9328], |
| [0.7011, 0.8011, 0.9011, 1.0011]]]> : tensor<1x3x4xf32> |
| ) : tensor<1x3x4xf32> |
| return |
| } |
| |
| func.func @causal_attention1x3x4() { |
| %init = tensor.empty() : tensor<1x3x4xf32> |
| %query = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2]]]> : tensor<1x3x4xf32> |
| |
| %key = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2]]]> : tensor<1x3x4xf32> |
| %value = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2]]]> : tensor<1x3x4xf32> |
| %mask = util.unfoldable_constant dense<[[[true, false, false], |
| [true, true, false], |
| [true, true, true]]]> : tensor<1x3x3xi1> |
| %scale = arith.constant 0.5 : f32 |
| %1 = iree_linalg_ext.attention {indexing_maps = [affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d4)>, |
| affine_map<(d0, d1, d2, d3, d4) -> ()>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d3)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d4)>]} |
| ins(%query, %key, %value, %scale, %mask : tensor<1x3x4xf32>, |
| tensor<1x3x4xf32>, tensor<1x3x4xf32>, f32, tensor<1x3x3xi1>) outs(%init : tensor<1x3x4xf32>) { |
| ^bb0(%arg0: f32): |
| iree_linalg_ext.yield %arg0 : f32 |
| } -> tensor<1x3x4xf32> |
| check.expect_almost_eq_const( |
| %1, |
| dense<[[[0.1000, 0.2000, 0.3000, 0.4000], |
| [0.3509, 0.4509, 0.5509, 0.6509], |
| [0.7011, 0.8011, 0.9011, 1.0011]]]> : tensor<1x3x4xf32> |
| ) : tensor<1x3x4xf32> |
| return |
| } |
| |
| func.func @attention1x4x4_i1_mask_all_ones() { |
| %init = tensor.empty() : tensor<1x4x4xf32> |
| %query = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2], |
| [1.3, 1.4, 1.5, 1.6]]]> : tensor<1x4x4xf32> |
| |
| %key = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2], |
| [1.3, 1.4, 1.5, 1.6]]]> : tensor<1x4x4xf32> |
| %value = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2], |
| [1.3, 1.4, 1.5, 1.6]]]> : tensor<1x4x4xf32> |
| |
| %mask = util.unfoldable_constant dense<[[[true, true, true, true], |
| [true, true, true, true], |
| [true, true, true, true], |
| [true, true, true, true]]]> : tensor<1x4x4xi1> |
| |
| %scale = arith.constant 0.5 : f32 |
| %1 = iree_linalg_ext.attention {indexing_maps = [affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d4)>, |
| affine_map<(d0, d1, d2, d3, d4) -> ()>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d3)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d4)>]} |
| ins(%query, %key, %value, %scale, %mask : tensor<1x4x4xf32>, |
| tensor<1x4x4xf32>, tensor<1x4x4xf32>, f32, tensor<1x4x4xi1>) outs(%init : tensor<1x4x4xf32>) { |
| ^bb0(%arg0: f32): |
| iree_linalg_ext.yield %arg0 : f32 |
| } -> tensor<1x4x4xf32> |
| check.expect_almost_eq_const( |
| %1, |
| dense<[[[0.798884, 0.898884, 0.998884, 1.09888], |
| [0.941939, 1.04194, 1.14194, 1.24194], |
| [1.05371, 1.15371, 1.25371, 1.35371], |
| [1.13295, 1.23295, 1.33295, 1.43295]]]> : tensor<1x4x4xf32> |
| ) : tensor<1x4x4xf32> |
| return |
| } |
| |
| func.func @softcap_attention1x3x4() { |
| %init = tensor.empty() : tensor<1x3x4xf32> |
| %query = util.unfoldable_constant dense<[[[-0.6523, 0.7802, -0.6770, -0.9474], |
| [ 0.6822, 0.8594, 0.1035, -0.5115], |
| [-0.7860, -0.5562, -0.2750, -0.0973]]]> : tensor<1x3x4xf32> |
| |
| %key = util.unfoldable_constant dense<[[[-0.3457, 0.5223, -0.0553, 0.5851], |
| [ 0.4057, -0.8352, 0.1931, 0.1678], |
| [ 0.1438, 0.7713, -0.6535, 0.9221]]]> : tensor<1x3x4xf32> |
| %value = util.unfoldable_constant dense<[[[-0.2679, 0.6395, 0.4953, 0.3751], |
| [ 0.8195, -0.5137, 0.2178, 0.9182], |
| [ 0.8644, 0.7615, 0.2934, 0.9721]]]> : tensor<1x3x4xf32> |
| %scale = arith.constant 1.0 : f32 |
| %1 = iree_linalg_ext.attention {indexing_maps = [affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d4)>, |
| affine_map<(d0, d1, d2, d3, d4) -> ()>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d4)>]} |
| ins(%query, %key, %value, %scale : tensor<1x3x4xf32>, |
| tensor<1x3x4xf32>, tensor<1x3x4xf32>, f32) outs(%init : tensor<1x3x4xf32>) { |
| ^bb0(%arg0: f32): |
| %tanh = math.tanh %arg0 : f32 |
| %cst = arith.constant 2.0 : f32 |
| %mul = arith.mulf %tanh, %cst : f32 |
| iree_linalg_ext.yield %mul : f32 |
| } -> tensor<1x3x4xf32> |
| check.expect_almost_eq_const( |
| %1, |
| dense<[[[0.331149, 0.576043, 0.380060, 0.6879], |
| [0.519539, 0.532548, 0.342270, 0.78571], |
| [0.442426, 0.125056, 0.329748, 0.735554]]]> : tensor<1x3x4xf32> |
| ) : tensor<1x3x4xf32> |
| return |
| } |
| |
| func.func @attention1x4x4() { |
| %init = tensor.empty() : tensor<1x4x4xf32> |
| %query = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2], |
| [1.3, 1.4, 1.5, 1.6]]]> : tensor<1x4x4xf32> |
| %key = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2], |
| [1.3, 1.4, 1.5, 1.6]]]> : tensor<1x4x4xf32> |
| %value = util.unfoldable_constant dense<[[[0.1, 0.2, 0.3, 0.4], |
| [0.5, 0.6, 0.7, 0.8], |
| [0.9, 1.0, 1.1, 1.2], |
| [1.3, 1.4, 1.5, 1.6]]]> : tensor<1x4x4xf32> |
| %scale = arith.constant 0.5 : f32 |
| %1 = iree_linalg_ext.attention {indexing_maps = [affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d4)>, |
| affine_map<(d0, d1, d2, d3, d4) -> ()>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d4)>]} |
| ins(%query, %key, %value, %scale : tensor<1x4x4xf32>, |
| tensor<1x4x4xf32>, tensor<1x4x4xf32>, f32) outs(%init : tensor<1x4x4xf32>) { |
| ^bb0(%arg0: f32): |
| iree_linalg_ext.yield %arg0 : f32 |
| } -> tensor<1x4x4xf32> |
| check.expect_almost_eq_const( |
| %1, |
| dense<[[[0.7989, 0.8989, 0.9989, 1.0989], |
| [0.9419, 1.0419, 1.1419, 1.2419], |
| [1.0537, 1.1537, 1.2537, 1.3537], |
| [1.1329, 1.2329, 1.3329, 1.4329]]]> : tensor<1x4x4xf32> |
| ) : tensor<1x4x4xf32> |
| return |
| } |
| |
| func.func @attention3x3x4() { |
| %init = tensor.empty() : tensor<3x3x4xf32> |
| %query = util.unfoldable_constant dense<[[[-1.5256, -0.7502, -0.6540, -1.6095], |
| [-0.6092, -0.9798, -1.6091, -0.7121], |
| [ 0.4676, -0.6970, -1.1608, 0.6995]], |
| [[ 0.8657, 0.2444, -0.6629, 0.8073], |
| [-0.1759, -2.2456, -1.4465, 0.0612], |
| [-0.7735, 0.1991, 0.0457, 0.1530]], |
| [[-0.1110, 0.2927, -0.1578, -0.0288], |
| [ 1.1422, 0.2486, -1.7754, -0.0255], |
| [ 1.6103, -0.7040, -0.1853, -0.9962]]]> : tensor<3x3x4xf32> |
| %key = util.unfoldable_constant dense<[[[-0.6092, -0.9798, -1.6091, -0.7121], |
| [-0.7773, -0.2515, -0.2223, 1.6871], |
| [ 0.4676, -0.6970, -1.1608, 0.6995]], |
| [[ 0.8657, 0.2444, -0.6629, 0.8073], |
| [-0.7981, -0.1316, 1.8793, -0.0721], |
| [-0.7735, 0.1991, 0.0457, 0.1530]], |
| [[-0.1110, 0.2927, -0.1578, -0.0288], |
| [ 1.1422, 0.2486, -1.7754, -0.0255], |
| [ 1.6103, -0.7040, -0.1853, -0.9962]]]> : tensor<3x3x4xf32> |
| %value = util.unfoldable_constant dense<[[[-1.5256, -0.7502, -0.6540, -1.6095], |
| [-0.6092, -0.9798, -1.6091, -0.7121], |
| [ 0.4676, -0.6970, -1.1608, 0.6995]], |
| [[ 0.8657, 0.2444, -0.6629, 0.8073], |
| [-0.1759, -2.2456, -1.4465, 0.0612], |
| [-0.7773, -0.2515, -0.2223, 1.6871]], |
| [[-0.1110, 0.2927, -0.1578, -0.0288], |
| [-0.5962, -1.0055, 0.4285, 1.4761], |
| [ 1.6103, -0.7040, -0.1853, -0.9962]]]> : tensor<3x3x4xf32> |
| %scale = arith.constant 0.5 : f32 |
| %1 = iree_linalg_ext.attention {indexing_maps = [affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d2)>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d3, d4)>, |
| affine_map<(d0, d1, d2, d3, d4) -> ()>, |
| affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d4)>]} |
| ins(%query, %key, %value, %scale : tensor<3x3x4xf32>, |
| tensor<3x3x4xf32>, tensor<3x3x4xf32>, f32) outs(%init : tensor<3x3x4xf32>) { |
| ^bb0(%arg0: f32): |
| iree_linalg_ext.yield %arg0 : f32 |
| } -> tensor<3x3x4xf32> |
| check.expect_almost_eq_const( |
| %1, |
| dense<[[[-1.2804, -0.7607, -0.7648, -1.3364], |
| [-1.0711, -0.7572, -0.8238, -1.0953], |
| [-0.4030, -0.7807, -1.1112, -0.3482]], |
| [[ 0.4245, -0.1012, -0.6484, 0.9162], |
| [ 0.1324, -0.2762, -0.6125, 1.0206], |
| [-0.1866, -0.9266, -0.7977, 0.8593]], |
| [[ 0.1917, -0.4658, 0.0510, 0.2561], |
| [-0.1054, -0.8358, 0.2544, 0.8461], |
| [ 0.9522, -0.7023, -0.0358, -0.3303]]]> : tensor<3x3x4xf32> |
| ) : tensor<3x3x4xf32> |
| return |
| } |
