QuantizeLinear - 10 vs 13

Files changed (1) hide show
  1. QuantizeLinear10 → QuantizeLinear13 +18 -7
QuantizeLinear10 → QuantizeLinear13 RENAMED
@@ -1 +1 @@
1
- The linear per-tensor/layer quantization operator. It consumes a high precision tensor, a scale, a zero point to compute the low precision / quantized tensor.
1
+ The linear quantization operator. It consumes a high precision tensor, a scale, and a zero point to compute the low precision / quantized tensor.
2
+ The scale factor and zero point must have same shape, and can be either a scalar for per-tensor / per layer quantization, or a 1-D tensor for per-axis quantization.
3
+ The quantization formula is y = saturate ((x / y_scale) + y_zero_point).
2
- The quantization formula is y = saturate ((x / y_scale) + y_zero_point). For saturation, it saturates to [0, 255] if it's uint8, or [-128, 127] if it's int8.
4
+ For saturation, it saturates to [0, 255] if it's uint8, or [-128, 127] if it's int8.
3
5
  For (x / y_scale), it's rounding to nearest ties to even. Refer to https://en.wikipedia.org/wiki/Rounding for details. 'y_zero_point' and 'y' must have same type.
6
+
7
+ **Attributes**
8
+
9
+ * **axis**:
10
+ (Optional) The axis of the quantization dimension of the input
11
+ tensor. Ignored for per-tensor quantization. Negative value means
12
+ counting dimensions from the back. Accepted range is [-r, r-1] where
13
+ r = rank(input).
4
14
  **Inputs**
5
15
  Between 2 and 3 inputs.
6
16
  * **x** (heterogeneous) - **T1**:
7
17
  N-D full precision Input tensor to be quantized.
8
18
  * **y_scale** (heterogeneous) - **tensor(float)**:
9
- Scale for doing quantization to get 'y'. It's a scalar, which means
19
+ Scale for doing quantization to get 'y'. It can be a scalar, which
10
- a per-tensor/layer quantization.
20
+ means per-tensor/layer quantization, or a 1-D Tensor for per-axis
21
+ quantization.
11
22
  * **y_zero_point** (optional, heterogeneous) - **T2**:
12
- Zero point for doing quantization to get 'y'. It's a scalar, which
23
+ Zero point for doing quantization to get 'y'. Shape must match
13
- means a per-tensor/layer quantization. Default value is uint8 typed
24
+ y_scale. Default is uint8 with zero point of 0 if it's not
14
- 0 if it's not specified.
25
+ specified.
15
26
  **Outputs**
16
27
  * **y** (heterogeneous) - **T2**:
17
28
  N-D quantized output tensor. It has same shape as input 'x'.
18
29
  **Type Constraints**
19
30
  * **T1** in (
20
31
  tensor(float),
21
32
  tensor(int32)
22
33
  ):
23
34
  Constrain 'x' to float or int32 tensor.
24
35
  * **T2** in (
25
36
  tensor(int8),
26
37
  tensor(uint8)
27
38
  ):
28
39
  Constrain 'y_zero_point' and 'y' to 8-bit integer tensor.