Specify operation CTCLoss-4 (#1189)
* Specify operation CTCLoss-4 Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com> * Correct documentation for CTCLoss after #1 review Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com> * Correct documentation for CTCLoss after #2 review Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com> * Correct documentation for CTCLoss after #3 review * Correct documentation for CTCLoss after #4 review Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com> * Correct layout for logits and add more description for unique attribute Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com> * Correct types for length and indices tensors Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com> * Correct formulas and punctuation Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
This commit is contained in:
Roman Kazantsev
GitHub
125
docs/ops/sequence/CTCLoss_4.md
Normal file
125
docs/ops/sequence/CTCLoss_4.md
Normal file
@@ -0,0 +1,125 @@
|
||||
## CTCLoss <a name="CTCLoss"></a>
|
||||
|
||||
**Versioned name**: *CTCLoss-4*
|
||||
|
||||
**Category**: Sequence processing
|
||||
|
||||
**Short description**: *CTCLoss* computes the CTC (Connectionist Temporal Classification) Loss.
|
||||
|
||||
**Detailed description**:
|
||||
|
||||
*CTCLoss* operation is presented in [Connectionist Temporal Classification - Labeling Unsegmented Sequence Data with Recurrent Neural Networks: Graves et al., 2016](http://www.cs.toronto.edu/~graves/icml_2006.pdf)
|
||||
|
||||
*CTCLoss* estimates likelyhood that a target `labels[i,:]` can occur (or is real) for given input sequence of logits `logits[i,:,:]`.
|
||||
Briefly, *CTCLoss* operation finds all sequences aligned with a target `labels[i,:]`, computes log-probabilities of the aligned sequences using `logits[i,:,:]`
|
||||
and computes a negative sum of these log-probabilies.
|
||||
|
||||
Input sequences of logits `logits` can have different lengths. The length of each sequence `logits[i,:,:]` equals `logit_length[i]`.
|
||||
A length of target sequence `labels[i,:]` equals `label_length[i]`. The length of the target sequence must not be greater than the length of corresponding input sequence `logits[i,:,:]`.
|
||||
Otherwise, the operation behaviour is undefined.
|
||||
|
||||
*CTCLoss* calculation scheme:
|
||||
|
||||
1. Compute probability of `j`-th character at time step `t` for `i`-th input sequence from `logits` using softmax formula:
|
||||
\f[
|
||||
p_{i,t,j} = \frac{\exp(logits[i,t,j])}{\sum^{K}_{k=0}{\exp(logits[i,t,k])}}
|
||||
\f]
|
||||
|
||||
2. For a given `i`-th target from `labels[i,:]` find all aligned paths.
|
||||
A path `S = (c1,c2,...,cT)` is aligned with a target `G=(g1,g2,...,gT)` if both chains are equal after decoding.
|
||||
The decoding extracts substring of length `label_length[i]` from a target `G`, merges repeated characters in `G` in case *preprocess_collapse_repeated* equal to True and
|
||||
finds unique elements in the order of character occurence in case *unique* equal to True.
|
||||
The decoding merges repeated characters in `S` in case *ctc_merge_repeated* equal to True and removes blank characters represented by `blank_index`.
|
||||
By default, `blank_index` is equal to `C-1`, where `C` is a number of classes including the blank.
|
||||
For example, in case default *ctc_merge_repeated*, *preprocess_collapse_repeated*, *unique* and `blank_index` a target sequence `G=(0,3,2,2,2,2,2,4,3)` of a length `label_length[i]=4` is processed
|
||||
to `(0,3,2,2)` and a path `S=(0,0,4,3,2,2,4,2,4)` of a length `logit_length[i]=9` is also processed to `(0,3,2,2)`, where `C=5`.
|
||||
There exist other paths that are also aligned with `G`, for instance, `0,4,3,3,2,4,2,2,2`. Paths checked for alignment with a target `label[:,i]` must be of length `logit_length[i] = L_i`.
|
||||
Compute probabilities of these aligned paths (alignments) as follows:
|
||||
\f[
|
||||
p(S) = \prod_{t=1}^{L_i} p_{i,t,ct}
|
||||
\f]
|
||||
|
||||
3. Finally, compute negative sum of log-probabilities of all found alignments:
|
||||
\f[
|
||||
CTCLoss = \minus \sum_{S} \ln p(S)
|
||||
\f]
|
||||
|
||||
**Note**: This calculation scheme does not provide steps for optimal implementation and primarily serves for better explanation.
|
||||
|
||||
**Attributes**
|
||||
|
||||
* *preprocess_collapse_repeated*
|
||||
|
||||
* **Description**: *preprocess_collapse_repeated* is a flag for a preprocessing step before loss calculation, wherein repeated labels in `labels[i,:]` passed to the loss are merged into single labels.
|
||||
* **Range of values**: True or False
|
||||
* **Type**: `boolean`
|
||||
* **Default value**: False
|
||||
* **Required**: *no*
|
||||
|
||||
* *ctc_merge_repeated*
|
||||
|
||||
* **Description**: *ctc_merge_repeated* is a flag for merging repeated characters in a potential alignment during the CTC loss calculation.
|
||||
* **Range of values**: True or False
|
||||
* **Type**: `boolean`
|
||||
* **Default value**: True
|
||||
* **Required**: *no*
|
||||
|
||||
* *unique*
|
||||
|
||||
* **Description**: *unique* is a flag to find unique elements for a target `labels[i,:]` before matching with potential alignments. Unique elements in the processed `labels[i,:]` are sorted in the order of their occurence in original `labels[i,:]`. For example, the processed sequence for `labels[i,:]=(0,1,1,0,1,3,3,2,2,3)` of length `label_length[i]=10` will be `(0,1,3,2)` in case *unique* equal to True.
|
||||
* **Range of values**: True or False
|
||||
* **Type**: `boolean`
|
||||
* **Default value**: False
|
||||
* **Required**: *no*
|
||||
|
||||
**Inputs**
|
||||
|
||||
* **1**: `logits` - Input tensor with a batch of sequences of logits. Type of elements is *T_F*. Shape of the tensor is `[N, T, C]`, where `N` is the batch size, `T` is the maximum sequence length and `C` is the number of classes including the blank. Required.
|
||||
|
||||
* **2**: `logit_length` - 1D input tensor of type *T1* and of a shape `[N]`. The tensor must consist of non-negative values not greater than `T`. Lengths of input sequences of logits `logits[i,:,:]`. Required.
|
||||
|
||||
* **3**: `labels` - 2D tensor with shape `[N, T]` of type *T2*. A length of a target sequence `labels[i,:]` is equal to `label_length[i]` and must contain of integers from a range `[0; C-1]` except `blank_index`. Required.
|
||||
|
||||
* **4**: `label_length` - 1D tensor of type *T1* and of a shape `[N]`. The tensor must consist of non-negative values not greater than `T` and `label_length[i] <= logit_length[i]` for all possible `i`. Required.
|
||||
|
||||
* **5**: `blank_index` - Scalar of type *T2*. Set the class index to use for the blank label. Default value is `C-1`. Optional.
|
||||
|
||||
**Output**
|
||||
|
||||
* **1**: Output tensor with shape `[N]`, negative sum of log-probabilities of alignments. Type of elements is *T_F*.
|
||||
|
||||
**Types**
|
||||
|
||||
* *T_F*: any supported floating point type.
|
||||
|
||||
* *T1*, *T2*: `int32` or `int64`.
|
||||
|
||||
**Example**
|
||||
|
||||
```xml
|
||||
<layer ... type="CTCLoss" ...>
|
||||
<input>
|
||||
<port id="0">
|
||||
<dim>8</dim>
|
||||
<dim>20</dim>
|
||||
<dim>128</dim>
|
||||
</port>
|
||||
<port id="1">
|
||||
<dim>8</dim>
|
||||
</port>
|
||||
<port id="2">
|
||||
<dim>8</dim>
|
||||
<dim>20</dim>
|
||||
</port>
|
||||
<port id="3">
|
||||
<dim>8</dim>
|
||||
</port>
|
||||
<port id="4"> <!-- blank_index value is: 120 -->
|
||||
</input>
|
||||
<output>
|
||||
<port id="0">
|
||||
<dim>8</dim>
|
||||
</port>
|
||||
</output>
|
||||
</layer>
|
||||
```
|
||||
Reference in New Issue
Block a user