Skip to content

Commit 146a73b

Browse files
vladimir-dudnikvladimir-dudnik
authored
correct output format for classification models (#2910)
1 parent e14d01e commit 146a73b

File tree

25 files changed

+40
-40
lines changed

25 files changed

+40
-40
lines changed

‎models/public/dla-34/README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -53,14 +53,14 @@ Channel order is `BGR`
5353
Object classifier according to ImageNet classes, name - `prob`, shape - `1, 1000`, output data format is `B, C`, where:
5454

5555
- `B` - batch size
56-
- `C` - predicted probabilities for each class in [0, 1] range
56+
- `C` - predicted probabilities for each class in logits format
5757

5858
### Converted model
5959

6060
Object classifier according to ImageNet classes, name - `prob`, shape - `1, 1000`, output data format is `B, C`, where:
6161

6262
- `B` - batch size
63-
- `C` - predicted probabilities for each class in [0, 1] range
63+
- `C` - predicted probabilities for each class in logits format
6464

6565
## Download a Model and Convert it into Inference Engine Format
6666

‎models/public/googlenet-v3-pytorch/README.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -59,7 +59,7 @@ Expected color order - `BGR`.
5959
Object classifier according to ImageNet classes, name - `prob`, shape - `1, 1000` in `B, C` format, where:
6060

6161
- `B` - batch size
62-
- `C` - vector of probabilities for each class in [0, 1] range
62+
- `C` - vector of probabilities for each class in logits format
6363

6464
## Download a Model and Convert it into Inference Engine Format
6565

‎models/public/googlenet-v3/README.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -50,7 +50,7 @@ Expected color order: `BGR`.
5050
Object classifier according to ImageNet classes, name: `InceptionV3/Predictions/Softmax`, shape: `1, 1001` in `B, C` format, where:
5151

5252
- `B` - batch size
53-
- `C` - vector of probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format.
53+
- `C` - vector of probabilities for all dataset classes in [0, 1] range (0 class is background).
5454

5555
## Download a Model and Convert it into Inference Engine Format
5656

‎models/public/hbonet-0.25/README.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -50,7 +50,7 @@ Expected color order: `BGR`.
5050
Object classifier according to ImageNet classes, shape: `1, 1000` in `B, C` format, where:
5151

5252
- `B` - batch size
53-
- `C` - vector of probabilities for all dataset classes.
53+
- `C` - vector of probabilities for all dataset classes in logits format.
5454

5555
## Download a Model and Convert it into Inference Engine Format
5656

‎models/public/hbonet-0.5/README.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -50,7 +50,7 @@ Expected color order: `BGR`.
5050
Object classifier according to ImageNet classes, shape: `1, 1000` in `B, C` format, where:
5151

5252
- `B` - batch size
53-
- `C` - vector of probabilities for all dataset classes.
53+
- `C` - vector of probabilities for all dataset classes in logits format.
5454

5555
## Download a Model and Convert it into Inference Engine Format
5656

‎models/public/hbonet-1.0/README.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -50,7 +50,7 @@ Expected color order: `BGR`.
5050
Object classifier according to ImageNet classes, shape: `1, 1000` in `B, C` format, where:
5151

5252
- `B` - batch size
53-
- `C` - vector of probabilities for all dataset classes.
53+
- `C` - vector of probabilities for all dataset classes in logits format.
5454

5555
## Download a Model and Convert it into Inference Engine Format
5656

‎models/public/mobilenet-v1-0.25-128/README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -49,11 +49,11 @@ Expected color order: `BGR`.
4949

5050
### Original Model
5151

52-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Reshape_1`.
52+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Reshape_1`.
5353

5454
### Converted Model
5555

56-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
56+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
5757

5858
- `B` - batch size
5959
- `C` - vector of probabilities.

‎models/public/mobilenet-v1-0.50-160/README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -49,11 +49,11 @@ Expected color order: `BGR`.
4949

5050
### Original Model
5151

52-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Reshape_1`.
52+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Reshape_1`.
5353

5454
### Converted Model
5555

56-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
56+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
5757

5858
- `B` - batch size
5959
- `C` - vector of probabilities.

‎models/public/mobilenet-v1-0.50-224/README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -49,11 +49,11 @@ Expected color order: `BGR`.
4949

5050
### Original Model
5151

52-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Reshape_1`.
52+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Reshape_1`.
5353

5454
### Converted Model
5555

56-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
56+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
5757

5858
- `B` - batch size
5959
- `C` - vector of probabilities.

‎models/public/mobilenet-v1-1.0-224-tf/README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -49,11 +49,11 @@ Expected color order: `BGR`.
4949

5050
### Original Model
5151

52-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Reshape_1`.
52+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Reshape_1`.
5353

5454
### Converted Model
5555

56-
Probabilities for all dataset classes (0 class is background). Probabilities are represented in logits format. Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
56+
Probabilities for all dataset classes in [0, 1] range (0 class is background). Name: `MobilenetV1/Predictions/Softmax`, shape: `1, 1001`, format: `B, C`, where:
5757

5858
- `B` - batch size
5959
- `C` - vector of probabilities.

0 commit comments

Comments
 (0)