Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 1 | Copyright 2017 The Android Open Source Project |
| 2 | |
| 3 | Licensed under the Apache License, Version 2.0 (the "License"); |
| 4 | you may not use this file except in compliance with the License. |
| 5 | You may obtain a copy of the License at |
| 6 | |
| 7 | http://www.apache.org/licenses/LICENSE-2.0 |
| 8 | |
| 9 | Unless required by applicable law or agreed to in writing, software |
| 10 | distributed under the License is distributed on an "AS IS" BASIS, |
| 11 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 12 | See the License for the specific language governing permissions and |
| 13 | limitations under the License. |
| 14 | ------------------------------------------------------------------ |
| 15 | |
| 16 | This directory contains models data for the Android Neural Networks API benchmarks. |
| 17 | |
| 18 | Included models: |
| 19 | |
| 20 | ------------------------------------------------------------------ |
| 21 | - mobilenet_float.tflite |
| 22 | MobileNet tensorflow lite model based on: |
| 23 | "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications" |
| 24 | https://arxiv.org/abs/1704.04861 |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 25 | Apache License, Version 2.0 |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 26 | |
Mika Raento | 38cb4be | 2018-08-01 13:26:44 +0100 | [diff] [blame] | 27 | Downloaded from |
| 28 | http://download.tensorflow.org/models/mobilenet_v1_2018_02_22/mobilenet_v1_1.0_224.tgz |
| 29 | on Aug 1 2018 and converted using ToT toco. |
| 30 | Golden output generated with ToT tensorflow (Linux, CPU). |
| 31 | |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 32 | ------------------------------------------------------------------ |
| 33 | - mobilenet_quantized.tflite |
| 34 | 8bit quantized MobileNet tensorflow lite model based on: |
| 35 | "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications" |
| 36 | https://arxiv.org/abs/1704.04861 |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 37 | Apache License, Version 2.0 |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 38 | |
Mika Raento | 38cb4be | 2018-08-01 13:26:44 +0100 | [diff] [blame] | 39 | Downloaded from |
| 40 | http://download.tensorflow.org/models/mobilenet_v1_2018_07_12/mobilenet_v1_1.0_224_quant.tgz |
| 41 | on Aug 1 2018. |
| 42 | Golden output generated with ToT tflite (Linux, CPU). |
| 43 | |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 44 | ------------------------------------------------------------------ |
| 45 | - hdrnet_float.tflite |
| 46 | Partial tensorflow lite model based on |
| 47 | "Deep Bilateral Learningfor Real-Time Image Enhancement" |
| 48 | https://groups.csail.mit.edu/graphics/hdrnet/ |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 49 | Apache License, Version 2.0 |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 50 | |
| 51 | It's partial because Pack and Transpose operations were not supported at the time |
| 52 | of model creation. |
| 53 | |
| 54 | ------------------------------------------------------------------ |
| 55 | - hdrnet_quantized.tflite |
| 56 | 8bit quantized partial tensorflow lite model based on |
| 57 | "Deep Bilateral Learning for Real-Time Image Enhancement" |
| 58 | https://groups.csail.mit.edu/graphics/hdrnet/ |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 59 | Apache License, Version 2.0 |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 60 | |
Przemyslaw Szczepaniak | 918fba2 | 2018-05-25 15:15:48 +0100 | [diff] [blame] | 61 | ------------------------------------------------------------------ |
| 62 | - resnet_float.tflite |
| 63 | ResNet tensorflow lite model based on: |
| 64 | "Deep Residual Learning for Image Recognition" |
| 65 | https://arxiv.org/abs/1512.03385 |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 66 | Apache License, Version 2.0 |
Przemyslaw Szczepaniak | 918fba2 | 2018-05-25 15:15:48 +0100 | [diff] [blame] | 67 | |
| 68 | ------------------------------------------------------------------ |
| 69 | - resnet_quantized.tflite |
| 70 | 8bit quantized ResNet tensorflow lite model based on: |
| 71 | "Deep Residual Learning for Image Recognition" |
| 72 | https://arxiv.org/abs/1512.03385 |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 73 | Apache License, Version 2.0 |
| 74 | |
Viet Dang | a1d44e8 | 2018-07-04 13:35:37 +0100 | [diff] [blame] | 75 | ------------------------------------------------------------------ |
| 76 | - tts_float.tflite |
| 77 | TTS tensorflow lite model based on: |
| 78 | "Fast, Compact, and High Quality LSTM-RNN Based Statistical Parametric Speech Synthesizers for |
| 79 | Mobile Devices" |
| 80 | https://ai.google/research/pubs/pub45379 |
| 81 | Apache License, Version 2.0 |
| 82 | |
| 83 | Note that the tensorflow lite model is the acoustic model in the paper. It is used because it is |
| 84 | much heavier than the duration model. |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 85 | |
| 86 | ------------------------------------------------------------------ |
| 87 | Input files: |
| 88 | ------------------------------------------------------------------ |
| 89 | - cup.input |
| 90 | Photo of cup by pszczepaniak@google.com, Apache License, Version 2.0 |
Viet Dang | 7069703 | 2018-09-07 16:25:18 +0100 | [diff] [blame] | 91 | - tts_float/arctic_*.input |
| 92 | Linguistic features and durations generated from text sentences from the CMU Arctic set |
| 93 | (http://www.festvox.org/cmu_arctic/cmuarctic.data), Apache License, Version 2.0 |
Przemyslaw Szczepaniak | 397b27f | 2018-06-04 16:00:25 +0100 | [diff] [blame] | 94 | ------------------------------------------------------------------ |
| 95 | |
Przemyslaw Szczepaniak | 918fba2 | 2018-05-25 15:15:48 +0100 | [diff] [blame] | 96 | |
Przemyslaw Szczepaniak | 6a16f22 | 2018-05-18 13:40:18 +0100 | [diff] [blame] | 97 | TODO(pszczepaniak): Update hdrnet to full model with pack and transpose |
| 98 | TODO(pszczepaniak): Provide at least 5 inputs outputs for each model |
| 99 | |