google/automl issues and pull requests

#1213 - I want to change to a regression model, how do I change it

Issue - State: open - Opened by Yang12334 5 days ago

#1212 - Efficientdet lite1 mAP

Issue - State: open - Opened by suke138 about 1 month ago

#1105 - Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = -1.000

Issue - State: open - Opened by ayennam almost 3 years ago - 2 comments

#695 - How to train the object detection model for negative samples

Issue - State: closed - Opened by RuchaTambe about 4 years ago - 6 comments

#688 - Memory leak in train_and_eval mode

Issue - State: closed - Opened by NikZak about 4 years ago - 6 comments

#100 - can you show the graphs?

Issue - State: closed - Opened by bitwangdan over 4 years ago - 1 comment

#99 - Convert coco data to tfrecord. Is this normal?Finished writing, skipped 26715 annotations.

Issue - State: closed - Opened by goldwater668 over 4 years ago - 4 comments

#98 - Could it produce json format box result?

Issue - State: closed - Opened by b03505036 over 4 years ago - 4 comments

#97 - det_post_process is very slow

Issue - State: closed - Opened by monghimng over 4 years ago - 17 comments
Labels: P1

#96 - The test speed of D6 on coco val2017 is very slow

Issue - State: closed - Opened by PKUbahuangliuhe over 4 years ago - 3 comments

#95 - Soft nms

Pull Request - State: closed - Opened by fsx950223 over 4 years ago - 5 comments
Labels: cla: yes

#94 - How to train on VOC or custom dataset?

Issue - State: closed - Opened by cyx6666 over 4 years ago - 14 comments

#93 - Use your own data to train the following problems?

Issue - State: closed - Opened by goldwater668 over 4 years ago - 21 comments

#92 - Questions about pretrained efficientdet

Issue - State: closed - Opened by handsomefengfeng over 4 years ago - 1 comment

#91 - ask for more information about loss when training on COCO 2017

Issue - State: closed - Opened by Alex220284 over 4 years ago - 7 comments

#90 - tensorflow.python.framework.errors_impl.InvalidArgumentError: 2 root error(s) found.

Issue - State: closed - Opened by YangZhangMizzou over 4 years ago - 11 comments

#89 - Today the model detects less object

Issue - State: closed - Opened by wwdok over 4 years ago - 17 comments

#88 - Small changes in README.md:)

Pull Request - State: closed - Opened by StepfenShawn over 4 years ago - 2 comments
Labels: cla: yes

#87 - How can i run model_inspect.py for video?

Issue - State: closed - Opened by levanpon98 over 4 years ago - 4 comments

#86 - Can EfficientDet train customized dataset?

Issue - State: closed - Opened by YangZhangMizzou over 4 years ago - 3 comments

#85 - OOM for GPU training

Issue - State: closed - Opened by b03505036 over 4 years ago - 21 comments

#84 - How can I see the log or tensor_board for the loss value?

Issue - State: closed - Opened by WonTaeYeon over 4 years ago - 3 comments

#83 - why the default efficientdet-d1 num_classes for coco is 90

Issue - State: closed - Opened by b03505036 over 4 years ago - 2 comments

#82 - inference.ServingDriver can 't work

Issue - State: closed - Opened by mmzmagic over 4 years ago - 3 comments

#81 - Predict Problem

Issue - State: closed - Opened by HK017 over 4 years ago - 5 comments

#80 - Problems in transforming coco data set？

Issue - State: closed - Opened by goldwater668 over 4 years ago - 3 comments

#79 - How to use AA during training ?

Issue - State: closed - Opened by b03505036 over 4 years ago - 1 comment

#78 - Update copyright year

Pull Request - State: closed - Opened by StepfenShawn over 4 years ago
Labels: cla: yes

#77 - Conversion to TRT with INT8 quantization.

Issue - State: closed - Opened by MartinBrazdil over 4 years ago - 4 comments

#76 - Is number of classes including background?

Issue - State: closed - Opened by b03505036 over 4 years ago - 2 comments

#75 - Box regression loss

Issue - State: closed - Opened by dkoguciuk over 4 years ago - 3 comments

#74 - I use the efficientdet-d0 model for reasoning, but the reasoning picture does not have the effect of rectangular box?

Issue - State: closed - Opened by goldwater668 over 4 years ago - 3 comments

#73 - module 'inference' has no attribute 'InferenceDriver'

Issue - State: closed - Opened by wwdok over 4 years ago - 5 comments

#72 - some questions about GPU Devices

Issue - State: closed - Opened by Guzaiwang over 4 years ago - 1 comment

#71 - How do I identify only one class of a picture, such as “person”?

Issue - State: closed - Opened by quuhua911 over 4 years ago - 2 comments

#70 - Doubt on inputs_offsets in pyramid config

Issue - State: closed - Opened by bluesky314 over 4 years ago - 1 comment

#69 - Image normalization

Issue - State: closed - Opened by ancorasir over 4 years ago - 1 comment

#68 - Custom dataset 0 mAP

Issue - State: closed - Opened by CraigWang1 over 4 years ago - 49 comments

#67 - best_eval metrics

Issue - State: closed - Opened by fsx950223 over 4 years ago

#66 - How can I save it as an ONNX model?

Issue - State: closed - Opened by WonTaeYeon over 4 years ago - 13 comments

#65 - Redundant bias

Issue - State: closed - Opened by rwightman over 4 years ago - 8 comments

#64 - ValueError: The passed save_path is not a valid checkpoint: efficientdet-d0

Issue - State: closed - Opened by nicolas-gervais over 4 years ago - 5 comments

#63 - The Code has been deleted?

Issue - State: closed - Opened by tabsun over 4 years ago - 1 comment

#62 - Does this code support CUDA version 10.1.243 + tensorflow-gpu2.1.0?

Issue - State: closed - Opened by goldwater668 over 4 years ago - 3 comments

#61 - Train error

Issue - State: closed - Opened by ckqsars over 4 years ago - 3 comments

#60 - Can this code eval with GPU？

Issue - State: closed - Opened by goldwater668 over 4 years ago - 7 comments

#59 - NotFoundError: Restoring from checkpoint failed.

Issue - State: closed - Opened by wangtianlong1994 over 4 years ago - 5 comments

#58 - tfrecord can‘t creat

Issue - State: closed - Opened by xiaopojizu123 over 4 years ago

#57 - tfrecord can‘t creat

Issue - State: closed - Opened by xiaopojizu123 over 4 years ago - 1 comment

#56 - File "main.py", line 318, in main current_step = int(os.path.basename(ckpt).split('-')[1]) IndexError: list index out of range

Issue - State: closed - Opened by goldwater668 over 4 years ago - 2 comments

#55 - support Saved model

Pull Request - State: closed - Opened by fsx950223 over 4 years ago - 4 comments
Labels: cla: yes

#54 - Does model support channels_first data_format?

Issue - State: closed - Opened by fsx950223 over 4 years ago

#53 - Display loss?

Issue - State: closed - Opened by CraigWang1 over 4 years ago - 5 comments

#52 - Remove dependency on deleted file

Pull Request - State: closed - Opened by CraigWang1 over 4 years ago - 6 comments
Labels: cla: yes

#51 - BFlops is not the same as the results reported in the paper.

Issue - State: closed - Opened by tingtingliangvs over 4 years ago - 2 comments

#50 - How to export a saved model ?

Issue - State: closed - Opened by Byronnar over 4 years ago - 2 comments

#49 - Train from public checkpoints

Pull Request - State: closed - Opened by Ely-S over 4 years ago - 1 comment
Labels: cla: yes

#48 - How long did it take to train efficientdet-d1?

Issue - State: closed - Opened by elv-xuwen over 4 years ago - 1 comment

#47 - fix typo in verbosity level

Pull Request - State: closed - Opened by Ely-S over 4 years ago
Labels: cla: yes

#46 - pretrained-model load ERROR!

Issue - State: closed - Opened by HK017 over 4 years ago - 5 comments

#45 - bbox loss always zero

Issue - State: closed - Opened by Root-9527 over 4 years ago - 3 comments

#44 - Is there a semantic segmentation network or instance segmentation betwork based on Efficient-Det?

Issue - State: closed - Opened by bilel-bj over 4 years ago - 6 comments

#43 - How to do prediction in video

Issue - State: closed - Opened by PiyalGeorge over 4 years ago - 5 comments

#42 - Test speed code available. you can modify it for predicting your own image.

Issue - State: closed - Opened by YanqingWu over 4 years ago - 10 comments

#41 - ValueError: Variable efficientnet-b6/stem/conv2d/kernel already exists, disallowed

Issue - State: closed - Opened by deepakcrk over 4 years ago - 2 comments

#40 - Finetune form pretrained checkpoint

Issue - State: closed - Opened by veryfocus over 4 years ago - 7 comments

#39 - Train error

Issue - State: closed - Opened by WonTaeYeon over 4 years ago - 5 comments

#38 - fix training command in README

Pull Request - State: closed - Opened by Ely-S over 4 years ago - 4 comments
Labels: cla: yes

#37 - error in change num_classes

Issue - State: closed - Opened by zhangchaosunshine over 4 years ago - 3 comments

#36 - How to modify EfficientDet for data different from 2D RGB images?

Issue - State: closed - Opened by bilel-bj over 4 years ago - 2 comments

#35 - NaN loss during trainning.

Issue - State: closed - Opened by YanqingWu over 4 years ago - 3 comments

#34 - How to train the code on my own dataset?

Issue - State: closed - Opened by bilel-bj over 4 years ago - 1 comment

#33 - CudNN failed to initialize when calling model_inspect.py

Issue - State: closed - Opened by Cyril9227 over 4 years ago - 6 comments

#32 - inferenced by pb

Issue - State: closed - Opened by Jilliansea over 4 years ago - 8 comments

#31 - Can you provide a tutorial on how to build a model ?

Issue - State: closed - Opened by Cyril9227 over 4 years ago - 1 comment

#30 - Conv Bn ReLU pattern

Issue - State: closed - Opened by romulus0914 over 4 years ago

#29 - GPU train

Issue - State: closed - Opened by silence0628 over 4 years ago - 28 comments

#28 - How can I use multi GPU?

Issue - State: closed - Opened by Sylvia-Vee over 4 years ago - 6 comments

#27 - Detections much worse than YOLOv3?

Issue - State: closed - Opened by AakashKumarNain over 4 years ago - 17 comments

#26 - why setting the box_loss_weight =50?

Issue - State: closed - Opened by HardLaugh over 4 years ago - 7 comments

#25 - Training

Issue - State: closed - Opened by Zumbalamambo over 4 years ago - 4 comments

#24 - W tensorflow/core/grappler/utils/graph_view.cc:830] No registered 'swish_f32' OpKernel for GPU devices compatible with node {{node model_1/swish_38/swish_f32}}

Issue - State: closed - Opened by bitwangdan over 4 years ago - 5 comments

#23 - Add saved model support

Issue - State: closed - Opened by fsx950223 over 4 years ago

#22 - How to convert train TFREcord for my own dataset

Issue - State: closed - Opened by HK017 over 4 years ago - 7 comments

#21 - What are the parameters for training only some classes?

Issue - State: closed - Opened by WonTaeYeon over 4 years ago - 2 comments

#20 - Not found: Key efficientnet-b0/blocks_0/conv2d/kernel not found in checkpoint

Issue - State: closed - Opened by CraigWang1 over 4 years ago - 12 comments

#19 - AssertionError: Bad argument number for Name: 3, expecting 4

Issue - State: closed - Opened by deepakcrk over 4 years ago - 4 comments

#18 - The test seems not using GPU

Issue - State: closed - Opened by PKUbahuangliuhe over 4 years ago - 1 comment

#17 - Using Efficentdet with TF serving

Issue - State: closed - Opened by akashdeepjassal over 4 years ago - 4 comments

#16 - Requested more than 0 entries, but params is empty

Issue - State: closed - Opened by deepakcrk over 4 years ago - 5 comments

#15 - 'NoneType' object has no attribute 'value'

Issue - State: closed - Opened by KaqiKaqi over 4 years ago - 6 comments

#14 - No module named 'tensorflow.compat.v2'

Issue - State: closed - Opened by KaqiKaqi over 4 years ago - 1 comment

#13 - EfficentDet Segmentation Model

Issue - State: closed - Opened by bluesky314 over 4 years ago - 1 comment

#12 - fix model_inspect command

Pull Request - State: closed - Opened by fsx950223 over 4 years ago

#11 - EfficientDet for EdgeTPU

Issue - State: closed - Opened by alexanderfrey over 4 years ago - 17 comments

#10 - Which version of tensorflow_addons to use?

Issue - State: closed - Opened by KaqiKaqi over 4 years ago - 1 comment

#9 - The output features of BiFPN Layer

Issue - State: closed - Opened by kevin20080525 over 4 years ago - 1 comment

#8 - [help] what is the 'training_file_pattern'?

Issue - State: closed - Opened by WonTaeYeon over 4 years ago - 3 comments

#7 - Why all new model writed by tf1.x???

Issue - State: closed - Opened by Lannister-Xiaolin over 4 years ago - 6 comments

#6 - How can I predict an image?

Issue - State: closed - Opened by levanpon98 over 4 years ago - 10 comments

GitHub / google/automl issues and pull requests