aichallenger / ai_challenger_2018 Goto Github PK

I ran the train_CNN.py and set CUDA_VISIBLE DEVICES, but it still did not run on GPU. In my understanding, keras with tensorflow backend can use GPU automatically.

Are there any specific examples submitted ？

a_embedding, _ = self.a_encoder(a_embeddings.view(-1, a_embeddings.size(2), a_embeddings.size(3)))
这个a_embeddings.size(3)是什么含义

Hello there, i would like to download the video understanding dataset, but found it not available on the official website(https://challenger.ai/dataset/mlsvd2018). Does anyone have any idea how or where i can get this dataset? Thank you for your kind help in advance.

在测试baseline的过程中，发现对应的问价并不存在，就试着找到一些如下：
能否给出更“清洁的”数据，并对训练验证用图片文件有比较清晰的命名或者编号？
../AgriculturalDisease_validationset/images/d879aaf6-d557-4607-b7a0-ecbb3761f152___RS_LB 4723.JPG doesn't exists!
../AgriculturalDisease_validationset/images/d93424d7-a8db-4cff-b534-2f47930950fa___RS_HL 1745.JPG doesn't exists!
../AgriculturalDisease_validationset/images/e777c06477039578fa0cfbf750f6a4a8.jpg doesn't exists!
../AgriculturalDisease_validationset/images/ff20152d-baf7-45bd-a1a1-a5f52f389275___UF.GRC_YLCV_Lab 01917.JPG doesn't exists!
../AgriculturalDisease_validationset/images/IMG_20180623_182710.jpg doesn't exists!
../AgriculturalDisease_validationset/images/ff158c0ad13534e1302a71bb5b54978a.jpg doesn't exists!
../AgriculturalDisease_validationset/images/fff1d02879899b8f3749a5573270208b.jpg doesn't exists!
../AgriculturalDisease_validationset/images/debf892cc124358c66df432bea6b7fbf.jpg doesn't exists!
../AgriculturalDisease_validationset/images/daef3f32df6f0503ab9d69737d82d26a.jpg doesn't exists!
../AgriculturalDisease_validationset/images/f9e4b1bc-36b0-40df-bb48-4e6ed48ec8b5___FREC_Scab 3514.JPG doesn't exists!
../AgriculturalDisease_validationset/images/d83de01988eb2f409f3785ee5f8d199b.jpg doesn't exists!
../AgriculturalDisease_validationset/images/e250d8157695486a23e5304d568ee160.jpg doesn't exists!
../AgriculturalDisease_validationset/images/f538eae64c334b8b6468f94e025a9722.jpg doesn't exists!
../AgriculturalDisease_validationset/images/fd8d4072-e0b3-4936-95c2-e2d6dc6d22b0___FREC_Pwd.M 0250.JPG doesn't exists!
../AgriculturalDisease_validationset/images/df5d63fb-d270-459b-9798-ad680374ad1a___UF.Citrus_HLB_Lab 1668.JPG doesn't exists!
../AgriculturalDisease_validationset/images/f42f177ddf65171277fcf4773da0a909.jpg doesn't exists!
../AgriculturalDisease_validationset/images/f8fdd1396856e03c9a9a979cfd0097d6.jpg doesn't exists!
../AgriculturalDisease_validationset/images/f3a192cdcf04db4a7783262bc24b7b81.jpg doesn't exists!
../AgriculturalDisease_validationset/images/IMG_20180623_190102.jpg doesn't exists!
../AgriculturalDisease_validationset/images/f021c403-3d7b-4369-a928-531ca45c3b08___RS_HL 9682.JPG doesn't exists!

Please add Nvidia-driver version for Short-video classification README.

Now, I cann't get an accuracy for my predicts because there is no chance to upload our results like several months ago.

So, is there any other ways to evalute our results?

I'm interested in Plant Disease Classification.

Thanks.

各位大佬， 有遇到官网提交 drivable area 和 alternative area 零分的情况吗。用了测试脚本mIOU还可以，但交上去就不对，显示零分。
测试脚本可以提供一张 train 图在ref的 案例吗，一直显示零分都不敢交了。。。

数据集有么？

File "train_CNN.py", line 132, in main
callbacks=[checkpointer])
File "/home/ubuntu/anaconda2/envs/py36/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper
return func(*args, **kwargs)
File "/home/ubuntu/anaconda2/envs/py36/lib/python3.6/site-packages/keras/engine/training.py", line 2224, in fit_generator
class_weight=class_weight)
File "/home/ubuntu/anaconda2/envs/py36/lib/python3.6/site-packages/keras/engine/training.py", line 1877, in train_on_batch
class_weight=class_weight)
File "/home/ubuntu/anaconda2/envs/py36/lib/python3.6/site-packages/keras/engine/training.py", line 1480, in _standardize_user_data
exception_prefix='target')
File "/home/ubuntu/anaconda2/envs/py36/lib/python3.6/site-packages/keras/engine/training.py", line 123, in _standardize_input_data
str(data_shape))
ValueError: Error when checking target: expected reshape_2 to have shape (40,) but got array with shape (17,)

What's mean about this RuntimeWarning? Is this reason to cause the mIOU is 0? Who can tell me what the right seg result pictures is like?

+1

Originally posted by @jz1998 in #32 (comment)

I have been trying to follow the tutorial in order to replicate the baseline model. I have converted the data to tfRecord format by using the old label format since the code doesn't work for the new label format. I have also downloaded the "faster_rcnn_resnet50_coco_2018_01_28" model from model zoo. Now I know I have to run the following line for training:

python train.py --logtostderr --train_dir path/to/train_dir --model_config_path path/to/model_config.pbtxt --train_config_path path/to/train_config.pbtxt --input_config_path path/to/train_input_config.pbtxt

Problem is that model_config.pbtxt, train_config.pbtxt and train_input_config.pbtxt files are not present in either this repository or the model file downloaded. How am I supposed to get these files?

2018-10-24 07:58:03,091 [INFO] (MainThread) start load data
2018-10-24 07:58:04,746 [INFO] (MainThread) start seg train data
Building prefix dict from the default dictionary ...
2018-10-24 07:58:04,747 [DEBUG] (MainThread) Building prefix dict from the default dictionary ...
Loading model from cache c:\users\user\appdata\local\temp\jieba.cache
2018-10-24 07:58:04,750 [DEBUG] (MainThread) Loading model from cache c:\users\user\appdata\local\temp\jieba.cache
Loading model cost 0.344 seconds.
2018-10-24 07:58:05,092 [DEBUG] (MainThread) Loading model cost 0.344 seconds.
Prefix dict has been built succesfully.
2018-10-24 07:58:05,092 [DEBUG] (MainThread) Prefix dict has been built succesfully.
2018-10-24 08:02:45,384 [INFO] (MainThread) complete seg train data
2018-10-24 08:02:45,384 [INFO] (MainThread) start train feature extraction
Traceback (most recent call last):
File "D:/work/SentimentanalysisCore-master/Baselines/main_train.py", line 52, in
vectorizer_tfidf.fit(content_train)
File "C:\Python27\lib\site-packages\sklearn\feature_extraction\text.py", line 1562, in fit
X = super(TfidfVectorizer, self).fit_transform(raw_documents)
File "C:\Python27\lib\site-packages\sklearn\feature_extraction\text.py", line 1012, in fit_transform
self.fixed_vocabulary_)
File "C:\Python27\lib\site-packages\sklearn\feature_extraction\text.py", line 934, in _count_vocab
values.extend(feature_counter.values())
MemoryError

Hi,

I can't download the dataset from the official website. Can you fix the problem?

"query_id": 289334, "alternatives": "不能|无法确定"
"query_id": 289730, "alternatives": "无法确定"

其中query_id为 289334和289730的问题候选答案不是三个，这样在inference的时候会报错，所以我自己给添加了，如下：
289334, "alternatives": "能|不能|无法确定"
289730, "alternatives": "打折|不打折|无法确定"
可以吗？

Traceback (most recent call last):
File "train.py", line 186, in
tf.app.run()
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/platform/app.py", line 125, in run
_sys.exit(main(argv))
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/deprecation.py", line 272, in new_func
return func(*args, **kwargs)
File "train.py", line 182, in main
graph_hook_fn=graph_rewriter_fn)
File "/usr/local/lib/python3.5/dist-packages/object_detection-0.1-py3.5.egg/object_detection/legacy/trainer.py", line 396, in train
include_global_step=False))
File "/usr/local/lib/python3.5/dist-packages/object_detection-0.1-py3.5.egg/object_detection/utils/variables_helper.py", line 126, in get_variables_available_in_checkpoi
nt ckpt_reader = tf.train.NewCheckpointReader(checkpoint_path)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 306, in NewCheckpointReader
return CheckpointReader(compat.as_bytes(filepattern), status)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/errors_impl.py", line 519, in exit
c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.DataLossError: not an sstable (bad magic number)

谢谢大家

最近想做研究，但是官网数据集无法下载，链接打不开

Since the WAD2018 challenge has ended the submission server on Berkeley's portal is closed. While testing images are available, there annotated versions are missing. Can the test labels be provided here or is there any other server which can be used for the purpose of evaluating a trained model on the BDD100k test data?

细粒度用户评论情感分析数据集2018-上传结果计算出错

opinion_questions_machine_reading_comprehension2018_baseline/preprocess.py", line 105

提示缺少data/word-count.obj和data/word2id.obj两个文件

如何处理这个，你们有试过这个baseline吗
iou是多少

Can anyone share links to download this dataset ?

Thanks.

49 def score(a, b):
50 return (b - a) / b

should be
def score(a, b):
return (b - a) / a
is it right ?

line 24 : --problems=$PROBLEM
Maybe --problem=$PROBLEM \

Please add baseline algorithm implementation for Short-video classification. Thanks.

I can't find the labels or evaluation script of the competition about sentiment_analysis .
where to get test labels for testing accuracy?