i'm trying to follow this tutorial with my own local data files:
CNTK tutorial
i have the following function to save my data array into a txt file feedable to CNTK:
# Save the data files into a format compatible with CNTK text reader
def savetxt(filename, ndarray):
dir = os.path.dirname(filename)
if not os.path.exists(dir):
os.makedirs(dir)
if not os.path.isfile(filename):
print("Saving", filename )
with open(filename, 'w') as f:
labels = list(map(' '.join, np.eye(11, dtype=np.uint).astype(str)))
for row in ndarray:
row_str = row.astype(str)
label_str = labels[row[-1]]
feature_str = ' '.join(row_str[:-1])
f.write('|labels {} |features {}\n'.format(label_str, feature_str))
else:
print("File already exists", filename)
i have 2 ndarrays of the following shape that i want to feed the model:
train.shape
(1976L, 15104L)
test.shape
(1976L, 15104L)
Then i try to implement the fucntion like this:
# Save the train and test files (prefer our default path for the data)
data_dir = os.path.join("C:/Users", 'myself', "OneDrive", "IA Project", 'data', 'train')
if not os.path.exists(data_dir):
data_dir = os.path.join("data", "IA Project")
print ('Writing train text file...')
savetxt(os.path.join(data_dir, "Train-128x118_cntk_text.txt"), train)
print ('Writing test text file...')
savetxt(os.path.join(data_dir, "Test-128x118_cntk_text.txt"), test)
print('Done')
and then i get the following error:
Writing train text file...
Saving C:/Users\A702628\OneDrive - Atos\Microsoft Capstone IA\Capstone data\train\Train-128x118_cntk_text.txt
---------------------------------------------------------------------------
IndexError Traceback (most recent call last)
<ipython-input-24-b53d3c69b8d2> in <module>()
6
7 print ('Writing train text file...')
----> 8 savetxt(os.path.join(data_dir, "Train-128x118_cntk_text.txt"), train)
9
10 print ('Writing test text file...')
<ipython-input-23-610c077db694> in savetxt(filename, ndarray)
12 for row in ndarray:
13 row_str = row.astype(str)
---> 14 label_str = labels[row[-1]]
15 feature_str = ' '.join(row_str[:-1])
16 f.write('|labels {} |features {}\n'.format(label_str, feature_str))
IndexError: list index out of range
Can somebody please tell me what's going wrong with this part of the code? And how could i fix it? Thank you very much in advance.
Since you're using your own input data -- are they labelled in the range 0 to 9? The labels array only has 10 entries in it, so that could cause an out-of-range problem.
Related
I have a problem loading data from text file in Octave.
My text file looks like this:
# Created by Octave 5.2.0, Wed May 05 16:07:02 2021 GMT <unknown#DESKTOP-HEVT6O6>
# name: x
# type: matrix
# rows: 1
# columns: 3600
4.8899999999999997 4.9000000000000004 4.9000000000000004 4.9100000000000001 4.9299999999999997 4.9249999999999998 ...
I need to load those float numbers in one matrix and plot them in time domain.
My code so far:
fs = 360;
Ts = 1/fs;
d = fileread('ecg.txt');
data = regexp(d(1,136:62328),' ','split');
data = str2double(data);
ed = length(data);
t = linspace(0,Ts,ed - 1);
figure(1)
plot(t,data(1,2:ed))
So My question is if there is another way to do it or if there is a better way to do it.
Your file is in Octave’s text data format. This is the default file format when saving variables to file with save. That is, that text file was saved in Octave using save ecg.txt x. The Octave command load ecg.txt will load the file, and re-create the x variable just like it was when it was saved.
Thus, to plot your data, just do
load ecg.txt
plot(x)
I'm working on a network using triplet mining for training. In order to make it work properly, I need my batches to contain several images of the same class. The problem I'm currently facing is that I have 751 classes, for a total of 12,937 pictures, and a batch size of 48 pictures. When shuffling the dataset using the command below, the odds to get pictures from the same class are really low, making the triplet mining inefficient.
dataset = dataset.shuffle(12937)
What I would need instead is a way of generating batches that contain a specific number of pictures for every class represented in this batch. As an example, let's say here that I want 12 classes per batch, there would be 4 pictures for each of them.
Another problem I'm facing is how would I shuffle this dataset at the end of every epoch so that I can have different batches that still follow the condition fixed above, that is 12 classes, 4 pictures for each one of them?
Is there any proper way to do it? I can't really find one. Please let me know if I'm unclear, and if you need further details.
================ EDIT ================
I've been trying a few things, and came up with something that would do what I want. The function would be the following:
counter = 0.
# Assuming a format such as (data, label)
def predicate(data, label):
global counter
allowed_labels = tf.constant([counter])
isallowed = tf.equal(allowed_labels, tf.cast(label, tf.float32))
reduced = tf.reduce_sum(tf.cast(isallowed, tf.float32))
counter += 1
return tf.greater(reduced, tf.constant(0.))
##tf.function
def custom_shuffle(train_dataset, batch_size, samples_per_class = 4, iterations_in_epoch = 100, database='market'):
assert batch_size%samples_per_class==0, F'batch size must be a {samples_per_class} multiple.'
if database == 'market':
class_nbr = 751
else:
raise Exception('Unsuported database yet')
all_datasets = [train_dataset.filter(predicate) for _ in range(class_nbr)] # Every element of this array is a dataset of one class
for i in range(iterations_in_epoch):
choice = tf.random.uniform(
shape=(batch_size//samples_per_class,),
minval=0,
maxval=class_nbr,
dtype=tf.dtypes.int64,
) # Which classes will be in batch
choice = tf.data.Dataset.from_tensor_slices(tf.concat([choice for _ in range(4)], axis=0)) # Exactly 4 picture from each class in the batch
batch = tf.data.experimental.choose_from_datasets(all_datasets, choice)
if i==0:
all_batches = batch
else:
all_batches = all_batches.concatenate(batch)
all_batches = all_batches.batch(batch_size)
return all_batches
It does what I want, however the returned dataset is extremely slow to iterate, making modele learning impossible. As per this thread, I understood that I needed to decorate custom_shuffle with #tf.function, as the one commented out. However, when doing so, it raises the following error:
Traceback (most recent call last):
File "training.py", line 137, in <module>
main()
File "training.py", line 80, in main
train_dataset = get_dataset(TRAINING_FILENAMES, IMG_SIZE, BATCH_SIZE, database=database, func_type='train')
File "E:\Morgan\TransReID_TF\tfr_to_dataset.py", line 260, in get_dataset
dataset = custom_shuffle(dataset, batch_size)
File "D:\Programs\Anaconda3\envs\AlignedReID_TF\lib\site-packages\tensorflow\python\eager\def_function.py", line 780, in __call__
result = self._call(*args, **kwds)
File "D:\Programs\Anaconda3\envs\AlignedReID_TF\lib\site-packages\tensorflow\python\eager\def_function.py", line 846, in _call
return self._concrete_stateful_fn._filtered_call(canon_args, canon_kwds) # pylint: disable=protected-access
File "D:\Programs\Anaconda3\envs\AlignedReID_TF\lib\site-packages\tensorflow\python\eager\function.py", line 1843, in _filtered_call
return self._call_flat(
File "D:\Programs\Anaconda3\envs\AlignedReID_TF\lib\site-packages\tensorflow\python\eager\function.py", line 1923, in _call_flat
return self._build_call_outputs(self._inference_function.call(
File "D:\Programs\Anaconda3\envs\AlignedReID_TF\lib\site-packages\tensorflow\python\eager\function.py", line 545, in call
outputs = execute.execute(
File "D:\Programs\Anaconda3\envs\AlignedReID_TF\lib\site-packages\tensorflow\python\eager\execute.py", line 59, in quick_execute
tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
tensorflow.python.framework.errors_impl.InternalError: No unary variant device copy function found for direction: 1 and Variant type_index: class tensorflow::data::`anonymous namespace'::DatasetVariantWrapper
[[{{node BatchDatasetV2/_206}}]] [Op:__inference_custom_shuffle_11485]
Function call stack:
custom_shuffle
Which I don't understand, and don't see how to fix.
Is there something I'm doing wrong?
PS: I'm aware the lack of minimal code to reproduce this behavior makes it hard to debug, I'll try to provide some as soon as possible.
I used np.genfromtxt to load a txt file.
Below is the code
datasets = np.genfromtxt('X&YTrainingsets1.txt',delimiter="")
It happens that I got an error
ValueError: Some errors were detected !
Line #1162 (got 5 columns instead of 7)
Line #1163 (got 2 columns instead of 7)
So I had to look through the txt file and I found this. So how do I solve this problem?. Thanks
I used
import pandas as pd
datasets = pd.read_csv('X&YTrainingsets1.txt',delimiter='\s+',index_col=False,header=None)
for i in range(0,len(datasets)):
if datasets.iloc[i].isnull().sum() == 2:
datasets.loc[i][5],datasets.loc[i][6] = datasets.loc[i+1][0],datasets.loc[i+1][1]
elif datasets.iloc[i].isnull().sum() == 1:
datasets.loc[i][6] = datasets.loc[i+1][0]
I am fairly new to IDL, and I am trying to write a code that will take a MODIS HDF file (level three data MOD14A1 and MYD14A1 to be specific), read the array, and then write the data from the array preferably to a csv file, but ASCII would work, too. I have code that will allow me to do this for one file, but I want to be able to do this for multiple files. Essentially, I want it to read one HDF array, write it to a csv, move to the next HDF file, and then write that array to the same csv file in the next row. Any help here would be greatly appreciated. I've supplied the code I have so far to do this with one file.
filename = dialog_pickfile(filter = filter, path = path, title = title)
csv_file = 'Data.csv'
sd_id = HDF_SD_START(filename, /READ)
; read "FirePix", "MaxT21"
attr_index = HDF_SD_ATTRFIND(sd_id, 'FirePix')
HDF_SD_ATTRINFO, sd_id, attr_index, DATA = FirePix
attr_index = HDF_SD_ATTRFIND(sd_id, 'MaxT21')
HDF_SD_ATTRINFO, sd_id, attr_index, DATA = MaxT21
index = HDF_SD_NAMETOINDEX(sd_id, 'FireMask')
sds_id = HDF_SD_SELECT(sd_id, index)
HDF_SD_GETDATA, sds_id, FireMask
HDF_SD_ENDACCESS, sds_id
index = HDF_SD_NAMETOINDEX(sd_id, 'MaxFRP')
sds_id = HDF_SD_SELECT(sd_id, index)
HDF_SD_GETDATA, sds_id, MaxFRP
HDF_SD_ENDACCESS, sds_id
HDF_SD_END, sd_id
help, FirePix
print, FirePix, format = '(8I8)'
print, MaxT21, format = '("MaxT21:", F6.1, " K")'
help, FireMask, MaxFRP
WRITE_CSV, csv_file, FirePix
After I run this, and choose the correct file, this is the output I am getting:
FIREPIX LONG = Array[8]
0 4 0 0 3 12 3 0
MaxT21: 402.1 K
FIREMASK BYTE = Array[1200, 1200, 8]
MAXFRP LONG = Array[1200, 1200, 8]
The "FIREPIX" array is the one I want stored into a csv.
Thanks in advance for any help!
Instead of using WRITE_CSV, it is fairly simple to use the primitive IO routines to write a comma-separated array, i.e.:
openw, lun, csv_file, /get_lun
; the following line produces a single line the output CSV file
printf, lun, strjoin(strtrim(firepix, 2), ', ')
; TODO: do the above line as many times as necessary
free_lun, sun
Hiya i have made a program that stores the player name and strength..Here is the code:
data = {
"PLAYER":name2,
"STRENGTH":str(round(strength, 2)),
}
with open("data2.txt", "w", encoding="utf-8") as file:
file.write(repr(data))
file.close()
So this stores the data so what to i do if i wanna append/change the value after a certain action usch as a 'BATTLE'
Is it possible the get the variable of 'STRENGTH' and then change the number?
At the moment to read data from the external file 'DATA1.txt'i am using this code:
with open("data1.txt", "r", encoding="utf-8") as file:
data_string = file.readline()
data = eval(data_string)
# (data["STRENGTH"])
S1 = (float(data["STRENGTH"]))
file.close()
Now i can do something with the variable --> 'S1'
Here is the external text file 'data1.txt'
{'PLAYER': 'Oreo', 'STRENGTH': '11.75'}
... But i wanna change the strength value after a "battle" many thanks
Maybe you're not understanding Python dict semantics?
Seems to me you're doing a lot of unnecessary things like S1 = (float(data['STRENGTH'])) to try to manipulate and change values when you could be doing really simple stuff.
>>> data = {'PLAYER': 'Oreo', 'STRENGTH': '11.75'}
>>> data['STRENGTH'] = float(data['STRENGTH'])
>>> data
{'PLAYER': 'Oreo', 'STRENGTH': 11.75}
>>> data['STRENGTH'] += 1
>>> data
{'PLAYER': 'Oreo', 'STRENGTH': 12.75}
Maybe you should give Native Data Types -- Dive Into Python 3 a read to see if it clears things up.