Computer to Machine Learning codes
Buy a computer with memory in SSD in pleace of a traditional HD is a good idea to run Machine Learning codes?
See also questions close to this topic
How can I compute the tensor in Pytorch efficiently?
I have a tensor
x.shape=(batch_size,10), now I want to take
x[i] = x[i]*x[i]*...*x[i] for i in range(batch_size)
Here is my code:
for i in range(batch_size): for k in range(1, 10): x[i] = x[i] * x[i][k]
But when I implement this in
loss.backward(), the speed of back-propagation is very slow. Why is it slow and is there any way to implement it efficiently?
How do I get all Gini indices in my decision tree?
I have made a decision tree using sklearn, here, under the SciKit learn DL package, viz.
How do I get the gini indices for all possible nodes at each step?
graphvizonly gives me the gini index of the node with the lowest gini index, ie the node used for split.
For example, the image below (from
graphviz) tells me the gini score of the Pclass_lowVMid right index which is 0.408, but not the gini index of the Pclass_lower or Sex_male at that step. I just know the Gini index of Pclass_lower and Sex_male must be greater than (0.408*0.7 + 0) but that's it.
Fighting against overfitting in an RNN model
We are currently trying to use an RNN model to build a classifier using text features. Our final accuracy on the training data is 87% but our accuracy on validation data flat out at 57% which is clearly overfitting. We think that the reason for overfitting is because of small data size since we only have about 4000 entries. What can we do to fix to the problem, we have also thought about doing data augmentation but all we can find is replacing words with synonyms which it wouldn't work in our case. Here's our code for the model and thank you in advance.
model = Sequential() model.add(Embedding(num_vocab+1,32)) model.add(SimpleRNN(64)) model.add(Dense(num_classes, activation='softmax')) model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['acc']) history = model.fit(f_train, cause_train, epochs=10, batch_size=50, validation_split=0.2)
- Internet explorer 11 (on windows 10) text stacking issue
micro pc using mobile screen
I'd like to get a micro-PC (a bar-of-soap sized "desktop" computer) or, alternatively, an HDMI stick (a computer that looks like a large pendrive and normally plugs into a TV) to use when going places. Is there a way to use a mobile phone (Android, Samsung Galaxy S7 Edge) as its screen? Either by connecting to its USB port or wirelessly? What about a bluetooth keyboard? I know there are ways to connect a mobile as a remote desktop, but the one I have tried (Google Chrome remote desktop) relies on first "calling Mom" (connecting to the home infrastructure of its maker), to find the list of my remotely-controllable PCs. Can I do remote desktopping when not connected to the Internet, say I'm on a desert island with just the micro-PC and my phone, no GSM coverage, no nothing? FWIW the micro PC will be running Windows (probably 10, maybe 7), and MS Office is a must for my purposes. TIA for all answers and suggestions. Should this not be the right question for this forum, apologies and please tell me where to go.
Alienware Aurora Crashes
I have an Alienware Aurora R4 which has been sporadically crashing for a year now. Sometimes more than others. I believe that I figured out if a USB cord or Ethernet cord gets pulled is when it crashes, all these ports specifically live on the same metal riser box. This specific riser on the back of the board has 1 Gigabit Ethernet and 2 USB 2.0 ports. I was going to disable it in the bios and get a PCI-e card for my Ethernet, I actually found a one star tech brand with a (realtech controller) which doubles and also has 3 USB 3.0 ports in addition to the Ethernet. I was hoping if I buy this and not use that riser that my problem will go away. I guess my question is there any way to repair this or at least confirm my suspicion that I am correct about the issue, generally if it crashes I will go back there and try to put pressure on the metal riser towards the board with the power unplugged, then boot the computer. I have gone months without a crash, generally when my computer is stationary and I haven't pulled it out or opened it up in a long time. It's a bit older, I would like to avoid replacing the motherboard, though it does have an i7 and is pretty fast. Any ideas? Am I on the right track if I buy this realtech PCI-e Ethernet / USB 3.0 card?
What is the best alternative for bandwidth limited applications for storage?
If have application that wants best bandwidth possible and I have to store my data either in a external server that is connected to my application over network or ssd in the system that is running the application, what will be the better option?
PS: I got asked this question in an interview and my answer was SSD but later realized that networks speeds can go to 1Tbps, so I am not sure which one is better now.
Performance Issue between Two Hard Driver
If you have two same computer with installed application Visual Studio 2017.
The diffence between them is the harddrive, the first one is a regular harddrive 1 TB and the second computer has a SSD harddrive with 1 TB.
How difference will be the performance issue when you use VS 2017?
About the installation of gentoo in samsung 960pro SSD
When I tried to install gentoo on my computer.I found that gentoo could not find the samsung 960pro SSD, but could detect the other two disks.If anyone else has had this problem.Please tell me how to solve it.Thank you very much.I may not express myself very clearly, because English is not my first mother tongue. Gentoo: install - amd64 - minimal - 20181129 t214502z. Iso