i am currently trying to tune the svm function in the e1071 package for R. my input is genomic data (that is each attribute takes a value in the set {-1, 0, 1}) and none of the four kernels currently offered in the package is really good for this kind of data --- i would like to use Hamming distance as my kernel instead.
the svm function, it seems, is written in C++. i have downloaded the source via
download.packages(pkgs = "e1071",
destdir = ".",
type = "source")
found the svm.cpp file containing code for the function and the corresponding kernel portion, where i can potentially add my own custom kernel. has anyone tried doing this? is it possible to do this? once i've finished modifying svm.cpp (provided i figure out how..), how do i make the package "see" the modified file?
You can modify the existing kernel.
I changed the return statement of radial kernel to make the changes..
You can try with that
Related
I'm trying implementing deep learning model into TensorRT runtime. The model conversion step is done quite OK and i'm pretty sure about it.
Now there's 2 parts i'm currently struggle with is memCpy data from host To Device (like openCV to Trt) and get the right output shape in order to get the right data. So my questions is:
How actually a shape of input dims relate with memory buffer. What is the difference when the model input dims is NCHW and NHWC, so when i read a openCV image, it's NHWC and also the model input is NHWC, do i have to re-arange the buffer data, if Yes then what's the actual consecutive memory format i have to do ?. Or simply what does the format or sequence of data that the engine are expecting ?
About the output (assume the input are correctly buffered), how do i get the right result shape for each task (Detection, Classification, etc..)..
Eg. an array or something look similar like when working with python .
I read Nvidia docs and it's not beginner-friendly at all.
//Let's say i have a model thats have a dynamic shape input dim in the NHWC format.
auto input_dims = nvinfer1::Dims4{1, 386, 342, 3}; //Using fixed H, W for testing
context->setBindingDimensions(input_idx, input_dims);
auto input_size = getMemorySize(input_dims, sizeof(float));
// How do i format openCV Mat to this kind of dims and if i encounter new input dim format, how do i adapt to that ???
And the expected output dims is something like (1,32,53,8) for example, the output buffer result in a pointer and i don't know what's the sequence of the data to reconstruct to expected array shape.
// Run TensorRT inference
void* bindings[] = {input_mem, output_mem};
bool status = context->enqueueV2(bindings, stream, nullptr);
if (!status)
{
std::cout << "[ERROR] TensorRT inference failed" << std::endl;
return false;
}
auto output_buffer = std::unique_ptr<int>{new int[output_size]};
if (cudaMemcpyAsync(output_buffer.get(), output_mem, output_size, cudaMemcpyDeviceToHost, stream) != cudaSuccess)
{
std::cout << "ERROR: CUDA memory copy of output failed, size = " << output_size << " bytes" << std::endl;
return false;
}
cudaStreamSynchronize(stream);
//How do i use this output_buffer to form right shape of output, (1,32,53,8) in this case ?
Could you please edit your question and tell us which model you're using if it's a commonly known NN, prehaps one we can download to test locally?
Then, the answer since it doesn't depend on the model (even though it would help to answer)
How actually a shape of input dims relate with memory buffer
If the input is NxCxHxW, you need to allocate N*C*H*W*sizeof(float) memory for that on your CPU and GPU. To be more precise, you need to allocate space on GPU for all the bindings and on CPU for only input and output bindings.
when i read a openCV image, it's NHWC and also the model input is NHWC, do i have to re-arange the buffer data
No, you do not have to re-arrange the buffer data. If you would have to change between NHWC and NCHW you can check this or google 'opencv NHWC to NHCW'.
Full working code example here, especially this function.
Or simply what does the format or sequence of data that the engine are expecting ?
This depends on how the neural network was trained. You should in general know exactly which kind of preprocessing and image data formats have been used to train the NN. You should even use the same libraries to load images and process them if possible. It's an open problem in ML: if you try to replicate results of some papers and use their models but they haven't open sourced the preprocessing you might get worse results. In the "worst" case you can implement both NHCW and NCHW and test which of them works.
About the output (assume the input are correctly buffered), how do i get the right result shape for each task (Detection, Classification, etc..).. Eg. an array or something look similar like when working with python .
This question clearly requires me to understand which NNs you are referring to. But I myself do the following:
Load the TensorRT .engine file in my code like this and deserialize like this
Print the bindings like this
Then I know the size of the input binding or bindings if there are many inputs, and the size of the output binding or bindings if there are many outputs.
This way you know the right result shape for each task. I hope this answered your question. If not, please add detailed comments and edit your post to be more precise. Thank you.
I read Nvidia docs and it's not beginner-friendly at all.
Yes I agree. You're better of searching TensorRT c++ (or Python) repositories from Github and studying their code. Have you seen TensorRT samples? It doesn't really take many lines of code to implement TensorRT inference.
In R the cumulative distribution function for the binomial distribution is called via an underlying C/C++ function called C_pbinom. I am trying to find the underlying code for this algorithm, so that I can find out what algorithm this function uses to compute the cumulative distribution function. Unfortunately, I have not been successful in finding the underlying code, nor any information on the algorithm that is used.
My question: How do I find the underlying C/C++ code for the function C_pbinom. Alternatively, is there any information source available showing the algorithm used by this function?
What I have done so far: Calling pbinom in R gives the following details:
function (q, size, prob, lower.tail = TRUE, log.p = FALSE)
.Call(C_pbinom, q, size, prob, lower.tail, log.p)
<bytecode: 0x000000000948c5a0>
<environment: namespace:stats>
I have located and opened the underlying NAMESPACE file in the stats library. This file lists various functions, including the pbinom function, but does not give code for the C_pbinom function, nor any pointer to where it can be found. I have also read a related answer on finding source code in R, and an article here on "compiling source codes", but neither has been of sufficient assistance to let me find the code. At this point I have his a dead end.
I went to the Github mirror for the R source code, searched for pbinom, and filtered to C: that got me here. The meat of the function is simply
pbeta(p, x + 1, n - x, !lower_tail, log_p)
This is invoking the incomplete beta function (= CDF of the Beta distribution): it means you need to in turn look up the pbeta function in the code: here, it says that the code is "a wrapper for TOMS708" , which is in src/nmath/toms708.c and described in a little more detail here (google "TOMS 708") ... original code here.
The full reference is here: Didonato and Morris, Jr.,
ACM Transactions on Mathematical Software (TOMS), Volume 18 Issue 3, Sept. 1992, Pages 360-373.
I'd like to build and train a multi-layer LSTM model (stateIsTuple=True) in python, and then load and use it in C++. But I'm having a hard time figuring out how to feed and fetch states in C++, mainly because I don't have string names which I can reference.
E.g. I put the initial state in a named scope such as
with tf.name_scope('rnn_input_state'):
self.initial_state = cell.zero_state(args.batch_size, tf.float32)
and this appears in the graph as below, but how can I feed to these in C++?
Also, how can I fetch the current state in C++? I tried the graph construction code below in python but I'm not sure if it's the right thing to do, because last_state should be a tuple of tensors, not a single tensor (though I can see that the last_state node in tensorboard is 2x2x50x128, which sounds like it just concatenated the states as I have 2 layers, 128 rnn size, 50 mini batch size, and lstm cell - with 2 state vectors).
with tf.name_scope('outputs'):
outputs, last_state = legacy_seq2seq.rnn_decoder(inputs, self.initial_state, cell, loop_function=loop if infer else None)
output = tf.reshape(tf.concat(outputs, 1), [-1, args.rnn_size], name='output')
and this is what it looks like in tensorboard
Should I concat and split the state tensors so there is only ever one state tensor going in and out? Or is there a better way?
P.S. Ideally the solution won't involve hard-coding the number of layers (or rnn size). So I can just have four strings input_node_name, output_node_name, input_state_name, output_state_name, and the rest is derived from there.
I managed to do this by manually concatenating the state into a single tensor. I'm not sure if this is wise, since this is how tensorflow used to handle states, but is now deprecating that and switching to tuple states. Instead of setting state_is_tuple=False and risking my code being obsolete soon, I've added extra ops to manually stack and unstack the states to and from a single tensor. Saying that, it works fine both in python and C++.
The key code is:
# setting up
zero_state = cell.zero_state(batch_size, tf.float32)
state_in = tf.identity(zero_state, name='state_in')
# based on https://medium.com/#erikhallstrm/using-the-tensorflow-multilayered-lstm-api-f6e7da7bbe40#.zhg4zwteg
state_per_layer_list = tf.unstack(state_in, axis=0)
state_in_tuple = tuple(
# TODO make this not hard-coded to LSTM
[tf.contrib.rnn.LSTMStateTuple(state_per_layer_list[idx][0], state_per_layer_list[idx][1])
for idx in range(num_layers)]
)
outputs, state_out_tuple = legacy_seq2seq.rnn_decoder(inputs, state_in_tuple, cell, loop_function=loop if infer else None)
state_out = tf.identity(state_out_tuple, name='state_out')
# running (training or inference)
state = sess.run('state_in:0') # zero state
loop:
feed = {'data_in:0': x, 'state_in:0': state}
[y, state] = sess.run(['data_out:0', 'state_out:0'], feed)
Here is the full code if anyone needs it
https://github.com/memo/char-rnn-tensorflow
I am running libsvm through weka. Its output accuracy looks good to me, so I am planning to write a svm model by myself. However, weka didn't generate any training parameter, such as number of support vector. Therefore i cannot do anything. Searching the web, i found somebody said it would generate some parameters like the following:
optimization finished, #iter = 27
nu = 0.058475864943863545
obj = -1.871013102744184, rho = -0.19357337828800944
nSV = 9, nBSV = 0 `enter code here`
Total nSV = 9
but how come i didn't see any of them? any step that i missed? please help me. Thanks a lot.
Weka writes the output you mentioned to stderr.
So if you have started weka.sh or weka.bat from a terminal (or "command window" if you are on Windows), you should see that output appear in your terminal window after clicking "classify"
If you want to have access to this information via scripts, you can
redirect the output to a file and read in that file.
Here is how to edit the startup file weka.sh / weka.bat.
Edit this line (it is probably the last line) in order to write log info to a file instead of the terminal window:
java -cp $CP -Xmx8092m weka.gui.GUIChooser 2>>/opt/weka-stable/weka.log &
You can also add a properties file to your home directory to add more fine-grained behaviour.
https://weka.wikispaces.com/Properties+file
(You probably can also access information via the Weka Java API somehow, but you did not ask for that)
I want to preview TeX formulas in my User Interface. After a long time searching, it seems to me that there is no other possibility than
write the formula into a .tex file
call tex with system() and write a dvi file
call e.g. dvipng with system() and write a png file
load this file into the GUI
clean up(erase all these files).
I think that the performance of this way doing it is not a problem, since there are only formulas to render and not whole documents. But setting up the environment automatically for the TeX system seems to be a bigger problem.
So, is there a possibility to include TeX as an API in my program?
Thanks a lot!
Couldn't you encapsulate those steps in a single shell script (i.e. which takes the formula and png filename as arguments)? The script could then also handle setting up the environment for TeX. Your program just calls the script with the system() call.
There's a C API for TeX called MimeTeX but the resulting image is... not a nice as it could be.
If you're OK with Java, there's JLatexMath
And if you'd like a WPF version, one is under development at WPFMath
I'm not sure, but think MathType's Component will overkill.
Also have a look at sideshare and see flash video to get more information about sitmo, mathMagig, Edoboard and their API tools.
good lucks.
For Edoboard and Tutorsbox.com we do the following:
Keep a blacklist of LaTeX commands to avoid:
TEX_BLACKLIST = ["\\def", "\\let", "\\futurelet",
"\\newcommand", "\\renewcommand", "\\else", "\\fi", "\\write",
"\\input", "\\include", "\\chardef", "\\catcode", "\\makeatletter",
"\\noexpand", "\\toksdef", "\\every", "\\errhelp", "\\errorstopmode",
"\\scrollmode", "\\nonstopmode", "\\batchmode", "\\read", "\\csname",
"\\newhelp", "\\relax", "\\afterground", "\\afterassignment",
"\\expandafter", "\\noexpand", "\\special", "\\command", "\\loop",
"\\repeat", "\\toks", "\\output", "\\line", "\\mathcode", "\\name",
"\\item", "\\section", "\\mbox", "\\DeclareRobustCommand", "\\[", "\\]"];
We then do system call "latex and textopng".
That as an API REST plus some caching and here you go :)
As an upgrade we will soon convert those LaTeX images as SVG.
LyX is a TeX based Document Processor. As the application is open source you can inspect the C++ code to see how they deal with the problem you described.