He Ma

@goffredogiordano The main error here is a TypeError rather than PyCUDA Error. The PyCUDA Error shows when code exits without proper context clean up. Not sure why the title of...

@DennisBang The `meta_clsloc.mat` file in the `ILSVRC2014_devkit/data` folder contains some meta data about the dataset, including ILSVRC2014_ID, WNID, words, gloss, num_children, wordiness_height and num_train_images (only the first 1000 synsets has...

@npit The downloaded weights are named in the same order as they appears in the `params` list. In this list, the weights of layers 2 (convpool_layer2) are named as W0_1...

@npit I haven't tried loading the snapshot for a while. It could give different result even because of newer version of cudnn. But I'm not sure. It's interesting to see...

@jiangqy What is your batch size and current learning rate?

@jiangqy @heipangpang Looks like you are running the single GPU train.py, then the problem is not related to weight exchanging. The cost should be around 6.9 initially. The unbounded cost...

@heipangpang Yes, this probably is the reason you got large cost. Make sure you set `use_data_layer` to False in config.yaml. Then the layer_input should be equal to x as shown...

@ducminhnguyen For computing with CPU, I think you need to run the code with THEANO_FLAGS=mode=FAST_RUN,device=cpu,floatX=float32 python train.py Follow the link below for Theano comparison on using CPU and GPU: http://deeplearning.net/software/theano/tutorial/using_gpu.html...

Hi @ksarker1205 , when the computing graph is compiled into a theano function, the forward and backward propagation happens each time the function is called at this line: https://github.com/uoguelph-mlrg/theano_alexnet/blob/master/train_funcs.py#L165 The...

@aryanbhardwaj Could you provide the full Traceback of the error? I just want to see the files that produce this error.