{
"cells": [
{
"cell_type": "code",
"execution_count": 21,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T13:53:26.552385Z",
"start_time": "2018-10-03T13:53:26.342168Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"array([ 0.85, -0.95, -1. , -0.65, 0.3 , -0.65, 0.55, -0.05])"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import h5py\n",
"from sklearn.model_selection import train_test_split\n",
"\n",
"#now load this dataset \n",
"h5f = h5py.File('./datasets/s8_sio2tio2_v2.h5','r')\n",
"X = h5f['sizes'][:]\n",
"Y = h5f['spectrum'][:]\n",
"\n",
"\n",
"#create a train - test split of the dataset\n",
"x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=0.4, random_state=42)\n",
"\n",
"x_test = (x_test - 50.0)/20.0\n",
"\n",
"x_test[9]\n"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T13:41:28.745978Z",
"start_time": "2018-10-03T13:41:26.725681Z"
}
},
"outputs": [],
"source": [
"import numpy as np\n",
"import mxnet as mx\n",
"\n",
"# Step1: Load the model in MXNet\n",
"\n",
"# Use the same prefix and epoch parameters we used in save_mxnet_model API.\n",
"sym, arg_params, aux_params = mx.model.load_checkpoint(prefix='my_mod', epoch=0)\n",
"\n",
"# We use the data_names and data_shapes returned by save_mxnet_model API.\n",
"mod = mx.mod.Module(symbol=sym, \n",
" data_names=['/input_12'], \n",
" context=mx.gpu(), \n",
" label_names=None)\n",
"mod.bind(for_training=False, \n",
" data_shapes=[('/input_12', (1,8))], \n",
" label_shapes=mod._label_shapes)\n",
"mod.set_params(arg_params, aux_params, allow_missing=True) "
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T15:57:53.584259Z",
"start_time": "2018-10-03T15:57:53.321389Z"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"255.54943084716797\n"
]
}
],
"source": [
"import scnets as scn\n",
"import mxnet as mx\n",
"from mxnet import nd\n",
"import time\n",
"\n",
"\n",
"bsize = 4000\n",
"y_true = nd.array(y_test[9], ctx=mx.gpu())\n",
"reps = 1000\n",
"\n",
"start = time.time()\n",
"for icv in np.arange(reps):\n",
" x_t = x_test[0:bsize]\n",
" res2 = mod.predict(x_t)\n",
" err = nd.abs(y_true - res2)/y_true\n",
" finerr = 100*nd.mean(err)\n",
"end = time.time()\n",
"print(1000*(end - start)) "
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T08:45:32.141435Z",
"start_time": "2018-10-03T08:45:31.874129Z"
}
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using MXNet backend\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"0.17115497589111328\n"
]
}
],
"source": [
"import scnets as scn\n",
"import mxnet as mx\n",
"from mxnet import nd\n",
"import time\n",
"\n",
"size = np.random.randint(30, 71, (1,8))\n",
"#spec_ac = snlay.calc_spectrum(size, mats, lams)\n",
"size = (size - 50.0)/20.0\n",
"#size = np.expand_dims(size, axis = 0)\n",
"size.shape\n",
"\n",
"y_true = nd.array(y_test[9], ctx=mx.gpu())\n",
"\n",
"reps = 1000\n",
"start = time.time()\n",
"#res1 = model.predict(size)\n",
"#data_iter = mx.io.NDArrayIter(size, None, 1)\n",
"for icv in np.arange(reps):\n",
" size = np.random.randint(30, 71, (1,8))\n",
" size = (size - 50.0)/20.0\n",
" res2 = mod.predict(size)\n",
" err = nd.abs(y_true - res2)/y_true\n",
" finerr = nd.sum(err)\n",
" \n",
"end = time.time()\n",
"print(1000*(end - start)/reps) "
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T08:45:41.210914Z",
"start_time": "2018-10-03T08:45:41.205195Z"
}
},
"outputs": [],
"source": [
"import scnets as scn\n",
"import mxnet as mx\n",
"from mxnet import nd\n",
"def give_loss(x_in, y_true, mxmod):\n",
" y_t = nd.array(y_true, ctx=mx.gpu())\n",
" res2 = mxmod.predict(x_in)\n",
" err = nd.abs(y_t - res2)/y_t\n",
" err2 = nd.sum(err).asscalar()\n",
" #ez2 = err2.copyto(mx.cpu())\n",
" return err2\n",
"\n",
"# err2cp = nd.zeros_like(err2, ctx=mx.cpu())"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T08:45:59.234753Z",
"start_time": "2018-10-03T08:45:57.287759Z"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"1.9373879432678223\n"
]
}
],
"source": [
"size = np.random.randint(30, 71, (1,8))\n",
"#spec_ac = snlay.calc_spectrum(size, mats, lams)\n",
"size = (size - 50.0)/20.0\n",
"\n",
"ans = 1.00\n",
"cd = give_loss(x_in=size, y_true=y_test[6], mxmod=mod)\n",
"\n",
"\n",
"reps = 1000\n",
"start = time.time()\n",
"for icv in np.arange(reps):\n",
" cc = give_loss(x_in=size, y_true=y_test[6], mxmod=mod)\n",
" #ans = cc.asscalar()\n",
"\n",
"end = time.time()\n",
"print(1000*(end - start)/reps) # "
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"ExecuteTime": {
"end_time": "2018-10-03T08:45:54.615638Z",
"start_time": "2018-10-03T08:45:54.610110Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"235.14212"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"cd"
]
},
{
"cell_type": "code",
"execution_count": 124,
"metadata": {
"ExecuteTime": {
"end_time": "2018-09-30T15:58:44.524220Z",
"start_time": "2018-09-30T15:58:44.518399Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"1.0"
]
},
"execution_count": 124,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x.__getitem__(0).asscalar()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.5.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}