Skip to content

Commit

Permalink
update outline
Browse files Browse the repository at this point in the history
  • Loading branch information
@aileli
aileli committed Oct 24, 2016
1 parent 5ddd436 commit da3ebef
Show file tree
Hide file tree
Showing 3 changed files with 40 additions and 46 deletions.
38 changes: 9 additions & 29 deletions python/outline.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,46 +8,26 @@
"\n",
"The outline of python notebooks.\n",
"\n",
"## Basic Concepts\n",
"## Tutorials\n",
"\n",
"- [MNIST](./tutorials/mnist.ipynb): Recognize handwritten digits with multilayer perceptrons and convolutional neural networks\n",
"- [Recognize image objects](./tutorials/predict_imagenet.ipynb) with pre-trained model on the full Imagenet dataset that containing more than 10M images and over 10K classes\n",
"- [Char-LSTM](./tutorials/char_lstm.ipynb): Generates Obama's speeches with character-level LSTM.\n",
"- [Matrix Factorization](./tutorials/matrix_factorization.ipynb): Recommend movies to users. \n",
"\n",
"These tutorials walk through how to use the basic components of MXNet.\n",
"## Basic Concepts\n",
"\n",
"- [NDArray](./basic/ndarray.ipynb): manipulating multi-dimensional array \n",
"- [Symbol](./basic/symbol.ipynb): symbolic expression for neural networks\n",
"- [Module](./basic/module.ipynb) : intermediate-level and high-level interface for neural network training and inference. \n",
"- [Loading data](./basic/data.ipynb) : feeding data into training/inference programs\n",
"- DOING. [Key-value Store]() : data communication for multi-device and multi-machines training \n",
"- DOING. [Initializer](./basic/initializer.ipynb) : various ways to intialize the parameters\n",
"- DOING. [Optimizer](./basic/optimizer.ipynb) : parameters updaters such as `sgd` updater\n",
"- TODO. Metric : various metric to evaluate the progress\n",
"- [Mixed programming](./basic/mixed.ipynb): developing training algorithms by using NDArray and Symbol together.\n",
" \n",
"\n",
"## Neural Networks\n",
"\n",
"How to implement various neural networks. \n",
"\n",
"TODO. Convolution Neural Networks\n",
"\n",
"Recurrent Neural networks\n",
"- [LSTM](./rnn/lstm.ipynb) build LSTM from scratch\n",
"\n",
"## How To\n",
"## How Tos\n",
"\n",
"- [Use pretrained models for prediction and feature extraction](./how_to/predict.ipynb) \n",
"\n",
"## Applications\n",
"\n",
"Complete examples for various applications. \n",
"\n",
"### Image Classification\n",
"\n",
"- [Convolutional neural network for Written Digit Recognition](./cnn/mnist.ipynb) train and predict on the mnist datasets\n",
"\n",
"### Recommendation Systems\n",
"\n",
"- [Matrix Factorization](./recommendation_systems/matrix_factorization.ipynb) writing a basic matrix factorization algorithm\n",
"- [Collaborative Deep Learning](./recommendation_systems/cdl/collaborative-dl.ipynb) state-of-the-art algorithm in KDD 15\n"
"- [Use pretrainde models for fine-tune](./how_to/finetune.ipynb)\n"
]
},
{
Expand Down
46 changes: 30 additions & 16 deletions python/tutorials/char_lstm.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -145,7 +145,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 4,
"metadata": {
"collapsed": false
},
Expand Down Expand Up @@ -192,7 +192,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 5,
"metadata": {
"collapsed": false
},
Expand All @@ -202,15 +202,21 @@
"output_type": "stream",
"text": [
"INFO:root:Start training with [gpu(0)]\n",
"INFO:root:Epoch[0] Batch [5]\tSpeed: 85.93 samples/sec\tTrain-Perplexity=64.713810\n",
"INFO:root:Epoch[0] Batch [10]\tSpeed: 71.62 samples/sec\tTrain-Perplexity=34.562343\n",
"INFO:root:Epoch[0] Batch [15]\tSpeed: 71.55 samples/sec\tTrain-Perplexity=26.872259\n",
"INFO:root:Epoch[0] Batch [20]\tSpeed: 71.57 samples/sec\tTrain-Perplexity=26.383636\n",
"INFO:root:Epoch[0] Batch [25]\tSpeed: 71.50 samples/sec\tTrain-Perplexity=26.499681\n",
"INFO:root:Epoch[0] Batch [30]\tSpeed: 71.49 samples/sec\tTrain-Perplexity=24.991160\n",
"INFO:root:Epoch[0] Batch [35]\tSpeed: 71.51 samples/sec\tTrain-Perplexity=23.675805\n",
"INFO:root:Epoch[0] Batch [40]\tSpeed: 71.55 samples/sec\tTrain-Perplexity=24.152127\n",
"INFO:root:Epoch[0] Batch [45]\tSpeed: 71.43 samples/sec\tTrain-Perplexity=23.883328\n"
"INFO:root:Epoch[0] Batch [20]\tSpeed: 74.77 samples/sec\tTrain-Perplexity=38.916073\n",
"INFO:root:Epoch[0] Batch [40]\tSpeed: 71.74 samples/sec\tTrain-Perplexity=24.287313\n",
"INFO:root:Epoch[0] Batch [60]\tSpeed: 71.72 samples/sec\tTrain-Perplexity=23.523363\n",
"INFO:root:Epoch[0] Batch [80]\tSpeed: 71.72 samples/sec\tTrain-Perplexity=23.010193\n",
"INFO:root:Epoch[0] Batch [100]\tSpeed: 71.81 samples/sec\tTrain-Perplexity=23.048955\n",
"INFO:root:Epoch[0] Batch [120]\tSpeed: 71.83 samples/sec\tTrain-Perplexity=22.809207\n",
"INFO:root:Epoch[0] Batch [140]\tSpeed: 71.83 samples/sec\tTrain-Perplexity=22.798320\n",
"INFO:root:Epoch[0] Batch [160]\tSpeed: 71.83 samples/sec\tTrain-Perplexity=22.409329\n",
"INFO:root:Epoch[0] Batch [180]\tSpeed: 71.88 samples/sec\tTrain-Perplexity=22.727832\n",
"INFO:root:Epoch[0] Batch [200]\tSpeed: 71.88 samples/sec\tTrain-Perplexity=22.405647\n",
"INFO:root:Epoch[0] Batch [220]\tSpeed: 71.87 samples/sec\tTrain-Perplexity=22.398161\n",
"INFO:root:Epoch[0] Batch [240]\tSpeed: 71.85 samples/sec\tTrain-Perplexity=22.343282\n",
"INFO:root:Epoch[0] Resetting Data Iterator\n",
"INFO:root:Epoch[0] Time cost=115.364\n",
"INFO:root:Saved checkpoint to \"obama-0001.params\"\n"
]
}
],
Expand Down Expand Up @@ -243,7 +249,7 @@
"\n",
"model.fit(X=data_train,\n",
" eval_metric=mx.metric.np(Perplexity),\n",
" batch_end_callback=mx.callback.Speedometer(batch_size, 5),\n",
" batch_end_callback=mx.callback.Speedometer(batch_size, 20),\n",
" epoch_end_callback=mx.callback.do_checkpoint(\"obama\"))"
]
},
Expand All @@ -258,7 +264,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 6,
"metadata": {
"collapsed": true
},
Expand Down Expand Up @@ -324,7 +330,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 7,
"metadata": {
"collapsed": true
},
Expand Down Expand Up @@ -356,11 +362,19 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 8,
"metadata": {
"collapsed": false
},
"outputs": [],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The United States of America. That's why I'm running for President.The first place we can do better than that they can afford to get the that they can afford to differ on the part of the political settlement. The second part of the problem is that the consequences would have to see the chance to starthe country that we can start by the challenges of the American people. The American people have been talking about how to compete with the streets of San Antonio who are serious about the courage to come together as one people. That the American people have been trying to get there. And they say\n"
]
}
],
"source": [
"seq_length = 600\n",
"input_ndarray = mx.nd.zeros((1,))\n",
Expand Down
2 changes: 1 addition & 1 deletion python/how_to/predict_demo.ipynb → python/tutorials/predict_imagenet.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
"source": [
"# Predict with pre-trained models\n",
"\n",
"This is a demo for predicting with a pre-trained model on the full imagenet dataset, which contains over 10 million images and 10 thousands classes. For more detailed explanation, please refer to [predict.ipynb](./predict.ipynb).\n",
"This is a demo for predicting with a pre-trained model on the full imagenet dataset, which contains over 10 million images and 10 thousands classes. For more detailed explanation, please refer to [how_to/predict.ipynb](../how_to/predict.ipynb).\n",
"\n",
"We first load the pre-trained model."
]
Expand Down

0 comments on commit da3ebef

Please sign in to comment.