"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# We can take out some of the outliers to get a more detailed view\n",
"plt.hist(row_size[row_size<1000])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We could do the same for the columns. This is just to make sure we don't have any super wide/big images, etc."
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(8178, 10357)"
]
},
"execution_count": 22,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Get the size of the training and test data sets\n",
"len(data.trn_ds), len(data.test_ds)"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(120,\n",
" ['affenpinscher',\n",
" 'afghan_hound',\n",
" 'african_hunting_dog',\n",
" 'airedale',\n",
" 'american_staffordshire_terrier'])"
]
},
"execution_count": 23,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Get the num of classes\n",
"len(data.classes), data.classes[:5]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Setting up the model\n",
"Now that we've done some data exploration, we can set up the model."
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [],
"source": [
"# Convenience method\n",
"def get_data(size, batch_size):\n",
" transforms = tfms_from_model(architecture, size, aug_tfms=transforms_side_on, max_zoom=1.1)\n",
" data = ImageClassifierData.from_csv(PATH, 'train', f'{PATH}labels.csv', test_name='test', num_workers=4,\n",
" val_idxs=val_idxs, suffix='.jpg', tfms=transforms, bs=batch_size)\n",
" return data if size>300 else data.resize(340, 'tmp')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Precompute\n",
"Set precompute to do and do our initial training for the final layers."
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "09b1ffa050fd42e5a160dda8fb3691d0",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, max=6), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"data = get_data(size, batch_size)"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [],
"source": [
"learn = ConvLearner.pretrained(architecture, data, precompute=True)"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "44c2750464804264924012874f58dea9",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=5), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"[0. 0.92608 0.39313 0.90853] \n",
"[1. 0.43449 0.36583 0.91092] \n",
"[2. 0.30892 0.30343 0.91331] \n",
"[3. 0.25296 0.31098 0.91906] \n",
"[4. 0.19217 0.30561 0.91667] \n",
"\n"
]
}
],
"source": [
"learn.fit(1e-2, 5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Augment\n",
"Next we turn precompute off so we can use data augmentations (transformations)."
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [],
"source": [
"from sklearn import metrics"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "83ece9a54156418aa5405735ecaaaf01",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, max=6), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"data = get_data(size, batch_size)"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [],
"source": [
"learn = ConvLearner.pretrained(architecture, data, precompute=True, ps=0.5)"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "2eb95b45c4ad482ba9396a6ea7506f9b",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=2), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"[0. 1.13687 0.4544 0.90079] \n",
"[1. 0.54595 0.39578 0.91236] \n",
"\n"
]
}
],
"source": [
"learn.fit(1e-2, 2)"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [],
"source": [
"learn.precompute=False"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "214265eb39604482a003a6a6b19ed457",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=5), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"[0. 0.4454 0.27167 0.92241] \n",
"[1. 0.40716 0.26172 0.92672] \n",
"[2. 0.38718 0.25009 0.92522] \n",
"[3. 0.35889 0.23875 0.93343] \n",
"[4. 0.33432 0.23403 0.93151] \n",
"\n"
]
}
],
"source": [
"# Use the learning rate finder to check the learning rate is optimal\n",
"# Tried 5 epochs to check the accuracy was getting better, then stuck with that\n",
"learn.fit(1e-2, 5, cycle_len=1)"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [],
"source": [
"# Save our weights\n",
"learn.save('224_pre')"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [],
"source": [
"learn.load('224_pre')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Increase size\n",
"Then we increase the size of the images.\n",
"\n",
"A tip Jeremy has is to train on smaller images, then train on larger images, so the images are different. It prevents overfitting. Also, training on the small images is faster, so do it first."
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "83598ce32fbf42c5818c57cad556af6a",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, max=6), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"learn.set_data(get_data(299, batch_size))\n",
"learn.freeze()"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "2bd07488b7ec46acab974f922191a503",
"version_major": 2,
"version_minor": 0
},
"text/html": [
"Failed to display Jupyter Widget of type HBox.
\n",
"\n",
" If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n",
" that the widgets JavaScript is still loading. If this message persists, it\n",
" likely means that the widgets JavaScript library is either not installed or\n",
" not enabled. See the Jupyter\n",
" Widgets Documentation for setup instructions.\n",
"
\n",
"\n",
" If you're reading this message in another frontend (for example, a static\n",
" rendering on GitHub or NBViewer),\n",
" it may mean that your frontend doesn't currently support widgets.\n",
"
\n"
],
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=1), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"[0. 0.34207 0.22987 0.93343] \n",
"\n"
]
}
],
"source": [
"# If the training loss is greater than the validation loss, we're underfiting\n",
"# We can up the cycle_mult parameter to correct this\n",
"# (We'd do 3 epochs normally)\n",
"learn.fit(1e-2, 1, cycle_len=1, cycle_mult=2)"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" \r"
]
}
],
"source": [
"# Then try using test time augmentation to see if that gives better results\n",
"log_preds, y = learn.TTA()"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.23592007515597008"
]
},
"execution_count": 40,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"probs = np.mean(np.exp(log_preds), 0)"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.9339530332681018"
]
},
"execution_count": 42,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"accuracy_np(probs, y)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We didn't unfreeze and try training more. This data is from image-net, that tuning the layers that were already pretrained on image-net anyway didn't make any difference."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Create the submission\n",
"Now that we've trained our model, we want to submit it to Kaggle. In the evaluation section of the competition, you can see the requested format:\n",
"\n",
"`id,affenpinscher,afghan_hound,..,yorkshire_terrier\n",
"000621fb3cbb32d8935728e48679680e,0.0083,0.0,...,0.0083\n",
"etc.`"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" id | \n",
" affenpinscher | \n",
" afghan_hound | \n",
" african_hunting_dog | \n",
" airedale | \n",
" american_staffordshire_terrier | \n",
" appenzeller | \n",
" australian_terrier | \n",
" basenji | \n",
" basset | \n",
" ... | \n",
" toy_poodle | \n",
" toy_terrier | \n",
" vizsla | \n",
" walker_hound | \n",
" weimaraner | \n",
" welsh_springer_spaniel | \n",
" west_highland_white_terrier | \n",
" whippet | \n",
" wire-haired_fox_terrier | \n",
" yorkshire_terrier | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 000621fb3cbb32d8935728e48679680e | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" ... | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
"
\n",
" \n",
" | 1 | \n",
" 00102ee9d8eb90812350685311fe5890 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" ... | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
"
\n",
" \n",
" | 2 | \n",
" 0012a730dfa437f5f3613fb75efcd4ce | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" ... | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
"
\n",
" \n",
" | 3 | \n",
" 001510bc8570bbeee98c8d80c8a95ec1 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" ... | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
"
\n",
" \n",
" | 4 | \n",
" 001a5f3114548acdefa3d4da05474c2e | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" ... | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
" 0.008333 | \n",
"
\n",
" \n",
"
\n",
"
5 rows × 121 columns
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" id | \n",
" affenpinscher | \n",
" afghan_hound | \n",
" african_hunting_dog | \n",
" airedale | \n",
" american_staffordshire_terrier | \n",
" appenzeller | \n",
" australian_terrier | \n",
" basenji | \n",
" basset | \n",
" ... | \n",
" toy_poodle | \n",
" toy_terrier | \n",
" vizsla | \n",
" walker_hound | \n",
" weimaraner | \n",
" welsh_springer_spaniel | \n",
" west_highland_white_terrier | \n",
" whippet | \n",
" wire-haired_fox_terrier | \n",
" yorkshire_terrier | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" de084b830010b6107215fef5d4a75b94 | \n",
" 0.000018 | \n",
" 0.000110 | \n",
" 0.000039 | \n",
" 0.000007 | \n",
" 0.000008 | \n",
" 0.000004 | \n",
" 0.000001 | \n",
" 7.672430e-07 | \n",
" 3.367392e-06 | \n",
" ... | \n",
" 1.198677e-06 | \n",
" 0.000002 | \n",
" 6.501280e-06 | \n",
" 1.887059e-06 | \n",
" 0.000010 | \n",
" 4.076534e-06 | \n",
" 0.000017 | \n",
" 0.000051 | \n",
" 0.000002 | \n",
" 1.166302e-06 | \n",
"
\n",
" \n",
" | 1 | \n",
" 6b423ca7020e70eb05732843c5d2bad1 | \n",
" 0.000029 | \n",
" 0.000325 | \n",
" 0.000021 | \n",
" 0.000039 | \n",
" 0.000005 | \n",
" 0.000002 | \n",
" 0.000039 | \n",
" 4.944084e-06 | \n",
" 1.226995e-05 | \n",
" ... | \n",
" 1.103837e-05 | \n",
" 0.000002 | \n",
" 1.619151e-06 | \n",
" 5.791285e-07 | \n",
" 0.000008 | \n",
" 8.745586e-06 | \n",
" 0.000345 | \n",
" 0.000012 | \n",
" 0.000052 | \n",
" 2.870508e-05 | \n",
"
\n",
" \n",
" | 2 | \n",
" 74aa7e201e0e93e13e87b986a7d31839 | \n",
" 0.000078 | \n",
" 0.000045 | \n",
" 0.000026 | \n",
" 0.003559 | \n",
" 0.000409 | \n",
" 0.000041 | \n",
" 0.000045 | \n",
" 1.390945e-04 | \n",
" 1.496679e-05 | \n",
" ... | \n",
" 5.939487e-05 | \n",
" 0.000068 | \n",
" 1.552802e-04 | \n",
" 7.329275e-05 | \n",
" 0.000020 | \n",
" 1.114965e-04 | \n",
" 0.000083 | \n",
" 0.000251 | \n",
" 0.243859 | \n",
" 3.139315e-05 | \n",
"
\n",
" \n",
" | 3 | \n",
" a079f72193264bc5685e5d28d7372680 | \n",
" 0.000064 | \n",
" 0.000024 | \n",
" 0.000049 | \n",
" 0.000095 | \n",
" 0.000048 | \n",
" 0.000002 | \n",
" 0.000177 | \n",
" 2.290131e-06 | \n",
" 1.020107e-05 | \n",
" ... | \n",
" 1.873471e-05 | \n",
" 0.000027 | \n",
" 1.161859e-05 | \n",
" 3.685182e-06 | \n",
" 0.000012 | \n",
" 2.911502e-05 | \n",
" 0.000159 | \n",
" 0.000013 | \n",
" 0.000835 | \n",
" 2.092581e-04 | \n",
"
\n",
" \n",
" | 4 | \n",
" 583f7580fa5fec1266331fcf83b76fd6 | \n",
" 0.000005 | \n",
" 0.000011 | \n",
" 0.000016 | \n",
" 0.000122 | \n",
" 0.000003 | \n",
" 0.000008 | \n",
" 0.000003 | \n",
" 1.663623e-06 | \n",
" 5.128858e-07 | \n",
" ... | \n",
" 9.443419e-07 | \n",
" 0.000006 | \n",
" 2.278522e-07 | \n",
" 1.967593e-07 | \n",
" 0.000003 | \n",
" 8.487252e-07 | \n",
" 0.000001 | \n",
" 0.000015 | \n",
" 0.000034 | \n",
" 8.924907e-07 | \n",
"
\n",
" \n",
"
\n",
"
5 rows × 121 columns
\n",
"
"
],
"text/plain": [
"/home/paperspace/fastai/courses/dl1/data/dogbreed/subm/submission.gz"
]
},
"execution_count": 53,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Get a link to the file to download\n",
"FileLink(f'{submission}submission.gz')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Predicting individual images\n",
"Sometimes you might want to predict just one image."
]
},
{
"cell_type": "code",
"execution_count": 54,
"metadata": {},
"outputs": [],
"source": [
"file_name = data.val_ds.fnames[0]"
]
},
{
"cell_type": "code",
"execution_count": 55,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'train/000bec180eb18c7604dcecc8fe0dba07.jpg'"
]
},
"execution_count": 55,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"file_name"
]
},
{
"cell_type": "code",
"execution_count": 56,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"execution_count": 56,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"Image.open(PATH+file_name).resize((150, 150))"
]
},
{
"cell_type": "code",
"execution_count": 57,
"metadata": {},
"outputs": [],
"source": [
"trn_tfms,val_tfms = tfms_from_model(architecture, size)"
]
},
{
"cell_type": "code",
"execution_count": 61,
"metadata": {},
"outputs": [],
"source": [
"# Prediction expects a mini-batch -- a collection, we by indexing with\n",
"# im[None] we essentially create a mini-batch just containing that one image\n",
"im = val_tfms(Image.open(PATH+file_name))\n",
"preds = learn.predict_array(im[None])\n",
"np.argmax(preds)"
]
},
{
"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": "3.6.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
![](\"https://raw.githubusercontent.com/pekoto/fast.ai/master/images/sym-links.jpg\")
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Dog Breeds\n",
"We will enter the [dog breed identification contest](https://www.kaggle.com/c/dog-breed-identification/).\n",
"\n",
"First get the data using the steps above."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"# Do our usual import and setup\n",
"%reload_ext autoreload\n",
"%autoreload 2\n",
"%matplotlib inline"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"# Set up our imports\n",
"from fastai.imports import *\n",
"from fastai.torch_imports import *\n",
"from fastai.transforms import *\n",
"from fastai.conv_learner import *\n",
"from fastai.model import *\n",
"from fastai.dataset import *\n",
"from fastai.sgdr import *\n",
"from fastai.plots import *"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"# Set up our paths, size, architecture, batch size, etc.\n",
"PATH=\"data/dogbreed/\"\n",
"size=224\n",
"architecture=resnext101_64\n",
"batch_size=58"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"After doing some initial set up, we do some data exploration to see what the downloaded data looks like."
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"label_csv = f'{PATH}labels.csv'\n",
"\n",
"# num of rows -1 (to account for the header)\n",
"n = len(list(open(label_csv)))-1\n",
"\n",
"# random 20% of rows to use as the validation set\n",
"# get cross validation indexes\n",
"val_idxs = get_cv_idxs(n)"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([2882, 4514, 7717, ..., 8922, 6774, 37])"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"val_idxs"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"labels.csv sample_submission.csv test tmp train\r\n"
]
}
],
"source": [
"# Take a look and see the files we have\n",
"!ls {PATH}"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"