Ranger-Mish-ImageWoof-Training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Below is the training settings used for ImageWoof 5 epoch, 128px.  You can kick off a run in the very next cell.\n",
    "#There are some other sample runs showing how some of the settings were adjusted...if you find better settings please let us know!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.969633</td>\n",
       "      <td>1.892525</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.872000</td>\n",
       "      <td>00:57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.745783</td>\n",
       "      <td>1.819594</td>\n",
       "      <td>0.468000</td>\n",
       "      <td>0.866000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.560068</td>\n",
       "      <td>1.549403</td>\n",
       "      <td>0.554000</td>\n",
       "      <td>0.934000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.429317</td>\n",
       "      <td>1.383480</td>\n",
       "      <td>0.634000</td>\n",
       "      <td>0.962000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.229336</td>\n",
       "      <td>1.179191</td>\n",
       "      <td>0.734000</td>\n",
       "      <td>0.970000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.927314</td>\n",
       "      <td>1.857006</td>\n",
       "      <td>0.422000</td>\n",
       "      <td>0.872000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.708304</td>\n",
       "      <td>1.728417</td>\n",
       "      <td>0.474000</td>\n",
       "      <td>0.914000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.556346</td>\n",
       "      <td>1.407383</td>\n",
       "      <td>0.628000</td>\n",
       "      <td>0.942000</td>\n",
       "      <td>00:25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.429676</td>\n",
       "      <td>1.316321</td>\n",
       "      <td>0.656000</td>\n",
       "      <td>0.960000</td>\n",
       "      <td>00:25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.234192</td>\n",
       "      <td>1.159229</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.972000</td>\n",
       "      <td>00:25</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.932611</td>\n",
       "      <td>1.826601</td>\n",
       "      <td>0.416000</td>\n",
       "      <td>0.892000</td>\n",
       "      <td>00:25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.713997</td>\n",
       "      <td>1.955363</td>\n",
       "      <td>0.374000</td>\n",
       "      <td>0.866000</td>\n",
       "      <td>00:25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.561075</td>\n",
       "      <td>1.433242</td>\n",
       "      <td>0.622000</td>\n",
       "      <td>0.934000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.421720</td>\n",
       "      <td>1.460905</td>\n",
       "      <td>0.622000</td>\n",
       "      <td>0.938000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.229185</td>\n",
       "      <td>1.180034</td>\n",
       "      <td>0.748000</td>\n",
       "      <td>0.960000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.977032</td>\n",
       "      <td>1.913467</td>\n",
       "      <td>0.374000</td>\n",
       "      <td>0.836000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.729828</td>\n",
       "      <td>1.893670</td>\n",
       "      <td>0.402000</td>\n",
       "      <td>0.884000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.567633</td>\n",
       "      <td>1.452216</td>\n",
       "      <td>0.578000</td>\n",
       "      <td>0.944000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.425000</td>\n",
       "      <td>1.361765</td>\n",
       "      <td>0.626000</td>\n",
       "      <td>0.942000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.229765</td>\n",
       "      <td>1.179944</td>\n",
       "      <td>0.732000</td>\n",
       "      <td>0.956000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.989221</td>\n",
       "      <td>2.063829</td>\n",
       "      <td>0.330000</td>\n",
       "      <td>0.832000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.737373</td>\n",
       "      <td>1.906402</td>\n",
       "      <td>0.352000</td>\n",
       "      <td>0.868000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.587891</td>\n",
       "      <td>1.449219</td>\n",
       "      <td>0.558000</td>\n",
       "      <td>0.938000</td>\n",
       "      <td>00:25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.433785</td>\n",
       "      <td>1.323854</td>\n",
       "      <td>0.654000</td>\n",
       "      <td>0.956000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.233859</td>\n",
       "      <td>1.180467</td>\n",
       "      <td>0.738000</td>\n",
       "      <td>0.966000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.734 0.75  0.748 0.732 0.738]\n",
      "0.74039996\n",
      "0.0073102703\n"
     ]
    }
   ],
   "source": [
    "#Run this to reproduce the results of our entry:\n",
    "%run train.py --run 5 --woof 1 --size 128 --bs 64 --mixup 0 --epoch 5 --lr 4e-3 --gpu 0 \\\n",
    "                --opt ranger --mom .95 --sched_type flat_and_anneal --ann_start 0.72"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Sample run below..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.984163</td>\n",
       "      <td>2.029806</td>\n",
       "      <td>0.330000</td>\n",
       "      <td>0.842000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.740235</td>\n",
       "      <td>1.942208</td>\n",
       "      <td>0.410000</td>\n",
       "      <td>0.868000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.623572</td>\n",
       "      <td>1.513153</td>\n",
       "      <td>0.548000</td>\n",
       "      <td>0.932000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.471418</td>\n",
       "      <td>1.411775</td>\n",
       "      <td>0.608000</td>\n",
       "      <td>0.938000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.269109</td>\n",
       "      <td>1.224273</td>\n",
       "      <td>0.708000</td>\n",
       "      <td>0.958000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.980169</td>\n",
       "      <td>1.957445</td>\n",
       "      <td>0.332000</td>\n",
       "      <td>0.834000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.720152</td>\n",
       "      <td>1.771204</td>\n",
       "      <td>0.428000</td>\n",
       "      <td>0.898000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.544014</td>\n",
       "      <td>1.408681</td>\n",
       "      <td>0.614000</td>\n",
       "      <td>0.952000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.414777</td>\n",
       "      <td>1.338698</td>\n",
       "      <td>0.652000</td>\n",
       "      <td>0.958000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.225740</td>\n",
       "      <td>1.153836</td>\n",
       "      <td>0.740000</td>\n",
       "      <td>0.968000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.940885</td>\n",
       "      <td>1.926415</td>\n",
       "      <td>0.360000</td>\n",
       "      <td>0.840000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.710998</td>\n",
       "      <td>1.874004</td>\n",
       "      <td>0.450000</td>\n",
       "      <td>0.874000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.575062</td>\n",
       "      <td>1.461225</td>\n",
       "      <td>0.598000</td>\n",
       "      <td>0.930000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.406095</td>\n",
       "      <td>1.314005</td>\n",
       "      <td>0.650000</td>\n",
       "      <td>0.968000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.221502</td>\n",
       "      <td>1.167559</td>\n",
       "      <td>0.738000</td>\n",
       "      <td>0.966000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.936900</td>\n",
       "      <td>1.887296</td>\n",
       "      <td>0.378000</td>\n",
       "      <td>0.858000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.706574</td>\n",
       "      <td>1.838660</td>\n",
       "      <td>0.412000</td>\n",
       "      <td>0.866000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.538803</td>\n",
       "      <td>1.465649</td>\n",
       "      <td>0.570000</td>\n",
       "      <td>0.938000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.417734</td>\n",
       "      <td>1.516062</td>\n",
       "      <td>0.610000</td>\n",
       "      <td>0.938000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.219298</td>\n",
       "      <td>1.146909</td>\n",
       "      <td>0.756000</td>\n",
       "      <td>0.968000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lr: 0.004; eff_lr: 0.004; size: 128; alpha: 0.99; mom: 0.95; eps: 1e-06\n",
      "/home/ubuntu/.fastai/data/imagewoof\n",
      "698  annealing start\n",
      "Using Ranger Optimizer...\n",
      "Using Ranger Optimizer...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>top_k_accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.984180</td>\n",
       "      <td>1.871148</td>\n",
       "      <td>0.410000</td>\n",
       "      <td>0.838000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.729023</td>\n",
       "      <td>1.808504</td>\n",
       "      <td>0.416000</td>\n",
       "      <td>0.906000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.579883</td>\n",
       "      <td>1.466045</td>\n",
       "      <td>0.576000</td>\n",
       "      <td>0.934000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.432573</td>\n",
       "      <td>1.329645</td>\n",
       "      <td>0.650000</td>\n",
       "      <td>0.956000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.258308</td>\n",
       "      <td>1.194339</td>\n",
       "      <td>0.734000</td>\n",
       "      <td>0.960000</td>\n",
       "      <td>00:24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.708 0.74  0.738 0.756 0.734]\n",
      "0.7352\n",
      "0.01552288\n"
     ]
    }
   ],
   "source": [
    "%run train.py --run 5 --woof 1 --size 128 --bs 64 --mixup 0 --epoch 5 --lr 4e-3 --gpu 0 \\\n",
    "                --opt ranger --mom .95 --sched_type flat_and_anneal --ann_start 0.72"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#possible ideas to test.....\n",
    "%run train.py --run 5 --woof 1 --size 128 --bs 64 --mixup 0 --epoch 20 --lr 3e-3 --gpu 0 \\\n",
    "                --opt ranger --sched_type flat_and_anneal --ann_start 0.7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run train.py --run 5 --woof 1 --size 128 --bs 64 --mixup 0 --epoch 20 --lr 5e-3 --gpu 0 \\\n",
    "                --opt ranger --sched_type flat_and_anneal --ann_start 0.7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run train.py --run 5 --woof 1 --size 128 --bs 64 --mixup 0 --epoch 20 --lr 8e-3 --gpu 0 \\\n",
    "                --opt ranger --sched_type flat_and_anneal --ann_start 0.7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run train.py --run 2 --woof 1 --size 128 --bs 64 --mixup 0 --epoch 5 --lr 2e-3 --gpu 0 \\\n",
    "                --opt adam --mom .95 --alpha .98 --sched_type flat_and_anneal --ann_start 0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run train.py --run 2 --woof 1 --size 128 --bs 64 --mixup 0 --sa 0 --epoch 80 --lr 8e-2 --gpu 0 \\\n",
    "                --opt Adams --mom .95 --alpha .98 --sched_type flat_and_anneal --ann_start 0.3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pytorch_fastai",
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}