workbench.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#i am a monitor notebook used to prototype on the fly >.<\n",
    "#i should be replaced with a proper dashboard,srsly.\n",
    "#right now i show some logs of mastering pong with a2c"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run stuff"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting job # 0 in a separate thread.\n"
     ]
    }
   ],
   "source": [
    "%%bash --bg\n",
    "#create several player processes. Each process plays games and saves results\n",
    "#the loop below spawns 10 players. If you are doing this on a laptop, reducing to 2-4 is okay\n",
    "for i in `seq 1 10`; \n",
    "do\n",
    "        THEANO_FLAGS=device=cpu,floatX=float32 python tinyverse atari.py play -b 3 &\n",
    "done  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode finished after 125 timesteps with reward=0.0\n",
      "Episode finished after 124 timesteps with reward=0.0\n",
      "Episode finished after 246 timesteps with reward=3.0\n",
      "Episode finished after 155 timesteps with reward=1.0\n",
      "Episode finished after 172 timesteps with reward=1.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "tinyverse/database.py:150: UserWarning: Params not found under key 'weights' (got None)\n",
      "  warn(str(exc))\n",
      "[2017-01-16 20:31:02,464] Making new env: BreakoutDeterministic-v0\n",
      "[2017-01-16 20:31:02,517] Clearing 6 monitor files from previous run (because force=True was provided)\n",
      "[2017-01-16 20:31:02,530] Starting new video recorder writing to /home/hedgedir/tinyversers/ezheverse/records/openaigym.video.0.22324.video000000.mp4\n",
      "[2017-01-16 20:31:03,633] Starting new video recorder writing to /home/hedgedir/tinyversers/ezheverse/records/openaigym.video.0.22324.video000001.mp4\n",
      "[2017-01-16 20:31:09,093] Finished writing results. You can upload them to the scoreboard via gym.upload('/home/hedgedir/tinyversers/ezheverse/records')\n"
     ]
    }
   ],
   "source": [
    "%%bash\n",
    "#play 5 games, show results\n",
    "python tinyverse atari.py eval -n 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting job # 2 in a separate thread.\n"
     ]
    }
   ],
   "source": [
    "%%bash --bg\n",
    "#create learner process on GPU. batch size 10\n",
    "THEANO_FLAGS=device=gpu,floatX=float32 python tinyverse atari.py train -b 10 &"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### View stuff"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline\n",
    "import theano\n",
    "import theano.tensor as T\n",
    "import pickle\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hedgedir/agentnet/agentnet/agent/recurrence.py:595: UserWarning: Warning: recurrent loop without unroll_scan got nonempty random state updates list. That happened because there is some source of randomness (e.g. dropout) inside recurrent step graph. To compile such graph, one must either call .get_automatic_updates() right after .get_output and pass these updates to a function, or use no_defalt_updates=True when compiling theano.function.\n",
      "  warn(\"Warning: recurrent loop without unroll_scan got nonempty random state updates list. That happened\"\n",
      "WARNING (theano.tensor.blas): We did not found a dynamic library into the library_dir of the library we use for blas. If you use ATLAS, make sure to compile it with dynamics library.\n",
      "[2017-01-17 07:39:42,515] We did not found a dynamic library into the library_dir of the library we use for blas. If you use ATLAS, make sure to compile it with dynamics library.\n"
     ]
    }
   ],
   "source": [
    "import atari\n",
    "from tinyverse.database import Database\n",
    "from agentnet.environment.session_batch import SessionBatchEnvironment\n",
    "\n",
    "\n",
    "db = Database()\n",
    "self=experiment=atari.make_experiment(db)\n",
    "agent = self.agent\n",
    "\n",
    "\n",
    "observation_shape=(1, 64, 64)  # same as env.observation_space.shape\n",
    "sequence_length=25  # how many steps to make before updating weights\n",
    "\n",
    "#make replay environment\n",
    "observations = T.tensor(theano.config.floatX,broadcastable=(False,)*(2+len(observation_shape)),\n",
    "                        name=\"observations[b,t,color,width,height]\")\n",
    "actions = T.imatrix(\"actions[b,t]\")\n",
    "rewards,is_alive = T.matrices(\"rewards[b,t]\",\"is_alive[b,t]\")\n",
    "prev_memory = [l.input_var for l in agent.agent_states.values()]\n",
    "\n",
    "\n",
    "\n",
    "replay = SessionBatchEnvironment(observations,\n",
    "                                 [observation_shape],\n",
    "                                 actions=actions,\n",
    "                                 rewards=rewards,\n",
    "                                 is_alive=is_alive,\n",
    "                                 preceding_agent_memories=prev_memory\n",
    "                                )\n",
    "\n",
    "#replay sessions\n",
    "_, _, h_seq, _, (logits_seq, V_seq) = agent.get_sessions(\n",
    "    replay,\n",
    "    session_length=sequence_length,\n",
    "    experience_replay=True,\n",
    "    unroll_scan=False,#speeds up compilation my 10x, slows down training by 20% (still faster than TF :P )\n",
    ")\n",
    "rng_updates = agent.get_automatic_updates()\n",
    "\n",
    "# compute pi(a|s) and log(pi(a|s)) manually [use logsoftmax]\n",
    "# we can't guarantee that theano optimizes logsoftmax automatically since it's still in dev\n",
    "logits_flat = logits_seq.reshape([-1,logits_seq.shape[-1]])\n",
    "policy_seq = T.nnet.softmax(logits_flat).reshape(logits_seq.shape)\n",
    "logpolicy_seq = T.nnet.logsoftmax(logits_flat).reshape(logits_seq.shape)\n",
    "\n",
    "#####################\n",
    "import lasagne\n",
    "from agentnet.learning import a2c\n",
    "# get policy gradient\n",
    "elwise_actor_loss,elwise_critic_loss = a2c.get_elementwise_objective(policy=logpolicy_seq,\n",
    "                                                                     treat_policy_as_logpolicy=True,\n",
    "                                                                     state_values=V_seq[:,:,0],\n",
    "                                                                     actions=replay.actions[0],\n",
    "                                                                     rewards=replay.rewards,\n",
    "                                                                     is_alive=replay.is_alive,\n",
    "                                                                     gamma_or_gammas=0.99,\n",
    "                                                                     n_steps=None,\n",
    "                                                                     n_steps_advantage=1,\n",
    "                                                                     return_separate=True)\n",
    "\n",
    "\n",
    "reg_entropy = 1e-2*T.mean(T.sum(policy_seq*logpolicy_seq,axis=-1))\n",
    "loss = elwise_actor_loss.mean() + 0.25*elwise_critic_loss.mean() + reg_entropy\n",
    "\n",
    "\n",
    "# Compute weight updates\n",
    "dV,dl = T.grad(-loss,[V_seq,logits_seq])\n",
    "#####################\n",
    "\n",
    "\n",
    "\n",
    "# compile train function\n",
    "inputs = [observations, actions, rewards, is_alive]+prev_memory\n",
    "eval_fun = theano.function(inputs, \n",
    "                           [policy_seq, V_seq,dV,logits_seq,dl],\n",
    "                           updates=rng_updates,on_unused_input='ignore')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "db.load_all_params(agent)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from agentnet.utils.persistence import save,load\n",
    "save(agent.policy,\"atari_a3c_weights.7h_train.pcl\") #save all layers that pi(a|s) and V(s) depend on\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[2017-01-17 07:41:10,388] Making new env: PongDeterministic-v0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['NOOP', 'FIRE', 'RIGHT', 'LEFT', 'RIGHTFIRE', 'LEFTFIRE']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import gym\n",
    "avg_reward_per_session = []\n",
    "gym.make(\"PongDeterministic-v0\").get_action_meanings()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n_sessions: 5409\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAioAAAFkCAYAAAAOihAyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrFj/Besbz2uolbf/AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4ab8aa7898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r\tV\t\tEnt\tLogits\n",
      "+0.0 -0.212 (+0.027) +1.745 | +0.09(+0.08) +0.15(+0.08) +0.56(-0.35) -0.54(+0.04) +0.00(+0.07) +0.11(+0.08)\n",
      "+0.0 -0.332 (+0.263) +1.728 | -0.08(-0.06) +0.03(-0.06) +0.68(+0.30) -0.49(-0.04) +0.21(-0.08) +0.05(-0.06)\n",
      "+0.0 -0.228 (+0.050) +1.763 | +0.35(+0.06) -0.08(+0.04) +0.38(+0.06) +0.02(+0.05) +0.11(+0.05) -0.34(-0.26)\n",
      "+0.0 -0.303 (+0.197) +1.773 | +0.11(+0.01) +0.11(+0.01) +0.15(+0.02) -0.41(+0.01) +0.26(-0.07) -0.04(+0.01)\n",
      "+0.0 -0.328 (+0.242) +1.758 | -0.10(-0.04) +0.01(-0.05) -0.11(+0.27) -0.37(-0.03) +0.09(-0.06) +0.45(-0.08)\n",
      "+0.0 -0.253 (+0.088) +1.695 | -0.25(+0.04) +0.05(+0.05) +0.19(+0.06) -0.37(+0.03) -0.50(+0.03) +0.73(-0.21)\n",
      "+0.0 -0.331 (+0.240) +1.674 | -0.22(+0.00) -0.28(+0.00) -0.02(+0.00) +0.04(+0.01) -0.36(+0.00) +0.88(-0.01)\n",
      "+0.0 -0.333 (+0.239) +1.744 | -0.65(-0.00) -0.15(-0.01) +0.34(-0.02) +0.01(-0.01) +0.09(-0.01) +0.37(+0.04)\n",
      "+0.0 -0.321 (+0.211) +1.683 | -0.63(+0.02) +0.03(-0.14) -0.20(+0.02) -0.32(+0.02) +0.77(+0.05) +0.30(+0.03)\n",
      "+0.0 -0.367 (+0.298) +1.706 | +0.12(+0.01) -0.15(-0.03) -0.63(+0.01) -0.19(+0.01) +0.66(+0.00) -0.20(+0.01)\n",
      "+0.0 -0.380 (+0.320) +1.727 | +0.48(-0.05) +0.13(-0.03) -0.61(-0.01) -0.37(-0.02) +0.37(+0.15) +0.06(-0.03)\n",
      "+0.0 -0.333 (+0.222) +1.715 | +0.03(+0.02) +0.03(+0.02) -0.58(+0.01) -0.06(-0.10) +0.71(+0.03) -0.04(+0.02)\n",
      "+0.0 -0.366 (+0.283) +1.729 | -0.01(+0.00) +0.09(-0.02) -0.20(+0.00) +0.07(+0.00) +0.82(-0.00) +0.17(+0.00)\n",
      "+0.0 -0.374 (+0.295) +1.680 | -0.09(+0.01) +0.06(+0.01) +0.15(+0.01) -0.38(+0.01) +0.93(-0.06) -0.19(+0.01)\n",
      "+0.0 -0.400 (+0.342) +1.628 | -0.29(+0.02) +0.21(+0.02) -0.50(+0.01) -0.46(+0.01) +1.00(-0.09) +0.07(+0.02)\n",
      "+0.0 -0.437 (+0.411) +1.445 | -0.17(+0.00) -0.34(+0.00) -0.10(+0.00) -0.89(+0.00) +1.38(-0.01) -0.29(+0.00)\n",
      "+0.0 -0.439 (+0.411) +1.611 | +0.08(+0.45) -0.21(-0.06) -0.42(-0.05) -0.25(-0.06) +1.03(-0.23) -0.43(-0.05)\n",
      "+0.0 -0.309 (+0.146) +1.749 | +0.31(-0.00) +0.39(-0.01) -0.02(-0.00) -0.61(+0.00) -0.05(+0.01) +0.06(-0.00)\n",
      "+0.0 -0.309 (+0.141) +1.704 | +0.50(+0.10) +0.35(+0.09) -0.57(+0.04) -0.17(+0.05) -0.45(+0.04) +0.52(-0.32)\n",
      "+0.0 -0.418 (+0.355) +1.658 | +0.80(-0.09) +0.57(-0.07) -0.57(-0.02) -0.41(-0.02) -0.38(+0.23) +0.16(-0.04)\n",
      "+0.0 -0.358 (+0.229) +1.686 | +0.42(-0.11) +0.59(+0.34) -0.31(-0.05) -0.69(-0.03) -0.44(-0.04) +0.42(-0.11)\n",
      "+0.0 -0.242 (-0.007) +1.703 | +0.05(+0.02) +0.48(+0.03) -0.32(+0.02) +0.03(-0.12) -0.80(+0.01) +0.55(+0.03)\n",
      "+0.0 -0.280 (+0.064) +1.762 | +0.09(+0.22) +0.18(-0.05) +0.38(-0.07) -0.20(-0.03) -0.22(-0.03) -0.27(-0.03)\n",
      "+0.0 -0.216 (-0.070) +1.657 | -0.09(+0.02) +0.43(-0.11) +0.69(+0.04) -0.98(+0.01) +0.34(+0.03) -0.60(+0.01)\n",
      "+0.0 -0.253 (+0.000) +1.643 | -0.60(+0.00) +0.28(-0.00) +0.86(-0.01) -0.49(+0.00) +0.12(+0.00) -0.46(+0.00)\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-11-865b8321f043>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     33\u001b[0m     \u001b[0mR\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     34\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m5000\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 35\u001b[0;31m         \u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mr\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0md\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mm\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloads\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mredis\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlindex\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'sessions'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     36\u001b[0m         \u001b[0mR\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m     \u001b[0mavg_reward_per_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mR\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/client.py\u001b[0m in \u001b[0;36mlindex\u001b[0;34m(self, name, index)\u001b[0m\n\u001b[1;32m   1203\u001b[0m         \u001b[0mend\u001b[0m \u001b[0mof\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1204\u001b[0m         \"\"\"\n\u001b[0;32m-> 1205\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute_command\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'LINDEX'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1206\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1207\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mlinsert\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwhere\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrefvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/client.py\u001b[0m in \u001b[0;36mexecute_command\u001b[0;34m(self, *args, **options)\u001b[0m\n\u001b[1;32m    571\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    572\u001b[0m             \u001b[0mconnection\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msend_command\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 573\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparse_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconnection\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcommand_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0moptions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    574\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mConnectionError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mTimeoutError\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    575\u001b[0m             \u001b[0mconnection\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdisconnect\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/client.py\u001b[0m in \u001b[0;36mparse_response\u001b[0;34m(self, connection, command_name, **options)\u001b[0m\n\u001b[1;32m    583\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mparse_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconnection\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcommand_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0moptions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    584\u001b[0m         \u001b[0;34m\"Parses a response from the Redis server\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 585\u001b[0;31m         \u001b[0mresponse\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconnection\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mread_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    586\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mcommand_name\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresponse_callbacks\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    587\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresponse_callbacks\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mcommand_name\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresponse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0moptions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/connection.py\u001b[0m in \u001b[0;36mread_response\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    575\u001b[0m         \u001b[0;34m\"Read the response from a previously sent command\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    576\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 577\u001b[0;31m             \u001b[0mresponse\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parser\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mread_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    578\u001b[0m         \u001b[0;32mexcept\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    579\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdisconnect\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/connection.py\u001b[0m in \u001b[0;36mread_response\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    236\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    237\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mread_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 238\u001b[0;31m         \u001b[0mresponse\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_buffer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreadline\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    239\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mresponse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    240\u001b[0m             \u001b[0;32mraise\u001b[0m \u001b[0mConnectionError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mSERVER_CLOSED_CONNECTION_ERROR\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/connection.py\u001b[0m in \u001b[0;36mreadline\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    166\u001b[0m         \u001b[0;32mwhile\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mendswith\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mSYM_CRLF\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    167\u001b[0m             \u001b[0;31m# there's more data in the socket that we need\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 168\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_read_from_socket\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    169\u001b[0m             \u001b[0mbuf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseek\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbytes_read\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    170\u001b[0m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbuf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreadline\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.5/site-packages/redis/connection.py\u001b[0m in \u001b[0;36m_read_from_socket\u001b[0;34m(self, length)\u001b[0m\n\u001b[1;32m    124\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    125\u001b[0m             \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 126\u001b[0;31m                 \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_sock\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrecv\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msocket_read_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    127\u001b[0m                 \u001b[0;31m# an empty string indicates the server shutdown the socket\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    128\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbytes\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "#MEGA-controller of doom. Draws all the stuff i wanted to see for some reason. Don't try to analize\n",
    "from IPython.display import clear_output\n",
    "while True:\n",
    "    if db.redis.exists(\"weights\"):\n",
    "        weights = db.loads(db.redis.get('weights'))\n",
    "        if any(map(np.any,map(np.isnan,weights))):\n",
    "            print(\"Weights broken!!!\")\n",
    "    \n",
    "    print(\"n_sessions:\",db.redis.llen(\"sessions\"))\n",
    "    plt.plot(avg_reward_per_session)\n",
    "    plt.grid()\n",
    "    plt.show()\n",
    "    \n",
    "    \n",
    "    from scipy.stats import entropy\n",
    "    db.load_all_params(experiment.agent,errors='raise')\n",
    "    s,a,r,alive,mem = db.loads(db.redis.lindex('sessions',np.random.randint(1000)))\n",
    "\n",
    "    pi,v,dv,logits,dlogits = eval_fun([s],[a],[r],[alive],*[m[None] for m in mem])\n",
    "    \n",
    "    print (\"r\\tV\\t\\tEnt\\tLogits\")\n",
    "    for lo_i,dlo_i,pi_i,v_i,dv_i,r_i in zip(\n",
    "        logits.reshape(-1,6),\n",
    "        dlogits.reshape(-1,6),\n",
    "        pi.reshape(-1,6),\n",
    "        v.reshape(-1),\n",
    "        dv.reshape(-1),\n",
    "        r.reshape(-1)):\n",
    "        logit_str = map(lambda l,dl:(\"%+.2f(%+.2f)\"%(l,100*dl)),lo_i,dlo_i)\n",
    "        print (\"%+.1f\"%r_i, \"%+.3f\"%v_i,\"(%+.3f)\"%(100*dv_i),\"%+.3f\"%entropy(pi_i),\"|\",\" \".join(logit_str))\n",
    "    \n",
    "\n",
    "    R=0\n",
    "    for i in range(5000)[::-1]:\n",
    "        s,a,r,d,m = db.loads(db.redis.lindex('sessions',i))\n",
    "        R += sum(r)\n",
    "    avg_reward_per_session.append(R)\n",
    "    clear_output(True)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "root      6326 90.6  0.7 79039712 928644 pts/11 Rl+ 06:42   0:44 python tinyverse atari.py train -b 10\r\n",
      "root      6475 90.9  0.2 71643972 347380 pts/8 Rl   06:42   0:28 python tinyverse atari.py play -b 1\r\n",
      "root      6476 87.8  0.2 71643852 348072 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6477 89.0  0.2 71643968 349932 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6478 87.4  0.2 71643552 347084 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6479 87.4  0.2 71643540 354384 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6480 84.1  0.2 71643848 347368 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6481 91.0  0.2 71643512 349676 pts/8 Rl   06:42   0:28 python tinyverse atari.py play -b 1\r\n",
      "root      6482 87.4  0.2 71643512 347188 pts/8 Sl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6483 86.0  0.2 71643848 348648 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6484 86.0  0.2 71643844 351244 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6485 90.8  0.2 71643840 347592 pts/8 Rl   06:42   0:28 python tinyverse atari.py play -b 1\r\n",
      "root      6486 97.8  0.2 71645432 348740 pts/8 Rl   06:42   0:30 python tinyverse atari.py play -b 1\r\n",
      "root      6487 89.5  0.2 71643848 349260 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6488 87.6  0.2 71643540 346468 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6489 85.3  0.2 71643588 349764 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6490 87.1  0.2 71643844 347828 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6491 87.3  0.2 71643388 346572 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6492 86.1  0.2 71643840 347660 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6493 87.7  0.2 71643856 352008 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6494 86.0  0.2 71643520 349620 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6495 91.6  0.2 71643856 347504 pts/8 Rl   06:42   0:28 python tinyverse atari.py play -b 1\r\n",
      "root      6496 88.1  0.2 71643576 347772 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6497 85.1  0.2 71643844 348192 pts/8 Rl   06:42   0:26 python tinyverse atari.py play -b 1\r\n",
      "root      6498 89.2  0.2 71643500 346368 pts/8 Rl   06:42   0:27 python tinyverse atari.py play -b 1\r\n",
      "root      6499 91.4  0.2 71643508 347040 pts/8 Rl   06:42   0:28 python tinyverse atari.py play -b 1\r\n",
      "root     11217  0.0  0.0   4448   796 pts/16   Ss+  06:43   0:00 /bin/sh -c ps aux | grep tinyverse\r\n",
      "root     11219  0.0  0.0  10472  2136 pts/16   R+   06:43   0:00 grep tinyverse\r\n"
     ]
    }
   ],
   "source": [
    "!ps aux | grep tinyverse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "!kill -9 $(ps aux | grep tinyverse | awk '{print $2}') #uncomment to hard kill all\n",
    "from tinyverse.database import Database\n",
    "db = Database()\n",
    "for key in db.redis.keys():\n",
    "    db.redis.delete(key)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "db.redis.delete(\"weights\")"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}