Rename utilities to help.

hallvardnmbu · Mar 12, 2024 · 08302f4 · 08302f4
1 parent af5b2c4
commit 08302f4
Show file tree

Hide file tree

Showing 13 changed files with 55 additions and 132 deletions.
diff --git a/breakout/example.ipynb b/breakout/example.ipynb
@@ -3,21 +3,30 @@
   {
    "cell_type": "markdown",
    "source": [
-    "# Visualise the pre-trained agent in action\n",
-    "\n",
-    "Modify the path to the weights and run the notebook."
+    "# Example of visualising the agent's training history performance"
    ],
    "metadata": {
     "collapsed": false
    },
    "id": "b3d8465ecb86eca7"
   },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "MODEL : Path to the pre-trained model\n",
+    "METRICS : Path to the training history, or None"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "fff872a8189754af"
+  },
   {
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "WEIGHTS = './_output/weights-15000.pth'\n",
-    "METRICS = './_output/metrics.csv'"
+    "MODEL = './results/model.pth'\n",
+    "METRICS = None"
    ],
    "metadata": {
     "collapsed": false
@@ -34,11 +43,9 @@
     "import gymnasium as gym\n",
     "import matplotlib.pyplot as plt\n",
     "\n",
-    "from DQN import VisionDeepQ\n",
-    "\n",
     "sys.path.append(\"../\")\n",
-    "from utilities.visualisation.plot import graph                                              # noqa\n",
-    "from utilities.visualisation.movie import movie                                             # noqa"
+    "from help.visualisation.plot import graph                                              # noqa\n",
+    "from help.visualisation.movie import movie                                             # noqa"
    ],
    "metadata": {
     "collapsed": false
@@ -49,48 +56,7 @@
   {
    "cell_type": "markdown",
    "source": [
-    "## Parameters"
-   ],
-   "metadata": {
-    "collapsed": false
-   },
-   "id": "4dddd56883444fab"
-  },
-  {
-   "cell_type": "code",
-   "outputs": [],
-   "source": [
-    "network = {\n",
-    "    \"input_channels\": 4, \"outputs\": 4,\n",
-    "    \"channels\": [32, 64, 64],\n",
-    "    \"kernels\": [8, 4, 3],\n",
-    "    \"padding\": [\"valid\", \"valid\", \"valid\"],\n",
-    "    \"strides\": [4, 2, 1],\n",
-    "    \"nodes\": [],\n",
-    "}\n",
-    "optimizer = {\n",
-    "    \"optimizer\": torch.optim.Adam,\n",
-    "    \"lr\": 1e-5,\n",
-    "    \"hyperparameters\": {}\n",
-    "}\n",
-    "shape = {\n",
-    "    \"original\": (1, 1, 210, 160),\n",
-    "    \"width\": slice(7, -7),\n",
-    "    \"height\": slice(31, -17),\n",
-    "    \"max_pooling\": 2,\n",
-    "}\n",
-    "skip = 4"
-   ],
-   "metadata": {
-    "collapsed": false
-   },
-   "id": "16867687f37ddbca",
-   "execution_count": null
-  },
-  {
-   "cell_type": "markdown",
-   "source": [
-    "## Setup"
+    "## Loading the agent and environment"
    ],
    "metadata": {
     "collapsed": false
@@ -101,13 +67,7 @@
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "value_agent = VisionDeepQ(\n",
-    "    network=network, optimizer=optimizer, shape=shape,\n",
-    "    exploration_rate=0.002,\n",
-    ")\n",
-    "\n",
-    "weights = torch.load(WEIGHTS, map_location=torch.device('cpu'))\n",
-    "value_agent.load_state_dict(weights)\n",
+    "agent = torch.load(MODEL, map_location=torch.device('cpu'))\n",
     "\n",
     "environment = gym.make('ALE/Breakout-v5', render_mode=\"rgb_array\",\n",
     "                       obs_type=\"grayscale\", frameskip=1, repeat_action_probability=0.0)\n",
@@ -132,7 +92,7 @@
   {
    "cell_type": "markdown",
    "source": [
-    "### Plotting the metrics from the csv-file created during training."
+    "### Training history (if specified)"
    ],
    "metadata": {
     "collapsed": false
@@ -143,7 +103,7 @@
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "graph(METRICS, title=\"Training history\", window=20) if METRICS else None\n",
+    "graph(METRICS, title=\"Breakout training history\", window=20) if METRICS else None\n",
     "plt.show() if METRICS else None"
    ],
    "metadata": {
@@ -155,7 +115,7 @@
   {
    "cell_type": "markdown",
    "source": [
-    "### Creating and saving a gif of the agent in action. The gif will be saved to the given path."
+    "### In action"
    ],
    "metadata": {
     "collapsed": false
@@ -166,7 +126,7 @@
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "movie(environment, value_agent, './_output/breakout.avi', fps=60)"
+    "movie(environment, agent, './results/breakout.mp4', fps=20)"
    ],
    "metadata": {
     "collapsed": false

diff --git a/cart-pole/DQN.ipynb b/cart-pole/DQN.ipynb
@@ -44,8 +44,8 @@
     "from DQN import DeepQ\n",
     "\n",
     "sys.path.append(\"../\")\n",
-    "from utilities.visualisation.plot import plot                                               # noqa\n",
-    "from utilities.visualisation.gif import gif2                                                # noqa"
+    "from help.visualisation.plot import plot                                                    # noqa\n",
+    "from help.visualisation.gif import gif2                                                     # noqa"
    ]
   },
   {

diff --git a/cart-pole/REINFORCE.ipynb b/cart-pole/REINFORCE.ipynb
@@ -43,8 +43,8 @@
     "from REINFORCE import PolicyGradient\n",
     "\n",
     "sys.path.append(\"../\")\n",
-    "from utilities.visualisation.plot import plot                                               # noqa\n",
-    "from utilities.visualisation.gif import gif2                                                # noqa"
+    "from help.visualisation.plot import plot                                                    # noqa\n",
+    "from help.visualisation.gif import gif2                                                     # noqa"
    ]
   },
   {

diff --git a/enduro/example.ipynb b/enduro/example.ipynb
@@ -3,21 +3,30 @@
   {
    "cell_type": "markdown",
    "source": [
-    "# Visualise the pre-trained agent in action\n",
-    "\n",
-    "Modify the path to the weights and run the notebook."
+    "# Example of visualising the agent's training history performance"
    ],
    "metadata": {
     "collapsed": false
    },
    "id": "b3d8465ecb86eca7"
   },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "MODEL : Path to the pre-trained model\n",
+    "METRICS : Path to the training history, or None"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "fbd0af6b11428abe"
+  },
   {
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "WEIGHTS = './_output/weights-0.pth'\n",
-    "METRICS = None #'./_output/metrics.csv'"
+    "MODEL = './results/model.pth'\n",
+    "METRICS = None"
    ],
    "metadata": {
     "collapsed": false
@@ -34,11 +43,11 @@
     "import gymnasium as gym\n",
     "import matplotlib.pyplot as plt\n",
     "\n",
-    "from DQN import VisionDeepQ\n",
+    "from train import SKIP\n",
     "\n",
     "sys.path.append(\"../\")\n",
-    "from utilities.visualisation.plot import visualise_csv                                       # noqa\n",
-    "from utilities.visualisation.gif import gif                                                  # noqa"
+    "from help.visualisation.plot import graph2                                                  # noqa\n",
+    "from help.visualisation.gif import gif                                                      # noqa"
    ],
    "metadata": {
     "collapsed": false
@@ -49,47 +58,7 @@
   {
    "cell_type": "markdown",
    "source": [
-    "## Parameters"
-   ],
-   "metadata": {
-    "collapsed": false
-   },
-   "id": "4dddd56883444fab"
-  },
-  {
-   "cell_type": "code",
-   "outputs": [],
-   "source": [
-    "network = {\n",
-    "    \"input_channels\": 2, \"outputs\": 9,\n",
-    "    \"channels\": [32, 64, 64],\n",
-    "    \"kernels\": [8, 4, 3],\n",
-    "    \"padding\": [\"valid\", \"valid\", \"valid\"],\n",
-    "    \"strides\": [4, 2, 1],\n",
-    "    \"nodes\": [512],\n",
-    "}\n",
-    "optimizer = {\n",
-    "    \"optimizer\": torch.optim.RMSprop,\n",
-    "    \"lr\": 0.0001,\n",
-    "    \"hyperparameters\": {}\n",
-    "}\n",
-    "shape = {\n",
-    "    \"original\": (1, 1, 210, 160),\n",
-    "    \"height\": slice(51, 155),\n",
-    "    \"width\": slice(8, 160)\n",
-    "}\n",
-    "skip = 4"
-   ],
-   "metadata": {
-    "collapsed": false
-   },
-   "id": "16867687f37ddbca",
-   "execution_count": null
-  },
-  {
-   "cell_type": "markdown",
-   "source": [
-    "## Setup"
+    "## Loading the agent and environment"
    ],
    "metadata": {
     "collapsed": false
@@ -100,13 +69,7 @@
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "value_agent = VisionDeepQ(\n",
-    "    network=network, optimizer=optimizer, shape=shape,\n",
-    "    exploration_rate=0.01,\n",
-    ")\n",
-    "\n",
-    "weights = torch.load(WEIGHTS, map_location=torch.device('cpu'))\n",
-    "value_agent.load_state_dict(weights)\n",
+    "agent = torch.load(MODEL, map_location=torch.device('cpu'))\n",
     "\n",
     "environment = gym.make('ALE/Enduro-v5', render_mode=\"rgb_array\",\n",
     "                       obs_type=\"grayscale\", frameskip=1, repeat_action_probability=0.0)\n",
@@ -131,7 +94,7 @@
   {
    "cell_type": "markdown",
    "source": [
-    "### Plotting the metrics from the csv-file created during training."
+    "### Training history (if specified)"
    ],
    "metadata": {
     "collapsed": false
@@ -142,7 +105,7 @@
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "visualise_csv(METRICS, title=\"Training history\", window=20) if METRICS else None\n",
+    "graph2(METRICS, title=\"Enduro training history\", window=20) if METRICS else None\n",
     "plt.show() if METRICS else None"
    ],
    "metadata": {
@@ -154,7 +117,7 @@
   {
    "cell_type": "markdown",
    "source": [
-    "### Creating and saving a gif of the agent in action. The gif will be saved to the given path."
+    "### In action"
    ],
    "metadata": {
     "collapsed": false
@@ -165,7 +128,7 @@
    "cell_type": "code",
    "outputs": [],
    "source": [
-    "gif(environment, value_agent, './_output/enduro-0.gif', skip, 25)"
+    "gif(environment, agent, './results/enduro.gif', SKIP, 25)"
    ],
    "metadata": {
     "collapsed": false

diff --git a/utilities/orion-hpc/README.md → help/orion-hpc/README.md b/utilities/orion-hpc/README.md → help/orion-hpc/README.md
diff --git a/...ies/orion-hpc/singularity/singularity.def → help/orion-hpc/singularity/singularity.def b/...ies/orion-hpc/singularity/singularity.def → help/orion-hpc/singularity/singularity.def
diff --git a/...ties/orion-hpc/singularity/singularity.sh → help/orion-hpc/singularity/singularity.sh b/...ties/orion-hpc/singularity/singularity.sh → help/orion-hpc/singularity/singularity.sh
diff --git a/utilities/orion-hpc/train.sh → help/orion-hpc/train.sh b/utilities/orion-hpc/train.sh → help/orion-hpc/train.sh
diff --git a/utilities/visualisation/gif.py → help/visualisation/gif.py b/utilities/visualisation/gif.py → help/visualisation/gif.py
@@ -31,7 +31,7 @@ def gif2(environment, agent, path="./live-preview.gif", duration=50):
     _ = imageio.mimsave(path, images, duration=duration)
 
 
-def gif(environment, agent, path="./live-preview.gif", skip=4, duration=50):
+def gif(environment, agent, path="./live-preview.gif", skip=1, duration=50):
     """
     Create a GIF of the agent playing the environment.
 

diff --git a/utilities/visualisation/movie.py → help/visualisation/movie.py b/utilities/visualisation/movie.py → help/visualisation/movie.py
@@ -4,7 +4,7 @@
 import torch
 
 
-def movie(environment, agent, path="./live-preview.mp4", skip=4, fps=50):
+def movie(environment, agent, path="./live-preview.mp4", skip=1, fps=50):
     """Created by Mistral Large."""
     states = agent.preprocess(environment.reset()[0])
     if hasattr(agent, "shape") and "reshape" in agent.shape:

diff --git a/utilities/visualisation/plot.py → help/visualisation/plot.py b/utilities/visualisation/plot.py → help/visualisation/plot.py
diff --git a/tetris/example.ipynb b/tetris/example.ipynb
@@ -41,8 +41,8 @@
     "from DQN import DeepQ\n",
     "\n",
     "sys.path.append(\"../\")\n",
-    "from utilities.visualisation.plot import graph                                              # noqa\n",
-    "from utilities.visualisation.gif import gif                                                 # noqa"
+    "from help.visualisation.plot import graph                                                   # noqa\n",
+    "from help.visualisation.gif import gif                                                      # noqa"
    ],
    "metadata": {
     "collapsed": false,

diff --git a/tetris/transfer-learning/DQN-ResNet18.ipynb b/tetris/transfer-learning/DQN-ResNet18.ipynb
@@ -34,8 +34,8 @@
     "from DQN import TransferDeepQ\n",
     "\n",
     "sys.path.append(\"../\")\n",
-    "from utilities.visualisation.plot import plot                                               # noqa\n",
-    "from utilities.visualisation.gif import gif                                                 # noqa"
+    "from help.visualisation.plot import plot                                               # noqa\n",
+    "from help.visualisation.gif import gif                                                 # noqa"
    ]
   },
   {