Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature/improve visual notebook #53

Open
wants to merge 11 commits into
base: master
Choose a base branch
from
Open

Feature/improve visual notebook #53

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
9258182
reqs and py file
MarkRoss-Eviden Mar 23, 2024
8ba8b9d
saved notebook version
MarkRoss-Eviden Mar 23, 2024
dd31164
update path and make dir
MarkRoss-Eviden Mar 23, 2024
5c01ad7
updated notbook
MarkRoss-Eviden Mar 23, 2024
a55cfb1
put tensor flow and cv2 in notebook and not reqs
MarkRoss-Eviden Mar 24, 2024
468420e
updated to make s3 optional
MarkRoss-Eviden Apr 2, 2024
bdd57f5
update kernel
MarkRoss-Eviden Apr 2, 2024
8f9cb9e
update kernel
MarkRoss-Eviden Apr 2, 2024
c4dc040
update kernel
MarkRoss-Eviden Apr 2, 2024
5a4f170
update ipynb metadata
MarkRoss-Eviden Apr 2, 2024
781527d
fixed s3 condition
MarkRoss-Eviden Apr 2, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
136 changes: 103 additions & 33 deletions Visual_Analysis.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,15 +40,17 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"# Make sure you have the required pre-reqs\n",
"\n",
"# import sys\n",
"\n",
"# !{sys.executable} -m pip install --upgrade -r requirements.txt"
"# !{sys.executable} -m pip install --upgrade -r requirements.txt\n",
"# !{sys.executable} -m pip install opencv-python\n",
"# !{sys.executable} -m pip install tensorflow"
]
},
{
Expand All @@ -63,32 +65,11 @@
},
{
"cell_type": "code",
"execution_count": 1,
"execution_count": null,
"metadata": {
"tags": []
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"2024-01-06 13:16:09.470934: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.\n",
"2024-01-06 13:16:09.515298: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.\n",
"2024-01-06 13:16:09.516233: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
"To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
"2024-01-06 13:16:10.292374: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/tensorflow/python/compat/v2_compat.py:107: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"non-resource variables are not supported in the long term\n"
]
}
],
"outputs": [],
"source": [
"import json\n",
"import os\n",
Expand All @@ -97,34 +78,123 @@
"import numpy as np\n",
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"from pathlib import Path\n",
"\n",
"import cv2\n",
"\n",
"import tensorflow.compat.v1 as tf\n",
"tf.disable_v2_behavior()\n",
"from tensorflow.compat.v1.io.gfile import GFile\n",
"\n",
"from deepracer.model import load_session, visualize_gradcam_discrete_ppo, rgb2gray"
"from deepracer.model import load_session, visualize_gradcam_discrete_ppo, rgb2gray\n",
"\n",
"import boto3\n",
"s3_resource = boto3.resource('s3')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Configure and load files\n",
"\n",
"Provide the paths where the image and models are stored. Also define which iterations you would like to review."
"# Use example files to understand how the notebook works\n",
"Only run this cell to use the example files"
]
},
{
"cell_type": "code",
"execution_count": 2,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# # Example / Alternative for logs on file-system\n",
"img_selection = 'logs/sample-model/pictures/*.png'\n",
MarkRoss-Eviden marked this conversation as resolved.
Show resolved Hide resolved
"model_path = 'logs/sample-model/model'\n",
"iterations = [15, 30, 48]"
"iterations = [15, 30, 48]\n",
"model_type = 'example'"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Advanced - Fetch your own models from S3 / Minio"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Login\n",
"\n",
"Login to AWS. There are several ways to log in:\n",
"1. On EC2 instance or Sagemaker Notebook with correct IAM execution role assigned.\n",
"2. AWS credentials available in `.aws/` through using the `aws configure` command. (DeepRacer-for-Cloud's `dr-start-loganalysis` supports this)\n",
"3. Setting the relevant environment variables by uncommenting the below section."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# os.environ[\"AWS_DEFAULT_REGION\"] = \"\" #<-Add your region\n",
"# os.environ[\"AWS_ACCESS_KEY_ID\"] = \"\" #<-Add your access key\n",
"# os.environ[\"AWS_SECRET_ACCESS_KEY\"] = \"\" #<-Add you secret access key\n",
"# os.environ[\"AWS_SESSION_TOKEN\"] = \"\" #<-Add your session key if you have one"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Configure S3 to get the models\n",
"\n",
"Depending on which way you are training your model, you will need a slightly different way to load the data.\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"tags": [
"parameters"
]
},
"outputs": [],
"source": [
"PREFIX='model-name' # Name of the model, without trailing '/'\n",
"BUCKET='bucket' # Bucket name is default 'bucket' when training locally\n",
"PROFILE=None # The credentials profile in .aws - 'minio' for local training\n",
"S3_ENDPOINT_URL=None # Endpoint URL: None for AWS S3, 'http://minio:9000' for local training\n",
"iterations = [1, 2, 3] #enter the numbers of your iterations you want to try (must exist in the model folder in S3)\n",
"img_selection = 'logs/sample-model/pictures/*.png' # replace with your own images as appropriate\n",
"model_type = 's3'"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# ### Configure and load files\n",
"#\n",
"if model_type=='s3':\n",
" model_path = 'logs/' + PREFIX\n",
" Path(model_path).mkdir(parents=True, exist_ok=True)\n",
" s3_resource.Object(BUCKET, PREFIX + '/model/model_metadata.json').download_file(\n",
" f'logs/{PREFIX}/model_metadata.json')\n",
" for i in iterations:\n",
" s3_resource.Object(BUCKET, PREFIX + '/model/model_' + str(i) + '.pb').download_file(\n",
" f'logs/{PREFIX}/' + 'model_' + str(i) + '.pb')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Load the models and pictures"
]
},
{
Expand Down Expand Up @@ -394,7 +464,7 @@
],
"source": [
"heatmaps = []\n",
"view_models = models_file_path[1:3]\n",
"view_models = models_file_path[0:len(iterations)]\n",
"\n",
"for model_file in view_models:\n",
" model, obs, model_out = load_session(model_file, my_sensor, False)\n",
Expand Down Expand Up @@ -468,7 +538,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.10"
"version": "3.9.16"
}
},
"nbformat": 4,
Expand Down
59 changes: 55 additions & 4 deletions Visual_Analysis.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,6 +47,8 @@
# import sys

# # !{sys.executable} -m pip install --upgrade -r requirements.txt
# # !{sys.executable} -m pip install opencv-python
# # !{sys.executable} -m pip install tensorflow
# -

#
Expand All @@ -62,6 +64,7 @@
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from pathlib import Path

import cv2

Expand All @@ -70,15 +73,63 @@
from tensorflow.compat.v1.io.gfile import GFile

from deepracer.model import load_session, visualize_gradcam_discrete_ppo, rgb2gray

import boto3
s3_resource = boto3.resource('s3')
# -

# ## Configure and load files
#
# Provide the paths where the image and models are stored. Also define which iterations you would like to review.
# # Use example files to understand how the notebook works
# Only run this cell to use the example files

# # Example / Alternative for logs on file-system
img_selection = 'logs/sample-model/pictures/*.png'
model_path = 'logs/sample-model/model'
iterations = [15, 30, 48]
model_type = 'example'

# # Advanced - Fetch your own models from S3 / Minio

# ### Login
#
# Login to AWS. There are several ways to log in:
# 1. On EC2 instance or Sagemaker Notebook with correct IAM execution role assigned.
# 2. AWS credentials available in `.aws/` through using the `aws configure` command. (DeepRacer-for-Cloud's `dr-start-loganalysis` supports this)
# 3. Setting the relevant environment variables by uncommenting the below section.

# +
# os.environ["AWS_DEFAULT_REGION"] = "" #<-Add your region
# os.environ["AWS_ACCESS_KEY_ID"] = "" #<-Add your access key
# os.environ["AWS_SECRET_ACCESS_KEY"] = "" #<-Add you secret access key
# os.environ["AWS_SESSION_TOKEN"] = "" #<-Add your session key if you have one
# -

# ### Configure S3 to get the models
#
# Depending on which way you are training your model, you will need a slightly different way to load the data.
#

# + tags=["parameters"]
PREFIX='model-name' # Name of the model, without trailing '/'
BUCKET='bucket' # Bucket name is default 'bucket' when training locally
PROFILE=None # The credentials profile in .aws - 'minio' for local training
S3_ENDPOINT_URL=None # Endpoint URL: None for AWS S3, 'http://minio:9000' for local training
iterations = [1, 2, 3] #enter the numbers of your iterations you want to try (must exist in the model folder in S3)
img_selection = 'logs/sample-model/pictures/*.png' # replace with your own images as appropriate
model_type = 's3'
# -

# ### Configure and load files
#
if model_type=='s3':
model_path = 'logs/' + PREFIX
Path(model_path).mkdir(parents=True, exist_ok=True)
s3_resource.Object(BUCKET, PREFIX + '/model/model_metadata.json').download_file(
f'logs/{PREFIX}/model_metadata.json')
for i in iterations:
s3_resource.Object(BUCKET, PREFIX + '/model/model_' + str(i) + '.pb').download_file(
f'logs/{PREFIX}/' + 'model_' + str(i) + '.pb')

# # Load the models and pictures

# Load the model metadata in, and define which sensor is in use.

Expand Down Expand Up @@ -154,7 +205,7 @@

# +
heatmaps = []
view_models = models_file_path[1:3]
view_models = models_file_path[0:len(iterations)]

for model_file in view_models:
model, obs, model_out = load_session(model_file, my_sensor, False)
Expand Down