id stringlengths 9 29 | image imagewidth (px) 1.02k 1.02k | image_path stringlengths 20 40 | topic stringlengths 11 29 | title stringlengths 9 27 | description stringlengths 84 118 | search_queries stringlengths 59 90 | embedding_text stringlengths 201 253 |
|---|---|---|---|---|---|---|---|
linear_regression_best_fit | images/linear_regression_best_fit.png | Linear Regression | Line of Best Fit | A scatter plot showing data points and a straight best-fit line used in linear regression. | linear regression, line of best fit, scatter plot, prediction, supervised learning | Linear Regression. Line of Best Fit. A scatter plot showing data points and a straight best-fit line used in linear regression. Keywords: linear regression, line of best fit, scatter plot, prediction, supervised learning. | |
support_vector_machine_margin | images/support_vector_machine_margin.png | Support Vector Machines | SVM Margin | A classification diagram showing two classes, a decision boundary, and the maximum margin in a support vector machine. | support vector machine, SVM, margin, decision boundary, classification | Support Vector Machines. SVM Margin. A classification diagram showing two classes, a decision boundary, and the maximum margin in a support vector machine. Keywords: support vector machine, SVM, margin, decision boundary, classification. | |
pca_components | images/pca_components.png | Principal Component Analysis | PCA Components | A two-dimensional point cloud with first and second principal component axes showing dimensionality reduction. | PCA, principal component analysis, dimensionality reduction, eigenvectors, components | Principal Component Analysis. PCA Components. A two-dimensional point cloud with first and second principal component axes showing dimensionality reduction. Keywords: PCA, principal component analysis, dimensionality reduction, eigenvectors, components. | |
decision_tree_split | images/decision_tree_split.png | Decision Trees | Decision Tree Split | A simple decision tree showing how features split data into branches and leaf predictions. | decision tree, feature split, branches, leaves, classification tree | Decision Trees. Decision Tree Split. A simple decision tree showing how features split data into branches and leaf predictions. Keywords: decision tree, feature split, branches, leaves, classification tree. | |
neural_network_layers | images/neural_network_layers.png | Neural Networks | Neural Network Layers | A neural network diagram with input neurons, hidden layers, weighted connections, and output neurons. | neural network, input layer, hidden layer, output layer, deep learning | Neural Networks. Neural Network Layers. A neural network diagram with input neurons, hidden layers, weighted connections, and output neurons. Keywords: neural network, input layer, hidden layer, output layer, deep learning. | |
gradient_descent_steps | images/gradient_descent_steps.png | Gradient Descent | Gradient Descent | A loss curve showing gradient descent steps moving downhill toward the minimum loss. | gradient descent, loss function, optimization, learning rate, minimum | Gradient Descent. Gradient Descent. A loss curve showing gradient descent steps moving downhill toward the minimum loss. Keywords: gradient descent, loss function, optimization, learning rate, minimum. | |
confusion_matrix | images/confusion_matrix.png | Model Evaluation | Confusion Matrix | A confusion matrix explaining true positives, false positives, false negatives, and true negatives. | confusion matrix, true positive, false positive, recall, precision | Model Evaluation. Confusion Matrix. A confusion matrix explaining true positives, false positives, false negatives, and true negatives. Keywords: confusion matrix, true positive, false positive, recall, precision. | |
knn_neighbors | images/knn_neighbors.png | K-Nearest Neighbors | KNN Classification | A K-nearest neighbors diagram showing a query point classified by nearby labeled points. | KNN, k nearest neighbors, query point, classification, distance | K-Nearest Neighbors. KNN Classification. A K-nearest neighbors diagram showing a query point classified by nearby labeled points. Keywords: KNN, k nearest neighbors, query point, classification, distance. | |
clustering_kmeans | images/clustering_kmeans.png | K-Means Clustering | K-Means Clusters | A clustering diagram showing three groups of points and their centroids in k-means clustering. | k-means, clustering, centroid, unsupervised learning, clusters | K-Means Clustering. K-Means Clusters. A clustering diagram showing three groups of points and their centroids in k-means clustering. Keywords: k-means, clustering, centroid, unsupervised learning, clusters. | |
dataset_split_workflow | images/dataset_split_workflow.png | Machine Learning Workflow | Train Test Split | A dataset split diagram showing training data used to fit a model and test data used for evaluation. | train test split, model validation, training data, testing data, machine learning workflow | Machine Learning Workflow. Train Test Split. A dataset split diagram showing training data used to fit a model and test data used for evaluation. Keywords: train test split, model validation, training data, testing data, machine learning workflow. | |
overfitting_underfitting | images/overfitting_underfitting.png | Model Generalization | Overfitting vs Underfitting | A comparison of underfitting, good fit, and overfitting curves on the same kind of data. | overfitting, underfitting, good fit, bias variance, model generalization | Model Generalization. Overfitting vs Underfitting. A comparison of underfitting, good fit, and overfitting curves on the same kind of data. Keywords: overfitting, underfitting, good fit, bias variance, model generalization. | |
roc_curve | images/roc_curve.png | Model Evaluation | ROC Curve | An ROC curve showing true positive rate versus false positive rate with an AUC label. | ROC curve, AUC, true positive rate, false positive rate, classifier evaluation | Model Evaluation. ROC Curve. An ROC curve showing true positive rate versus false positive rate with an AUC label. Keywords: ROC curve, AUC, true positive rate, false positive rate, classifier evaluation. | |
naive_bayes | images/naive_bayes.png | Naive Bayes | Naive Bayes | A probability diagram showing how Naive Bayes combines prior and likelihood evidence to classify a sample. | naive bayes, probability, prior, likelihood, posterior, classification | Naive Bayes. Naive Bayes. A probability diagram showing how Naive Bayes combines prior and likelihood evidence to classify a sample. Keywords: naive bayes, probability, prior, likelihood, posterior, classification. | |
cnn_convolution | images/cnn_convolution.png | Convolutional Neural Networks | CNN Convolution | A convolutional neural network diagram showing an image, a sliding filter, a feature map, and classification. | CNN, convolution, filter, feature map, image classification | Convolutional Neural Networks. CNN Convolution. A convolutional neural network diagram showing an image, a sliding filter, a feature map, and classification. Keywords: CNN, convolution, filter, feature map, image classification. | |
attention_mechanism | images/attention_mechanism.png | Transformers | Attention Mechanism | A transformer attention diagram showing query, key, value vectors and weighted attention scores. | attention mechanism, transformer, query key value, self attention, weights | Transformers. Attention Mechanism. A transformer attention diagram showing query, key, value vectors and weighted attention scores. Keywords: attention mechanism, transformer, query key value, self attention, weights. |
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Sample Knowledge Base Dataset
This folder contains a small starter dataset for the audio-to-image retrieval project.
It has 15 educational diagram images and metadata that can be used to build a Pinecone vector index.
Files
data/sample_knowledge_base/
train/
images/
*.png
metadata.csv
README.md
The images are generated educational diagrams for machine learning topics.
What Each Image Record Needs
Each image should have these fields:
| Field | Required | Why it matters |
|---|---|---|
id |
Yes | Unique identifier for the image/vector. Use this as the Pinecone vector id. |
file_name |
Yes | Used by Hugging Face imagefolder datasets to locate the image file. |
image_path |
Yes | Stored in Pinecone metadata so the app can find the matching dataset image. |
topic |
Recommended | Broad concept, such as Linear Regression or Neural Networks. |
title |
Recommended | Human-readable title for the diagram. |
description |
Yes | Main text used to explain the image and help retrieval. |
search_queries |
Recommended | Extra keywords and phrases users may say. |
embedding_text |
Yes | Final text that should be converted into a CLIP embedding and inserted into Pinecone. |
Why image_path Matters
The app retrieves metadata from Pinecone, then uses image_path to find the actual image in the dataset.
Example Pinecone metadata:
{
"image_path": "images/linear_regression_best_fit.png",
"description": "A scatter plot showing data points and a straight best-fit line used in linear regression."
}
The Hugging Face dataset row should have the same image_path.
Recommended Pinecone Vector Design
For each row:
vector id: id
vector values: OpenCLIP embedding of embedding_text
namespace: text_embeddings
metadata:
image_path
description
topic
title
If using OpenCLIP ViT-B-32, your Pinecone index should use:
dimension: 512
metric: cosine
Loading Locally With Hugging Face Datasets
You can test this dataset locally with:
from datasets import load_dataset
dataset = load_dataset(
"imagefolder",
data_dir="data/sample_knowledge_base",
split="train",
)
print(dataset[0])
The dataset should include an image column plus the metadata fields from metadata.csv.
Dataset Topics
The 15 included images are:
- Linear Regression - Line of Best Fit
- Support Vector Machines - SVM Margin
- Principal Component Analysis - PCA Components
- Decision Trees - Decision Tree Split
- Neural Networks - Neural Network Layers
- Gradient Descent - Loss Minimization
- Model Evaluation - Confusion Matrix
- K-Nearest Neighbors - KNN Classification
- K-Means Clustering - Cluster Centroids
- Machine Learning Workflow - Train Test Split
- Model Generalization - Overfitting vs Underfitting
- Model Evaluation - ROC Curve
- Naive Bayes - Prior, Likelihood, Posterior
- Convolutional Neural Networks - CNN Convolution
- Transformers - Attention Mechanism
- Downloads last month
- 45