Edge Impulse Learn API
The Learn API from Edge Impulse — 21 operation(s) for learn.
Operations 26
Each operation below carries the questions people ask an LLM about it and the instructions they give an agent to run it. Generated by API Evangelist overlay
GET
/api/{projectId}/training/{learnId}/x
Download processed features for a learning block · Download data
#
Ask an LLM
“How do I download the processed feature data a learning block trains on?”
“Can I get the DSP output used as training input for a learn block?”
Tell an agent
Download the processed X data for learn block {learnId} in project {projectId}.
Export the training features of learning block {learnId} in project {projectId}.
Ask an LLM
“Where can I download the training labels for a learning block?”
“Do all learn blocks support exporting their Y labels?”
Tell an agent
Download the labels (Y data) for learn block {learnId} in project {projectId}.
Get the label array for learning block {learnId} of project {projectId}.
GET
/api/{projectId}/training/anomaly/{learnId}
Get an anomaly block's configuration · Anomaly information
#
Ask an LLM
“What does my anomaly detection block depend on and how is it configured?”
“How can I inspect the settings of an anomaly block in my impulse?”
Tell an agent
Show the anomaly block {learnId} information for project {projectId}.
Get dependencies and settings of anomaly block {learnId} in project {projectId}.
POST
/api/{projectId}/training/anomaly/{learnId}
Set an anomaly block's minimum confidence · Anomaly settings
#
Ask an LLM
“How do I change the anomaly threshold on my anomaly detection block?”
“Can I set a minimum confidence score for anomaly detection?”
Tell an agent
Set the minimum confidence of anomaly block {learnId} in project {projectId} to {minimumConfidenceRating}.
Update anomaly block {learnId} settings in project {projectId}.
GET
/api/{projectId}/training/anomaly/{learnId}/metadata
Get metadata for a trained anomaly block · Anomaly metadata
#
Ask an LLM
“What metadata is available after my anomaly block has been trained?”
“Where can I see the trained results of an anomaly detection block?”
Tell an agent
Get metadata for trained anomaly block {learnId} in project {projectId}.
Show the training metadata of anomaly block {learnId} for project {projectId}.
GET
/api/{projectId}/training/anomaly/{learnId}/gmm/metadata
Get the GMM model metadata of an anomaly block · Anomaly GMM metadata
#
Ask an LLM
“Can I get the raw Gaussian mixture model parameters from my anomaly block?”
“Where are the GMM component details for a trained anomaly detector?”
Tell an agent
Get the Gaussian mixture model metadata for anomaly block {learnId} in project {projectId}.
Show raw GMM metadata for anomaly block {learnId} of project {projectId}.
GET
/api/{projectId}/training/keras/{learnId}
Get a Keras block's configuration · Keras information
#
Ask an LLM
“How is my Keras neural network block currently configured?”
“What does a Keras learning block depend on in my impulse?”
Tell an agent
Show the Keras block {learnId} information for project {projectId}.
Get the settings and dependencies of Keras block {learnId} in project {projectId}.
Ask an LLM
“How do I change the number of training cycles and learning rate for my neural network?”
“Can I set batch size and train/test split on a Keras block?”
Tell an agent
Set Keras block {learnId} in project {projectId} to {trainingCycles} training cycles at learning rate {learningRate}.
Change the batch size of Keras block {learnId} in project {projectId} to {batchSize}.
GET
/api/{projectId}/training/keras/{learnId}/metadata
Get metadata for a trained Keras model · Keras metadata
#
Ask an LLM
“What accuracy and model details did my trained Keras block produce?”
“Can I fetch trained neural network metadata without the label list?”
Tell an agent
Get metadata for trained Keras block {learnId} in project {projectId}.
Show the trained Keras block {learnId} metadata for project {projectId} excluding labels.
GET
/api/{projectId}/training/keras/{learnId}/data-explorer/features
Get data explorer embeddings from a Keras block · Get data explorer features
#
Ask an LLM
“How can I visualize my dataset using embeddings from my trained neural network?”
“Where do I get the t-SNE layout for the data explorer?”
Tell an agent
Get the data explorer features for Keras block {learnId} in project {projectId}.
Compute the t-SNE view of my raw dataset using embeddings from Keras block {learnId} in project {projectId}.
POST
/api/{projectId}/training/keras/{learnId}/files
Replace a Keras block's files from a zip · Upload Keras files
#
Ask an LLM
“Can I overwrite all of a Keras block's files with a zip archive?”
“What happens to existing Keras block files when I upload a replacement zip?”
Tell an agent destructive · confirm first
Replace the files of Keras block {learnId} in project {projectId} with zip {zip}.
Overwrite Keras block {learnId} in project {projectId} using the contents of {zip}.
POST
/api/{projectId}/training/keras/{learnId}/addFiles
Add files to a Keras block from a zip · Add Keras files
#
Ask an LLM
“How do I add extra files to a Keras block without replacing what is there?”
“Can I append the contents of a zip to my Keras block's files?”
Tell an agent
Add the files in zip {zip} to Keras block {learnId} in project {projectId}.
Merge {zip} into the existing files of Keras block {learnId} for project {projectId}.
GET
/api/{projectId}/training/keras/{learnId}/download-export
Download an exported Keras block · Download Keras export
#
Ask an LLM
“Where do I download a Keras block after exporting it?”
“Can I fetch the exported Keras block code once the export finishes?”
Tell an agent
Download the exported Keras block {learnId} from project {projectId}.
Get the Keras block export archive for block {learnId} in project {projectId}.
GET
/api/{projectId}/training/keras/{learnId}/download-data
Download the exported data of a Keras block · Download Keras data export
#
Ask an LLM
“How do I download the training data that was exported with a Keras block?”
“Where is the dataset export for a Keras block once it is ready?”
Tell an agent
Download the exported data for Keras block {learnId} in project {projectId}.
Fetch the Keras block {learnId} data export from project {projectId}.
GET
/api/{projectId}/learn-data/{learnId}/model/{modelDownloadId}
Download a trained model file · Download trained model
#
Ask an LLM
“How do I download the trained TensorFlow model from a learning block?”
“Can I get the cluster centroids from a trained block?”
Tell an agent
Download model {modelDownloadId} for learn block {learnId} in project {projectId}.
Get the trained model file {modelDownloadId} from learning block {learnId} of project {projectId}.
GET
/api/{projectId}/training/anomaly/{learnId}/features/get-graph
Plot a sample of anomaly-trained features · Trained features
#
Ask an LLM
“Can I graph the features my anomaly block was trained on along two axes?”
“Which feature axes can I choose when charting trained anomaly features?”
Tell an agent
Get trained anomaly features for block {learnId} in project {projectId} plotted on axes {featureAx1} and {featureAx2}.
Graph anomaly block {learnId} training features in project {projectId} using feature {featureAx1} vs {featureAx2}.
GET
/api/{projectId}/training/anomaly/{learnId}/features/get-graph/classification/{sampleId}
Score one sample against an anomaly block · Trained features for sample
#
Ask an LLM
“How do I see the anomaly features and score for a single sample?”
“Can I run the DSP and anomaly classifier on just one sample?”
Tell an agent
Get trained anomaly features for sample {sampleId} using block {learnId} in project {projectId}.
Classify sample {sampleId} with anomaly block {learnId} in project {projectId}.
GET
/api/{projectId}/pretrained-model
Get info on an uploaded pretrained model · Get pretrained model
#
Ask an LLM
“What input and output tensors does my uploaded pretrained model have?”
“Can I check the pretrained model I uploaded earlier?”
Tell an agent
Show the uploaded pretrained model info for project {projectId}.
Get the input/output tensor details of the pretrained model in impulse {impulseId} of project {projectId}.
POST
/api/{projectId}/pretrained-model/upload
Upload a pretrained model · Upload a pretrained model
#
Ask an LLM
“How do I bring my own TFLite or ONNX model into Edge Impulse?”
“Can I include representative features when uploading a pretrained model?”
Tell an agent
Upload pretrained model {modelFileName} of type {modelFileType} to project {projectId}.
Upload {modelFile} as {modelFileName} ({modelFileType}) to project {projectId} and profile it on device {device}.
POST
/api/{projectId}/pretrained-model/save
Save input and model settings for a pretrained model · Save parameters for pretrained model
#
Ask an LLM
“How do I tell Edge Impulse what input format and output type my uploaded model uses?”
“Will saving pretrained model parameters overwrite my current impulse?”
Tell an agent
Save pretrained model input {input} and model config {model} for project {projectId}.
Configure the uploaded model in project {projectId} with input settings {input} and output settings {model}.
POST
/api/{projectId}/pretrained-model/test
Test a pretrained model with raw features · Test pretrained model
#
Ask an LLM
“Can I try my uploaded model on a set of raw feature values?”
“How do I check a pretrained model's prediction from a feature array?”
Tell an agent
Run the pretrained model in project {projectId} on features {features}.
Test pretrained model {modelInfo} in project {projectId} with raw features {features}.
POST
/api/{projectId}/pretrained-model/profile
Profile latency and memory of a pretrained model · Profile pretrained model
#
Ask an LLM
“How much RAM, ROM and latency will my uploaded model use on the target device?”
“Does profiling a pretrained model use the project's selected latency device?”
Tell an agent
Profile the pretrained model in project {projectId}.
Estimate latency, RAM and ROM for the pretrained model in impulse {impulseId} of project {projectId}.
GET
/api/{projectId}/pretrained-model/download/{pretrainedModelDownloadType}
Download a pretrained model file · Download pretrained model
#
Ask an LLM
“Can I download the pretrained model file I uploaded back out of my project?”
“Which file types can I download for an uploaded pretrained model?”
Tell an agent
Download the {pretrainedModelDownloadType} file of the pretrained model in project {projectId}.
Get pretrained model download {pretrainedModelDownloadType} for impulse {impulseId} in project {projectId}.
POST
/api/{projectId}/training/keras/{learnId}/profile
Start profiling a Keras block · Start a profile job for a Keras learn block
#
Ask an LLM
“How do I get on-device latency and memory estimates for my trained neural network?”
“When does a Keras profiling job actually run for the selected device?”
Tell an agent
Start a profiling job for Keras block {learnId} in project {projectId}.
Profile neural network block {learnId} of project {projectId} on the selected latency device.
POST
/api/{projectId}/training/anomaly/{learnId}/profile
Start profiling an anomaly block · Start a profile job for an anomaly learn block
#
Ask an LLM
“How do I estimate on-device performance of my anomaly detection block?”
“Can I kick off profiling for an anomaly block that has no device data yet?”
Tell an agent
Start a profiling job for anomaly block {learnId} in project {projectId}.
Profile anomaly detection block {learnId} of project {projectId} on the selected latency device.
POST
/api/{projectId}/pretrained-model/test-image
Test a pretrained model with an image · Test pretrained model using image data
#
Ask an LLM
“Can I test my uploaded image model by sending a picture?”
“What format does the image need to be in to test a pretrained model?”
Tell an agent
Test the pretrained model in project {projectId} on base64 image {imageFileBase64}.
Run image {imageFileBase64} through pretrained model {model} with input settings {input} in project {projectId}.