Skip to content

Commit

Permalink
docs: quickstart (GITBOOK-32)
Browse files Browse the repository at this point in the history
  • Loading branch information
axiomofjoy authored and gitbook-bot committed Mar 16, 2023
1 parent 5c0b990 commit 15df6bd
Show file tree
Hide file tree
Showing 2 changed files with 56 additions and 7 deletions.
14 changes: 7 additions & 7 deletions docs/how-to/install-and-import-phoenix.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,18 +4,18 @@ description: How to fly with Phoenix

# Install and Import Phoenix

Install Phoenix in your Jupyter or Colab environment with
In your Jupyter or Colab environment, run

```
pip install arize-phoenix
```

{% hint style="info" %}
Phoenix is supported on Python ≥3.8, <3.11.
{% endhint %}
to install Phoenix and its dependencies. Once installed, import Phoenix in your notebook with

Once installed, import Phoenix with

```
```python
import phoenix as px
```

{% hint style="info" %}
Phoenix is supported on Python ≥3.8, <3.11.
{% endhint %}
49 changes: 49 additions & 0 deletions docs/quickstart.md
Original file line number Diff line number Diff line change
@@ -1,2 +1,51 @@
---
description: Get started with Phoenix in three easy steps
---

# Quickstart

Follow along in your local Jupyter environment or in the Colab linked below to take your first flight with Phoenix.

TODO: Add Colab link

## Install and Import Phoenix

In your Jupyter or Colab environment, run

```
pip install arize-phoenix
```

to install Phoenix and its dependencies. Once installed, import Phoenix inside of your notebook with

```python
import phoenix as px
```

## Launch Phoenix with Example Datasets

Run the following to download an example dataset and launch the app.

```python
datasets = px.load_datasets("sentiment_classification_language_drift")
session = px.launch_app(datasets.primary, datasets.reference)
session.view()
```

## Find the Root Cause of Your Model's Performance Issue

This example contains inference data on both training and production examples from a model that classifies the sentiment of product reviews as positive, negative, or neutral. You've noticed a dip in your model's performance in production that you want to investigate.

1. Navigate to the "Embeddings" tab and click on "text\_embedding".
2. The graph at the top of the page measures drift over time. Click on a period of high drift and examine the point cloud below.
3. Click on the first cluster on the left. This cluster contains all production data points, meaning that your model is seeing data in production the likes of which it never saw during training.
4. Use the panel on the right to examine the data points in this cluster.

<details>

<summary>Question: What's gone wrong with your model in production?</summary>

**Answer:** Your model was fine-tuned on examples of labeled product reviews in English. In production, however, your model is encountering product reviews in Spanish whose sentiment it cannot correctly predict.

</details>

0 comments on commit 15df6bd

Please sign in to comment.