# Zero-shot Image Classification with OpenAI's CLIP

> A deep dive on OpenAI's CLIP for zero-shot image classification.

State-of-the-art (SotA) computer vision (CV) models are characterized by a _restricted_ understanding of the visual world based on their training data [1].

These models can perform _very well_ on specific tasks and datasets, but they do not generalize well. They cannot handle new classes or images beyond the domain they have been trained with.

This brittleness can be a problem for building niche image classification use-cases, such as defect detection in agriculture to identifying false banknotes to fight fraud. It can be extraordinarily hard to gather labeled datasets that are large enough to fine-tune CV models with traditional methods for those specialized use cases.

Ideally, a CV model should learn the contents of images without excessive focus on the specific labels it is initially trained to understand. With an image of a dog, the model should understand that a dog is in the image. But it would be a lot more useful if it could also understand that there are trees in the background, that it is daytime, and that the dog is on a grassy field.

Unfortunately, the result of classification training is the opposite. Models learn to push their internal representations of dogs into the same “dog vector space” and cats into the same “cat vector space”. All that matters is the binary yes/no as to whether an image aligns with a class.

![Dog and cat vectors](https://cdn.sanity.io/images/vr8gru94/production/e56c69c5bd52b98509324e0e02ae0344d897547b-1980x1140.png)


Retraining classification models is an option, but training requires significant time and capital investment for gathering a classification dataset _and_ the act of model training itself.

Fortunately, OpenAI’s CLIP has proved itself as an incredibly flexible classification model that often requires _zero_ retraining. In this chapter, we will explore zero-shot image classification using CLIP.

[Video](https://www.youtube.com/watch?v=98POYg2HZqQ)


## N-Shot Learning

Before diving into CLIP, let’s take a moment to understand what exactly “zero-shot” is and its significance in ML.

The concept derives from N-shot learning. Here we define _N_ as the number of samples required to train a model to begin making predictions in a new domain or on a new task.

Many SotA models today are pretrained on vast amounts of data like ResNet or BERT. These pretrained models are then _fine-tuned_ for a specific task and domain. For example, a ResNet model can be pretrained with [ImageNet](https://www.pinecone.io/learn/series/image-search/imagenet/) and then fine-tuned for clothing categorization.

Models like ResNet and BERT are called _“many-shot”_ learners because we need _many_ training samples to reach acceptable performance during that final fine-tuning step.

Many-shot learning is only possible when we have compute, time, and data to allow us to fine-tune our models. Ideally, we want to maximize model performance while minimizing N-shot requirements.

Zero-shot is the natural _best-case scenario_ for a model as it means we require _zero_ training samples before shifting it to a new domain or task.

CLIP may not be breaking SotA performance benchmarks on specific datasets. Still, it is proving to be a massive leap forward in zero-shot performance across various tasks in both image and text modalities.

---

_The point of CLIP is not SotA performance. However, it’s worth noting that CLIP did beat the previous SotA results on the STL10 benchmark despite never being trained on that dataset._

---

The zero-shot adaptability of CLIP was found to work across many domains and different tasks. We will be talking about image classification in this article, but it can also be used in [multi-modal search/recommendation](https://www.pinecone.io/learn/series/image-search/clip/), object detection, and likely many more as of yet unknown tasks.

## How CLIP Makes Zero-Shot So Effective?

**C**ontrastive **L**anguage-**I**mage **P**retraining (CLIP) is a primarily transformer-based model released by OpenAI in 2021 [1].

CLIP consists of two models, as discussed in more depth in the [previous chapter](https://www.pinecone.io/learn/series/image-search/clip/). The version of CLIP we use here consists of a text transformer for encoding text embeddings and a vision transformer (ViT) for encoding image embeddings.

![Encoding image embeddings](https://cdn.sanity.io/images/vr8gru94/production/46b85e640919609e98a79a8142a1181e57d12799-2803x1335.png)


Both CLIP models are optimized during pretraining to align similar text and images in vector space. It does this by taking image-text pairs and pushing their output vectors nearer in vector space while separating the vectors of non-pairs.

![Text and image encoders](https://cdn.sanity.io/images/vr8gru94/production/b7e270212d57ff9dd2da60f8fac00c6dddef9c3a-1165x671.png)


It distinguishes itself from typical classification models for several reasons. First, OpenAI trained it on a **huge dataset of 400M text-image pairs** that were scraped from across the internet.

There are three primary benefits here:

1. CLIP requires just **image-text pairs** rather than specific class labels thanks to the _contrastive_ rather than _classification_ focused training function. This type of data is abundant in today’s social-media-centric world.
2. The large dataset size means CLIP can build a strong understanding of general textual concepts displayed within images.
3. Text descriptors often describe various features of an image, not just one. Meaning a more holistic representation of images (and text) can be built.

These benefits of CLIP are the primary factors that have led to its outstanding zero-shot performance.

The authors of CLIP draw a great example of this by comparing a ResNet-101 model trained specifically on ImageNet to CLIP when both are applied to other datasets derived from ImageNet.

![ResNet-101 fine-tuned on ImageNet vs. zero-shot CLIP, source [1].](https://cdn.sanity.io/images/vr8gru94/production/e58fd51aacd84d301d1cad0bb13dcd5296beb4df-2486x1398.png)


In this comparison, we can see that despite ResNet-101 training for ImageNet, its performance on _similar_ datasets is much worse than CLIP on the same tasks. CLIP outperforms a SotA model trained for ImageNet on slightly modified ImageNet tasks.

When applying a ResNet model to other domains, a standard approach is to use a “linear probe”. That is where the ResNet learned features (from the last few layers) are fed into a linear classifier that is then fine-tuned for a specific dataset. We would regard this as few to many-shot learning.

In the CLIP paper, linear probe ResNet-50 was compared to zero-shot CLIP. In one scenario, zero-shot CLIP outperforms linear probing across many tasks.

![Zero-shot CLIP performance compared to ResNet with linear probe, source [1].](https://cdn.sanity.io/images/vr8gru94/production/86c758f560e227b83e43813377754a54265cdf49-521x554.png)


Despite CLIP not being trained for these specific tasks, it outperforms a ResNet-50 with a linear probe. However, it’s worth noting that zero-shot did not outperform linear probing when given more training samples.

## Zero-Shot Classification

How exactly does CLIP do zero-shot classification? We know that the image and text encoder creates a 512-dimensional image and text vector that map to the same vector space.

Considering this vector space alignment, what if we wrote the dataset classes as text sentences?

Given a task where we must identify whether a photo contains a car, bird, or cat: we could create and encode three text classes:

`"a photo of a car" -> T_1 "a photo of a bird" -> T_2 "a photo of a cat" -> T_3`

Each of these _“classes”_ are output from the text encoder as vectors $T_1$, $T_2$, and $T_3$ respectively. Given an photo of a cat, we encode it with the ViT model to create vector $I_1$. When we calculate the similarity of these vectors with cosine similarity, we expect $sim(T_3,I_1)$ to return the highest score.

![We reformat the one-word classes into sentences and encode them into label-based text vectors. When we encode and compare an image of a cat, we expect to share the highest similarity with *"a photo of a cat"*.](https://cdn.sanity.io/images/vr8gru94/production/d9a6ebbc9a2f3334ec57a6b54d90155043c07595-1292x447.png)


We format the one-word classes into sentences because we expect CLIP saw more _sentence-like_ text during pretraining. For ImageNet it was reported that a 1.3 percentage point improvement in accuracy was achieved using the same prompt template of `"a photo of a {label}"` [1].

Prompt templates don’t always improve performance and they should be tested for each dataset. For the dataset we use below, we found it to _reduce_ performance by 1.5 percentage points.

### Python Implementation

Let’s move on to an applied example of CLIP for zero-shot classification. We will use the `frgfm/imagenette` dataset via Hugging Face _Datasets_.

```json
{
  "_key": "e2af9a7c45f7",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 2,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"Dataset({\\n\",\n       \"    features: ['image', 'label'],\\n\",\n       \"    num_rows: 3925\\n\",\n       \"})\"\n      ]\n     },\n     \"execution_count\": 2,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# import the imagenette dataset\\n\",\n    \"from datasets import load_dataset\\n\",\n    \"\\n\",\n    \"imagenette = load_dataset(\\n\",\n    \"    'frgfm/imagenette',\\n\",\n    \"    '320px',\\n\",\n    \"    split='validation',\\n\",\n    \"    revision=\\\"4d512db\\\"\\n\",\n    \")\\n\",\n    \"# show dataset info\\n\",\n    \"imagenette\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 3,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"{0, 1, 2, 3, 4, 5, 6, 7, 8, 9}\"\n      ]\n     },\n     \"execution_count\": 3,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# check labels in the dataset\\n\",\n    \"set(imagenette['label'])\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

The dataset contains _10_ labels, all stored as integer values. To perform classification with CLIP we need the _text content_ of these labels. Most Hugging Face datasets include the mapping to text labels inside the the dataset `info`:

```json
{
  "_key": "3e355923d37c",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 4,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"['tench',\\n\",\n       \" 'English springer',\\n\",\n       \" 'cassette player',\\n\",\n       \" 'chain saw',\\n\",\n       \" 'church',\\n\",\n       \" 'French horn',\\n\",\n       \" 'garbage truck',\\n\",\n       \" 'gas pump',\\n\",\n       \" 'golf ball',\\n\",\n       \" 'parachute']\"\n      ]\n     },\n     \"execution_count\": 4,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# labels names \\n\",\n    \"labels = imagenette.info.features['label'].names\\n\",\n    \"labels\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 5,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"['a photo of a tench',\\n\",\n       \" 'a photo of a English springer',\\n\",\n       \" 'a photo of a cassette player',\\n\",\n       \" 'a photo of a chain saw',\\n\",\n       \" 'a photo of a church',\\n\",\n       \" 'a photo of a French horn',\\n\",\n       \" 'a photo of a garbage truck',\\n\",\n       \" 'a photo of a gas pump',\\n\",\n       \" 'a photo of a golf ball',\\n\",\n       \" 'a photo of a parachute']\"\n      ]\n     },\n     \"execution_count\": 5,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# generate sentences\\n\",\n    \"clip_labels = [f\\\"a photo of a {label}\\\" for label in labels]\\n\",\n    \"clip_labels\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

Before we can compare labels and photos, we need to initialize CLIP. We will use the CLIP implementation found via Hugging Face transformers.

```json
{
  "_key": "7d1aa702009e",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 6,\n   \"metadata\": {},\n   \"outputs\": [],\n   \"source\": [\n    \"# initialization\\n\",\n    \"from transformers import CLIPProcessor, CLIPModel\\n\",\n    \"\\n\",\n    \"model_id = \\\"openai/clip-vit-base-patch32\\\"\\n\",\n    \"\\n\",\n    \"processor = CLIPProcessor.from_pretrained(model_id)\\n\",\n    \"model = CLIPModel.from_pretrained(model_id)\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 7,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"'cpu'\"\n      ]\n     },\n     \"execution_count\": 7,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"import torch\\n\",\n    \"\\n\",\n    \"# if you have CUDA set it to the active device like this\\n\",\n    \"device = \\\"cuda\\\" if torch.cuda.is_available() else \\\"cpu\\\"\\n\",\n    \"# move the model to the device\\n\",\n    \"model.to(device)\\n\",\n    \"\\n\",\n    \"device\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

Text transformers cannot read text directly. Instead, they need a set of integer values known as token IDs (or `input_ids`), where each unique integer represents a word or sub-word (known as a _token_).

We create these token IDs alongside another tensor called the _attention mask_ (used by the transformer’s attention mechanism) using the `processor` we just initialized.

```json
{
  "_key": "4b5ba6ecbab9",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 8,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"tensor([49406,   320,  1125,   539,   320,  1149,   634, 49407])\"\n      ]\n     },\n     \"execution_count\": 8,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# create label tokens\\n\",\n    \"label_tokens = processor(\\n\",\n    \"    text=clip_labels,\\n\",\n    \"    padding=True,\\n\",\n    \"    images=None,\\n\",\n    \"    return_tensors='pt'\\n\",\n    \").to(device)\\n\",\n    \"\\n\",\n    \"label_tokens['input_ids'][0][:10]\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

Using these transformer-readable tensors, we create the label text embeddings like so:

```json
{
  "_key": "2973ee98f902",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 9,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"(10, 512)\"\n      ]\n     },\n     \"execution_count\": 9,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# encode tokens to sentence embeddings\\n\",\n    \"label_emb = model.get_text_features(**label_tokens)\\n\",\n    \"# detach from pytorch gradient computation\\n\",\n    \"label_emb = label_emb.detach().cpu().numpy()\\n\",\n    \"label_emb.shape\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 10,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"(-1.9079254, 6.1716247)\"\n      ]\n     },\n     \"execution_count\": 10,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"label_emb.min(), label_emb.max()\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

The vectors that CLIP outputs _are not normalized_, meaning dot product similarity will give inaccurate results _unless_ the vectors are normalized beforehand. We do that like so:

```json
{
  "_key": "813dbc723309",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 11,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"(-0.8760544, 0.8961743)\"\n      ]\n     },\n     \"execution_count\": 11,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"import numpy as np\\n\",\n    \"\\n\",\n    \"# normalization\\n\",\n    \"label_emb = label_emb / np.linalg.norm(label_emb, axis=0)\\n\",\n    \"label_emb.min(), label_emb.max()\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

_(Alternatively, you can use cosine similarity)_

All we have left is to work through the same process with the images from our dataset. We will test this with a single image first.

```json
{
  "_key": "6ccc7a6fa75f",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 12,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"image/png\": \"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\",\n      \"text/plain\": [\n       \"<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=328x320 at 0x17F1B1A90>\"\n      ]\n     },\n     \"execution_count\": 12,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"imagenette[0]['image']\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 13,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"torch.Size([1, 3, 224, 224])\"\n      ]\n     },\n     \"execution_count\": 13,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"image = processor(\\n\",\n    \"    text=None,\\n\",\n    \"    images=imagenette[0]['image'],\\n\",\n    \"    return_tensors='pt'\\n\",\n    \")['pixel_values'].to(device)\\n\",\n    \"image.shape\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

After processing the image, we return a single (`1`) three-color channel (`3`) image width of `224` pixels and a height of `224` pixels. We must process incoming images to normalize and resize them to fit the input size requirements of the ViT model.

We can create the image embedding with:

```json
{
  "_key": "036b0ba94353",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 14,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"torch.Size([1, 512])\"\n      ]\n     },\n     \"execution_count\": 14,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"img_emb = model.get_image_features(image)\\n\",\n    \"img_emb.shape\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 15,\n   \"metadata\": {},\n   \"outputs\": [],\n   \"source\": [\n    \"img_emb = img_emb.detach().cpu().numpy()\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

From here, all we need to do is calculate the dot product similarity between our image embedding and the ten label text embeddings. The highest score is our predicted class.

```json
{
  "_key": "31d92d794d93",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 16,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"(1, 10)\"\n      ]\n     },\n     \"execution_count\": 16,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"scores = np.dot(img_emb, label_emb.T)\\n\",\n    \"scores.shape\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 17,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"2\"\n      ]\n     },\n     \"execution_count\": 17,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# get index of highest score\\n\",\n    \"pred = np.argmax(scores)\\n\",\n    \"pred\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 18,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"'cassette player'\"\n      ]\n     },\n     \"execution_count\": 18,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"# find text label with highest score\\n\",\n    \"labels[pred]\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

Label _2_, i.e., _“cassette player”_ is our correctly predicted winner. We can repeat this logic over the entire `frgfm/imagenette` dataset to get the classification accuracy of CLIP.

```json
{
  "_key": "7025b1b29c9b",
  "_type": "colabBlock",
  "jsonContent": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 19,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"application/vnd.jupyter.widget-view+json\": {\n       \"model_id\": \"002eb535f2cc4e4cb404f5ee94e68fad\",\n       \"version_major\": 2,\n       \"version_minor\": 0\n      },\n      \"text/plain\": [\n       \"  0%|          | 0/123 [00:00<?, ?it/s]\"\n      ]\n     },\n     \"metadata\": {},\n     \"output_type\": \"display_data\"\n    }\n   ],\n   \"source\": [\n    \"from tqdm.auto import tqdm\\n\",\n    \"\\n\",\n    \"preds = []\\n\",\n    \"batch_size = 32\\n\",\n    \"\\n\",\n    \"for i in tqdm(range(0, len(imagenette), batch_size)):\\n\",\n    \"    i_end = min(i + batch_size, len(imagenette))\\n\",\n    \"    images = processor(\\n\",\n    \"        text=None,\\n\",\n    \"        images=imagenette[i:i_end]['image'],\\n\",\n    \"        return_tensors='pt'\\n\",\n    \"    )['pixel_values'].to(device)\\n\",\n    \"    img_emb = model.get_image_features(images)\\n\",\n    \"    img_emb = img_emb.detach().cpu().numpy()\\n\",\n    \"    scores = np.dot(img_emb, label_emb.T)\\n\",\n    \"    preds.extend(np.argmax(scores, axis=1))\"\n   ]\n  },\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 20,\n   \"metadata\": {},\n   \"outputs\": [\n    {\n     \"data\": {\n      \"text/plain\": [\n       \"0.9870063694267516\"\n      ]\n     },\n     \"execution_count\": 20,\n     \"metadata\": {},\n     \"output_type\": \"execute_result\"\n    }\n   ],\n   \"source\": [\n    \"true_preds = []\\n\",\n    \"for i, label in enumerate(imagenette['label']):\\n\",\n    \"    if label == preds[i]:\\n\",\n    \"        true_preds.append(1)\\n\",\n    \"    else:\\n\",\n    \"        true_preds.append(0)\\n\",\n    \"\\n\",\n    \"sum(true_preds) / len(true_preds)\"\n   ]\n  }\n ],\n \"metadata\": {\n  \"kernelspec\": {\n   \"display_name\": \"Python 3.8.13 ('image-embeds')\",\n   \"language\": \"python\",\n   \"name\": \"python3\"\n  },\n  \"language_info\": {\n   \"codemirror_mode\": {\n    \"name\": \"ipython\",\n    \"version\": 3\n   },\n   \"file_extension\": \".py\",\n   \"mimetype\": \"text/x-python\",\n   \"name\": \"python\",\n   \"nbconvert_exporter\": \"python\",\n   \"pygments_lexer\": \"ipython3\",\n   \"version\": \"3.8.13\"\n  },\n  \"orig_nbformat\": 4,\n  \"vscode\": {\n   \"interpreter\": {\n    \"hash\": \"9ec8fc8fb845fc3e050bf8bf651a355c069bbfeddee31167baf4bc42b6050476\"\n   }\n  }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 2\n}"
}
```

That gives us an impressive zero-shot accuracy of _98.7%_. CLIP proved to be able to accurately predict image classes with little more than some minor reformating of text labels to create sentences.

Zero-shot image classification with CLIP is a fascinating use case for high-performance image classification with minimal effort and zero fine-tuning required.

Before CLIP, this was not possible. Now that we have CLIP, it is almost _too_ easy. The multi-modality and contrastive pretraining techniques have enabled a technological leap forward.

From multi-modal search, zero-shot image classification, and object detection to industry-changing tools like OpenAI’s Dall-E and Stable Diffusion, CLIP has opened the floodgates to many new use-cases that were previously blocked by insufficient data or compute.

## Resources

[1] A. Radford, et. al., [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) (2021)