Public
Edited
Oct 22, 2023
4 forks
2 stars
browse_image.image()
score = cosineSimilarity(
textEmbeddings.text_embeds.data,
embeddings.image_embeds.data
)
html`<textarea style="width: 100%; height: 10em">${JSON.stringify(
Array.from(textEmbeddings.text_embeds.data)
)}</textarea>`
transformers = import("https://cdn.jsdelivr.net/npm/@xenova/transformers")
processor = await transformers.AutoProcessor.from_pretrained(
"Xenova/clip-vit-base-patch16"
)
vision_model = await transformers.CLIPVisionModelWithProjection.from_pretrained(
"Xenova/clip-vit-base-patch16"
)
default_image = await transformers.RawImage.read(
"https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/football-match.jpg"
)
image = browse_image
? await transformers.RawImage.fromBlob(await browse_image.blob())
: await transformers.RawImage.read(
"https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/football-match.jpg"
)
image
image_inputs = await processor(image)
embeddings = await vision_model(image_inputs)
embeddings.image_embeds.data
tokenizer = await transformers.AutoTokenizer.from_pretrained(
"Xenova/clip-vit-base-patch16"
)
text_model = await transformers.CLIPTextModelWithProjection.from_pretrained(
"Xenova/clip-vit-base-patch16"
)
text_inputs = tokenizer(texts, { padding: true, truncation: true })
textEmbeddings = await text_model(text_inputs)
function dotProduct(a, b) {
let sum = 0;
for (let i = 0; i < a.length; i++) {
sum += a[i] * b[i];
}
return sum;
}
function magnitude(a) {
return Math.sqrt(dotProduct(a, a));
}
function cosineSimilarity(a, b) {
if (a.length !== b.length) {
throw new Error("Vectors must have the same dimensions");
}
const magA = magnitude(a);
const magB = magnitude(b);
if (magA === 0 || magB === 0) {
throw new Error(
"Magnitude of one of the vectors is zero, cannot calculate cosine similarity"
);
}
return dotProduct(a, b) / (magA * magB);
}
texts = [text]
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.