Skip to content

Move docs to new embeddinggemma-300m model #228

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Sep 25, 2025
Merged

Move docs to new embeddinggemma-300m model #228

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
233b328
Update docs-vectorize to the embeddinggemma model
mhart Sep 25, 2025
dee0a7b
Create social-mammals-chew.md
mhart Sep 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 7 additions & 0 deletions .changeset/social-mammals-chew.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
---
'workers-observability': minor
'workers-bindings': minor
'docs-vectorize': minor
---

Updated the model for docs search to embeddinggemma-300m
6 changes: 3 additions & 3 deletions apps/docs-vectorize/wrangler.jsonc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"vars": {
Expand Down Expand Up @@ -72,7 +72,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"analytics_engine_datasets": [
Expand Down Expand Up @@ -103,7 +103,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"analytics_engine_datasets": [
Expand Down
6 changes: 3 additions & 3 deletions apps/workers-bindings/wrangler.jsonc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"workers_dev": false,
Expand Down Expand Up @@ -92,7 +92,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"vars": {
Expand Down Expand Up @@ -141,7 +141,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"analytics_engine_datasets": [
Expand Down
6 changes: 3 additions & 3 deletions apps/workers-observability/wrangler.jsonc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"vars": {
Expand Down Expand Up @@ -98,7 +98,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"analytics_engine_datasets": [
Expand Down Expand Up @@ -136,7 +136,7 @@
"vectorize": [
{
"binding": "VECTORIZE",
"index_name": "docs-bge-base"
"index_name": "docs-embeddinggemma-v1"
}
],
"vars": {
Expand Down
16 changes: 8 additions & 8 deletions packages/mcp-common/src/tools/docs-vectorize.tools.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,6 +42,7 @@ export function registerDocsTools(agent: CloudflareMcpAgentNoAccount, env: Requi
.map((result) => {
return `<result>
<url>${result.url}</url>
<title>${result.title}</title>
<text>
${result.text}
</text>
Expand Down Expand Up @@ -93,10 +94,8 @@ ${result.text}
}

async function queryVectorize(ai: Ai, vectorizeIndex: VectorizeIndex, query: string, topK: number) {
// Recommendation from: https://huggingface.co/BAAI/bge-base-en-v1.5#model-list
const [queryEmbedding] = await getEmbeddings(ai, [
'Represent this sentence for searching relevant passages: ' + query,
])
// Recommendation from: https://ai.google.dev/gemma/docs/embeddinggemma/model_card#prompt_instructions
const [queryEmbedding] = await getEmbeddings(ai, ['task: search result | query: ' + query])

const { matches } = await vectorizeIndex.query(queryEmbedding, {
topK,
Expand All @@ -108,6 +107,7 @@ async function queryVectorize(ai: Ai, vectorizeIndex: VectorizeIndex, query: str
similarity: Math.min(match.score, 1),
id: match.id,
url: sourceToUrl(String(match.metadata?.filePath ?? '')),
title: String(match.metadata?.title ?? ''),
text: String(match.metadata?.text ?? ''),
}))
}
Expand All @@ -123,15 +123,15 @@ function sourceToUrl(path: string) {
)
}

async function getEmbeddings(ai: Ai, strings: string[]) {
async function getEmbeddings(ai: Ai, strings: string[]): Promise<number[][]> {
const response = await doWithRetries(() =>
ai.run('@cf/baai/bge-base-en-v1.5', {
// @ts-expect-error embeddinggemma not in types yet
ai.run('@cf/google/embeddinggemma-300m', {
text: strings,
// @ts-expect-error pooling not in types yet
pooling: 'cls',
})
)

// @ts-expect-error embeddinggemma not in types yet
return response.data
}

Expand Down