# Custom embeddings vs. embedding function

**URL:** <https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773>\
**Category:** Support\
**Created:** [October 3, 2023, 2:53pm UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773 "2023-10-03T14:53:57Z")\
**Posts on this page:** 8\
**Page:** 1

<div class="post-metadata">

**Author:** ![Adam\_Hughes](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/adam_hughes/32/124_2.png) [@Adam\_Hughes](https://forum.weaviate.io/u/Adam_Hughes)\
**Post date:** [October 3, 2023, 2:53pm UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/1 "2023-10-03T14:53:57Z")

</div>

In Chroma, I can create an embedding function and pass it as:

```
    self.collection = self.client.create_collection(
        name="Foo"
        embedding_function=myembedder
    )

```

Is this possible in weaviate as well? I understand that weaviate [supports numerous builtin vectorizers](https://weaviate.io/developers/weaviate/modules/retriever-vectorizer-modules/text2vec-huggingface), but hadn’t seen any examples of a client-created vectorizer.

---

<div class="post-metadata">

**Author:** ![jphwang](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/jphwang/32/38_2.png) [@jphwang](https://forum.weaviate.io/u/jphwang)\
**Post date:** [October 4, 2023, 10:35am UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/2 "2023-10-04T10:35:36Z")

</div>

Hi @Adam_Hughes 👋.

You can pass in the raw vector upon update like this: [Quickstart Tutorial | Weaviate - vector database](https://weaviate.io/developers/weaviate/quickstart#option-2-custom-vectors)

So, you could have your vectorizer function pre-generate vectors and use them at import time, or you could have the vectorizer work as a part of the import process.

---

<div class="post-metadata">

**Author:** ![sebawita](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/sebawita/32/5_2.png) [@sebawita](https://forum.weaviate.io/u/sebawita)\
**Post date:** [October 4, 2023, 5:49pm UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/3 "2023-10-04T17:49:11Z")

</div>

Here is a quick example, of how to use Weaviate without a vectorizer in Python:

> Note #1: the example uses Python client `v3` – we are about to release v4, which will change the syntax 😉

> Note #2: I haven’t tested the code, some parts are a bit made up like `"grab_your_data()"`

## Connect to Weaviate (docker example)

```auto
import weaviate

client = weaviate.Client(
    url = "http://localhost:8080", # Replace with your endpoint
)

```

## Create collection

```auto
client.schema.create_class({
    "class": "Foo",
    "vectorizer": "none",
    "vectorIndexConfig": {
        "distance": "cosine" # make sure to provide the distance metric to search through your vectors
    },
})

```

## Insert object

```auto
data = grab_your_data() # Load data

client.batch.configure(batch_size=10) # Configure batch

with client.batch as batch: # Configure a batch process
    for item in data: # Loop through your data objects

        # construct the properties for your object
        properties = {
            "a": item["A_Field"], # this should correspond to your data structure
            "b": item["B_Field"], 
        }
        vector_value=item["vector"] # grab the vector from your object

        batch.add_data_object(
            class_name="Foo",
            data_object=properties,
            vector=vector_value # Your vector goes here
        )

```

## Vector query

```auto
response = (
    client.query
    .get("Foo", ["a", "b"])
    .with_near_vector({ 
        "vector": [-0.0125526935, -0.021168863, -0.01076519, -0.02589537, -0.0070362035, 0.019870078, -0.010001986, -0.019120263, 0.00090044655, -0.017393013, 0.021302758]
    }) # your query vector goes here ^^^
    .with_limit(5)
    .do()
)

```

---

<div class="post-metadata">

**Author:** ![sebawita](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/sebawita/32/5_2.png) [@sebawita](https://forum.weaviate.io/u/sebawita)\
**Post date:** [October 5, 2023, 10:16am UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/4 "2023-10-05T10:16:10Z")

</div>

Here is a tested [recipe – Jupyter Notebook](https://github.com/weaviate/recipes/blob/main/data-with-vectors/vector_search.ipynb) that shows how to:

1. Create a new collection without a vectorizer
2. Insert data with vectors
3. Perform vector search
4. Perform near object search

---

<div class="post-metadata">

**Author:** ![Adam\_Hughes](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/adam_hughes/32/124_2.png) [@Adam\_Hughes](https://forum.weaviate.io/u/Adam_Hughes)\
**Post date:** [October 6, 2023, 2:42pm UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/5 "2023-10-06T14:42:24Z")

</div>

Thank you for these great examples. I really can’t overstate how helpful they are!.

(tangent) My last point of confusion, conceptually, is vectorizing a search query. Imagine I had a custom vectorizer and used it to vectorize 50 wikipedia articles.

```
 myvectorizer.vectorize([article1, article2, ...])

```

Then I wanted to pass a user query like “Tell me about the Red Baron”.

I’d simply pass the user query directly into the same vectorizer, right? Then use `nearVector`? (psuedocode)

```
query_vec = myvectorizer.vectorize("Tell me about the Red Baron")
client.nearVector(query_vec)

```

Is this the approach you’d use in my situation? Or is there more to it than just vectorizing the search query.

---

<div class="post-metadata">

**Author:** ![jphwang](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/jphwang/32/38_2.png) [@jphwang](https://forum.weaviate.io/u/jphwang)\
**Post date:** [October 6, 2023, 3:51pm UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/6 "2023-10-06T15:51:21Z")

</div>

Hi @Adam_Hughes - sounds like you’ve got it!

The Weaviate `near_text` function does exactly what you describe, but in an integrated way. So your pseudocode is absolutely correct. 🙂

---

<div class="post-metadata">

**Author:** ![Francesco\_Gianferrar](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/francesco_gianferrar/32/552_2.png) [@Francesco\_Gianferrar](https://forum.weaviate.io/u/Francesco_Gianferrar)\
**Post date:** [October 7, 2023, 6:43pm UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/7 "2023-10-07T18:43:38Z")

</div>

Is there a way to pass your own vector values for a clip like model (not using a vectorizer)? I need to pass, on insert, both the image and the text vector embeddings that I calculate offline in my GPU box…

Thanks!

---

<div class="post-metadata">

**Author:** ![jphwang](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.weaviate.io/jphwang/32/38_2.png) [@jphwang](https://forum.weaviate.io/u/jphwang)\
**Post date:** [October 17, 2023, 10:48am UTC](https://forum.weaviate.io/t/custom-embeddings-vs-embedding-function/773/8 "2023-10-17T10:48:43Z")

</div>

Hi @Francesco_Gianferrar - yes, you can use `near_vector` queries to find objects most similar to a given vector.

> **[Similarity / Vector search | Weaviate - vector database](https://weaviate.io/developers/weaviate/search/similarity#a-vector)**
>
> Overview

Even if you did use a Weaviate vectorizer, you can still use this function - as long as the vectors are compatible, of course. 🙂