Choose an embedding model
Before configuring your backend or running any scripts, decide which embedding model you want to use.
The embedding model will convert each movie plot summary into a fixed-size vector. The dimension of that vector depends on your chosen model, and you must configure this number in your backend.
Clone the repo
If you haven’t done so already, start by cloning the project repository:
git clone https://bitbucket.org/chromawallet/vector-db-movie-demo.git
cd vector-db-movie-demo
Create your .env file
Navigate to the python/ folder and create your .env file by copying the template:
cd python
cp .env.template .env
You will edit this file in the next step.
Select your model
Open the .env file in your editor and pick one of the available EMBEDDING_MODEL options by uncommenting the relevant line.
For example:
EMBEDDING_MODEL=sentence-transformers/all-mpnet-base-v2
This field controls which sentence embedding model will be used when vectorizing movie summaries.
Make sure to note the dimensions of the model you choose, as you'll need it in the next lesson to set up your Rell module.
Feel free to substitute any other model from the Hugging Face Hub, provided it's compatible with the sentence-transformers library and produces dense vectors.
Model overview
| Model | Dimensions | Notes |
|---|---|---|
all-MiniLM-L6-v2 | 384 | Small and fast |
all-mpnet-base-v2 | 768 | Balanced, a good default choice |
all-mpnet-large-v2 | 1024 | High quality but slower |
intfloat/e5-large-v2 | 1024 | Instruction-tuned, best for search |
BAAI/bge-large-en-v1.5 | 1024 | Alternative model with excellent performance |
Choosing the right model
Larger models can give better semantic results but require more memory and processing time.
If you're using a GPU, you can likely use any of the models listed above without issues.
If you're running on CPU, you may want to stick to a smaller model to avoid slow performance.
Next, you’ll use the model’s dimensions to configure your Rell module in chromia.yml.