Minor issue with the Quickstart

#13
by whither-compendium - opened

Following the Quickstart users are likely to hit errors as it is expected one has pillow installed.

Quick Start

Install the sentence-transformers library:

pip install -U sentence-transformers transformers

Generate text embeddings:

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("google/embeddinggemma-2")

query = "What causes the northern lights?"
document = "The northern lights are caused by charged particles from the sun."

query_emb = model.encode(query, prompt_name="SearchQuery")
doc_emb = model.encode(document, prompt_name="Document")
print(model.similarity(query_emb, doc_emb))

Following this from a fresh install and users will likely get errors:

...
ImportError: EmbeddingGemma2Processor requires the PIL library but it was not found in your environment. You can install it with pip:
`pip install pillow`. Please note that you may need to restart your runtime after installation.

To install the required dependencies, run:
pip install -U "sentence-transformers[image]"

Simple fix on user side sure, but I feel dependency should have been included in the Quickstart

Google org

Hi @whither-compendium

Thanks for pointing this out! We have noted this internally. Good catch - since embeddinggemma-2 is a multimodal model, its processor requires pillow, which explains why clean environments run into this ImportError. Appreciate you sharing it!

Adding to this: Pillow alone isn't enough. With sentence-transformers 6.3.0 and transformers 5.23.0 on CPU, text-only use of SentenceTransformer("google/embeddinggemma-2") fails in two ways:

  • Without torchvision or Pillow: Could not load any image processor class for google/embeddinggemma-2 ... Missing optional dependencies: torchvision, Pillow.
  • With Pillow only: EmbeddingGemma2VideoProcessor requires the Torchvision library but it was not found in your environment.

pip install -U "sentence-transformers[image]" installs both and fixes it, so the Quick Start line could be:

pip install -U "sentence-transformers[image]" transformers

AI disclosure: an AI coding agent (Claude Code) ran the reproduction and drafted this comment at my direction. I reviewed it.

Sign up or log in to comment