Sentence Transformers Segmentation fault

Question

I get a Segmentation fault error when calling model.encode on a SentenceTransformer model:

Segmentation fault
root@0ac58308616e:/app# /usr/local/lib/python3.10/multiprocessing/resource_tracker.py:224: UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown
  warnings.warn('resource_tracker: There appear to be %d '

The environment is Docker:

FROM python:3.8-slim-buster

RUN apt-get update && apt-get install -y \
    software-properties-common \
    build-essential \
    pkg-config \
    ninja-build \
    libopenblas-dev \
    python3-pip \
    curl

COPY . .
CMD ["bash"]

root@0ac58308616e:/app# python -c "import torch; print(torch.__version__);"
2.1.0
root@0ac58308616e:/app# python -c "import transformers; print(transformers.__version__);"
4.34.1
root@0ac58308616e:/app# python -c "import sentence_transformers; print(sentence_transformers.__version__);"
2.2.2

Code to reproduce:

from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]

model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
embeddings = model.encode(sentences)
print(embeddings)

This happens also using transformers library directly:

from transformers import AutoTokenizer, AutoModel
import torch
import torch.nn.functional as F

#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)


# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/all-MiniLM-L6-v2',cache_dir='models')
model = AutoModel.from_pretrained('sentence-transformers/all-MiniLM-L6-v2',cache_dir='models')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)

# Perform pooling
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])

# Normalize embeddings
sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1)

print("Sentence embeddings:")
print(sentence_embeddings)