embeddings.search hangs indefinitely on Mac

[tf13]

Running OS 12.7 (Monterrey) on a MacBook Air (M1, 2020) with Python 3.11.5

Script below — it should be effectively identical to the first example at https://neuml.hashnode.dev/tutorial-series-on-txtai

It failed in Spyder and also from the command line in Terminal (both with python3 scriptname.py and line by line after running python3 on the command line).

I then tried the same script on an Linux box I have access to, running Ubuntu 22.04.3 LTS. It worked fine as-is from the command line.

from txtai import Embeddings

# temp data
data = [
  "US tops 5 million confirmed virus cases",
  "Canada's last fully intact ice shelf has suddenly collapsed, forming a Manhattan-sized iceberg",
  "Beijing mobilises invasion craft along coast as Taiwan tensions escalate",
  "The National Park Service warns against sacrificing slower friends in a bear attack",
  "Maine man wins $1M from $25 lottery ticket",
  "Make huge profits without work, earn up to $100,000 a day"
]

# create embeddings
embeddings = Embeddings(path="sentence-transformers/nli-mpnet-base-v2")

# create index for list of texts
embeddings.index(data)

print(f"{'Query':20} Best Match")
print("-" * 50)


for query in ("feel good story", "climate change", "public health story", "war",
              "wildlife", "asia", "lucky", "dishonest junk"):
    # Extract uid of first result
    # search result format: (uid, score)
    uid = embeddings.search(query)[0][0]

    # Print text
    print(f"{query:20} {data[uid]}")

[davidmezzetti]

What happens if you change this line:

embeddings = Embeddings(path="sentence-transformers/nli-mpnet-base-v2")

to

embeddings = Embeddings(path="sentence-transformers/nli-mpnet-base-v2", gpu=False)

[tf13]

It froze just as before. This time I left it running to see whether it would ever resolve. More than 11 hours later, it is still hung after printing the headers.

[davidmezzetti]

Ok, not exactly sure what could be going on. Outside of managing the build script, I don't run txtai in a macOS environment frequently.

Perhaps you can try something things as seen in the build script such as a different Python version or setting the OMP_NUM_THREADS=1 environment variable. There are also a couple ideas in the FAQ

[davidmezzetti]

Closing this due to inactivity. Please re-open or open a new issue if there are further questions.

[tf13]

Sorry, I should have reported back: OMP_NUM_THREADS =1 does seem to do the trick.