Similarity Search & Embeddings Tutorial Not Working

[georgesphicas]

I've been trying to get the NAIP Inference and Similarity Search with Clay tutorial to work in my environment, but I am running into problems.

I was wondering if I could please have some help getting them to work? If it helps, the part of the similarity search that I am getting stuck on is with the iteration through the chips step. When I run this block of code in my jupyter notebook, I get the following output:

Asset height 247 is not a multiple of highest resolution height 12270.

We've tried to resive the target width and height and we continue to get the same error.

Do the images have to be of a certain size/resolution for the process to work? If so, do you have insight as to why the tutorial fails even when following each step as it's written?

[srmsoumya]

Thanks for starting this thread @georgesphicas - Can you link to the tutorial notebook you are trying & the step at which you are facing the issue please?

[georgesphicas]

Hello @srmsoumya ,

the tutorial notebook I am trying is https://clay-foundation.github.io/model/tutorials/v1-inference-simsearch-naip-stacchip.html

the step that has the issue is this one:
for item in items_list[:2]:
print(f"Working on {item}")

# Index the chips in the item
indexer = NoStatsChipIndexer(item)

# Obtain the item bounds and centroid
bounds, centroid_x, centroid_y = get_bounds_centroid(item.assets["image"].href)
print(
    f"Bbox coordinates: {bounds}, centroid coordinates: {centroid_x}, {centroid_y}"
)

# Instantiate the chipper
chipper = Chipper(indexer, asset_blacklist=["metadata"])

# Get 5 randomly sampled chips from the total
# number of chips within this item's entire image
for chip_id in random.sample(range(0, len(chipper)), 5):
    chip_images.append(chipper[chip_id]["image"])
    chip_bounds.append(bounds)

Thank you!