Spaces:

evaluate-metric

/

bleurt

like 10

Running

I am simply pasting the example given in the card -
"""
import evaluate

predictions = ["hello there", "general kenobi"]
references = ["hello there", "general kenobi"]
bleurt = evaluate.load("bleurt", module_type="metric", checkpoint="bleurt-tiny")
results = bleurt.compute(predictions=predictions, references=references)
print(results)
"""

but get the below error -
JIT compilation failed.
[[{{node bert/embeddings/LayerNorm/batchnorm/Rsqrt}}]] [Op:__inference_pruned_6256]

I am pretty new here - could someone help me