Update README.md

README.md

@@ -140,7 +140,7 @@ The model was trained in a distributed fashion using a single node with 4 16GB N
 
 #### Software
 
-As discussed above, we propose a solution to a set of lonnstanding issues in the 🤗 Evaluate library. While awaiting review on our proposal, we temporarily define a new 
+As discussed above, we propose a solution to a set of longstanding issues in the 🤗 Evaluate library. While awaiting review on our proposal, we temporarily define a new 
 set of evaluation metrics by subclassing the 🤗 Evaluate `Metric` to introduce more general multilabel classification accuracy, precision, f1, and recall metrics.
 
 Training utilized PyTorch, Apache Spark, 🤗 Transformers, 🤗 Tokenizers, 🤗 Evaluate, 🤗 Datasets, and more in an Azure Databricks execution environment.