Update README.md

README.md

@@ -122,6 +122,37 @@ Using a fixed threshold of 0.5 to convert the scores to binary predictions for e
 
 ### Use with ONNXRuntime
 
+The input to the model is called `logits`, and there is one output per label. Each output produces a 2d array, with 1 row per input row, and each row having 2 columns - the first being a proba output for the negative case, and the second being a proba output for the positive case.
+
 ```python
-pass
+# Assuming you have embeddings from BAAI/bge-small-en-v1.5 for the input sentences
+# E.g. produced from sentence-transformers E.g. huggingface.co/BAAI/bge-small-en-v1.5
+#      or from an ONNX version E.g. huggingface.co/Xenova/bge-small-en-v1.5
+
+print(sentences.shape)  # E.g. a batch of 1 sentence
+> (1, 384)
+
+import onnxruntime as ort
+
+sess = ort.InferenceSession(
+  "path_to_model_dot_onnx",
+  providers=['CPUExecutionProvider'],
+)
+
+outputs = [o.name for o in sess.get_outputs()]
+preds_onnx = sess.run(_outputs, {'logits': _label_embeddings})
+# preds_onnx is a list with 28 entries, each with a numpy array of shape (1, 2)
+
+print(outputs[0])
+# surprise
+print(preds_onnx[0])
+# array([[0.97136074, 0.02863926]], dtype=float32)
 ```
+
+### Commentary on the dataset
+
+Some labels (E.g. gratitude) when considered independently perform very strongly, whilst others (E.g. relief) perform very poorly.
+
+This is a challenging dataset. Labels such as relief do have much fewer examples in the training data (less than 100 out of the 40k+, and only 11 in the test split).
+
+But there is also some ambiguity and/or labelling errors visible in the training data of go_emotions that is suspected to constrain the performance. Data cleaning on the dataset to reduce some of the mistakes, ambiguity, conflicts and duplication in the labelling would produce a higher performing model.