Added code to map outputs back to the labels
Browse filesI added code to extract the probabilities and map them back to the class labels. I was a bit confused about how to do it and it took me a minute to figure it out. So I thought this might be helpful to others.
README.md
CHANGED
@@ -78,4 +78,20 @@ tokenizer = AutoTokenizer.from_pretrained("reachosen/autotrain-sdohv7-3701198597
|
|
78 |
inputs = tokenizer("The Patient is homeless", return_tensors="pt")
|
79 |
|
80 |
outputs = model(**inputs)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
81 |
```
|
|
|
78 |
inputs = tokenizer("The Patient is homeless", return_tensors="pt")
|
79 |
|
80 |
outputs = model(**inputs)
|
81 |
+
|
82 |
+
# Extract logits from the outputs
|
83 |
+
logits = outputs.logits
|
84 |
+
|
85 |
+
# Apply softmax to get probabilities
|
86 |
+
probabilities = F.softmax(logits, dim=1)
|
87 |
+
|
88 |
+
# get class mapping
|
89 |
+
configs_fp = "path to model files" + "config.json"
|
90 |
+
with open(configs_fp, 'r') as configs_file:
|
91 |
+
class_mapping = json.load(configs_file)['id2label']
|
92 |
+
|
93 |
+
# Create the class_probs dictionary
|
94 |
+
class_probs = {class_mapping[str(i)]: probabilities[0][i].item() for i in range(len(probabilities[0]))}
|
95 |
+
|
96 |
+
|
97 |
```
|