Add SetFit model

README.md

@@ -8,11 +8,12 @@ tags:
 metrics:
 - accuracy
 widget:
-- text: Can you name three different types of weather?
-- text: Which state is known as the Sunshine State?
-- text: If a dozen eggs cost $3, how much do 3 dozen eggs cost?
-- text: What is the name of the mountain range on the west coast?
-- text: Can you name two different habitats where animals live?
+- text: Can you name three different types of fruits?
+- text: What is the capital city of your state?
+- text: If 2 apples cost $1, how much would 5 apples cost?
+- text: John had 8 marbles. He lost 4 marbles and then got 3 new ones. How many marbles
+    does John have now?
+- text: What is the name of the civil rights leader who said 'I have a dream'?
 pipeline_tag: text-classification
 inference: true
 base_model: sentence-transformers/all-MiniLM-L6-v2
@@ -60,10 +61,10 @@ The model has been trained using an efficient few-shot learning technique that i
 - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
 
 ### Model Labels
-| Label    | Examples                                                                                                                                                                                                               |
-|:---------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
-| math     | <ul><li>'Which is more: 7 or 9?'</li><li>"How can you measure the amount of liquid in a cup if you don't have a measuring cup?"</li><li>'What is the name of a shape with 5 equal sides and 5 equal angles?'</li></ul> |
-| non_math | <ul><li>'Can you name two different holidays celebrated in the springtime?'</li><li>'Who discovered gravity?'</li><li>'Energy'</li></ul>                                                                               |
+| Label    | Examples                                                                                                                                                                                                                                                                            |
+|:---------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| math     | <ul><li>'Which unit would you use to measure how much milk you need for your cereal: cups or gallons?'</li><li>'What is the volume of a cube with side length 4 cm?'</li><li>'If school starts at 8:30 AM and ends at 3:15 PM, how many hours are there in a school day?'</li></ul> |
+| non_math | <ul><li>'What is the name of the long river that runs through the middle of the US?'</li><li>'What do we call the action of objects changing their position?'</li><li>'What is the currency used in Japan?'</li></ul>                                                               |
 
 ## Evaluation
 
@@ -90,7 +91,7 @@ from setfit import SetFitModel
 # Download from the 🤗 Hub
 model = SetFitModel.from_pretrained("serdarcaglar/primary-school-math-question")
 # Run inference
-preds = model("Which state is known as the Sunshine State?")
+preds = model("What is the capital city of your state?")
 ```
 
 <!--
@@ -122,12 +123,12 @@ preds = model("Which state is known as the Sunshine State?")
 ### Training Set Metrics
 | Training set | Min | Median  | Max |
 |:-------------|:----|:--------|:----|
-| Word count   | 1   | 12.5232 | 33  |
+| Word count   | 1   | 12.5378 | 33  |
 
 | Label    | Training Sample Count |
 |:---------|:----------------------|
 | math     | 141                   |
-| non_math | 96                    |
+| non_math | 97                    |
 
 ### Training Hyperparameters
 - batch_size: (16, 16)
@@ -150,18 +151,18 @@ preds = model("Which state is known as the Sunshine State?")
 ### Training Results
 | Epoch  | Step | Training Loss | Validation Loss |
 |:------:|:----:|:-------------:|:---------------:|
-| 0.0017 | 1    | 0.3751        | -               |
-| 0.0843 | 50   | 0.1394        | -               |
-| 0.1686 | 100  | 0.0114        | -               |
-| 0.2530 | 150  | 0.0045        | -               |
-| 0.3373 | 200  | 0.0026        | -               |
-| 0.4216 | 250  | 0.0023        | -               |
-| 0.5059 | 300  | 0.0014        | -               |
-| 0.5902 | 350  | 0.0014        | -               |
-| 0.6745 | 400  | 0.0004        | -               |
-| 0.7589 | 450  | 0.0008        | -               |
-| 0.8432 | 500  | 0.0007        | -               |
-| 0.9275 | 550  | 0.0008        | -               |
+| 0.0017 | 1    | 0.3115        | -               |
+| 0.0840 | 50   | 0.1498        | -               |
+| 0.1681 | 100  | 0.0127        | -               |
+| 0.2521 | 150  | 0.0056        | -               |
+| 0.3361 | 200  | 0.0019        | -               |
+| 0.4202 | 250  | 0.0007        | -               |
+| 0.5042 | 300  | 0.0016        | -               |
+| 0.5882 | 350  | 0.0019        | -               |
+| 0.6723 | 400  | 0.0005        | -               |
+| 0.7563 | 450  | 0.0009        | -               |
+| 0.8403 | 500  | 0.0009        | -               |
+| 0.9244 | 550  | 0.0008        | -               |
 
 ### Framework Versions
 - Python: 3.10.12

config_setfit.json

@@ -1,7 +1,7 @@
 {
+  "normalize_embeddings": false,
   "labels": [
     "math",
     "non_math"
-  ],
-  "normalize_embeddings": false
+  ]
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0e330952e79f9c57d40ea47b439bfaf3667cf66eff839aae46f3645bebaab310
+oid sha256:7c383fc411cebd9c2fd39b6391cd0c511b5c703e0c0b1cba4b768b9206d2cdb3
 size 90864192

model_head.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0a327dca371ff12fbf0b79715aef5a8ea9da7b5df410bdaecd27092fc692fb16
+oid sha256:54154be49871204f7afa49967c5b22cf6d928cb8250214a377450976b36b7891
 size 3967