Add SetFit model
Browse files- README.md +24 -33
- config_setfit.json +2 -2
- model.safetensors +1 -1
- model_head.pkl +1 -1
README.md
CHANGED
@@ -9,21 +9,18 @@ base_model: sentence-transformers/all-roberta-large-v1
|
|
9 |
metrics:
|
10 |
- accuracy
|
11 |
widget:
|
12 |
-
- text: '
|
13 |
-
|
|
|
|
|
|
|
|
|
|
|
14 |
- text: ' But even then, I didn''t have any other problems outside school I still
|
15 |
had no friends at European school, I haven''t had any walks which I had constantly
|
16 |
with my friends back in Ukraine'
|
17 |
-
- text: ' I
|
18 |
-
|
19 |
-
me off'
|
20 |
-
- text: ' I literally do not leave the house, I will sometimes go in the garden but
|
21 |
-
no further So I just ran outside and back to the car'
|
22 |
-
- text: ' I hated even going outside and seeing other people, because I just knew
|
23 |
-
how awkward and weird they would think I am for the way I looked and the way I
|
24 |
-
walked I hated even going outside and seeing other people, because I just knew
|
25 |
-
how awkward and weird they would think I am for the way I looked and the way I
|
26 |
-
walked'
|
27 |
pipeline_tag: text-classification
|
28 |
inference: true
|
29 |
model-index:
|
@@ -38,7 +35,7 @@ model-index:
|
|
38 |
split: test
|
39 |
metrics:
|
40 |
- type: accuracy
|
41 |
-
value: 0.
|
42 |
name: Accuracy
|
43 |
---
|
44 |
|
@@ -70,19 +67,19 @@ The model has been trained using an efficient few-shot learning technique that i
|
|
70 |
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
|
71 |
|
72 |
### Model Labels
|
73 |
-
| Label | Examples
|
74 |
-
|
75 |
-
|
|
76 |
-
|
|
77 |
-
|
|
78 |
-
|
|
79 |
|
80 |
## Evaluation
|
81 |
|
82 |
### Metrics
|
83 |
| Label | Accuracy |
|
84 |
|:--------|:---------|
|
85 |
-
| **all** | 0.
|
86 |
|
87 |
## Uses
|
88 |
|
@@ -102,7 +99,7 @@ from setfit import SetFitModel
|
|
102 |
# Download from the 🤗 Hub
|
103 |
model = SetFitModel.from_pretrained("Omar-Nasr/setfitmodel")
|
104 |
# Run inference
|
105 |
-
preds = model(" I
|
106 |
```
|
107 |
|
108 |
<!--
|
@@ -134,14 +131,14 @@ preds = model(" I literally do not leave the house, I will sometimes go in the g
|
|
134 |
### Training Set Metrics
|
135 |
| Training set | Min | Median | Max |
|
136 |
|:-------------|:----|:-------|:----|
|
137 |
-
| Word count | 7 |
|
138 |
|
139 |
| Label | Training Sample Count |
|
140 |
|:------|:----------------------|
|
141 |
-
| 0.0 |
|
142 |
-
| 1.0 |
|
143 |
-
| 2.0 |
|
144 |
-
| 3.0 |
|
145 |
|
146 |
### Training Hyperparameters
|
147 |
- batch_size: (16, 16)
|
@@ -163,13 +160,7 @@ preds = model(" I literally do not leave the house, I will sometimes go in the g
|
|
163 |
### Training Results
|
164 |
| Epoch | Step | Training Loss | Validation Loss |
|
165 |
|:------:|:----:|:-------------:|:---------------:|
|
166 |
-
| 0.
|
167 |
-
| 0.6667 | 50 | 0.0497 | - |
|
168 |
-
| 1.3333 | 100 | 0.0012 | - |
|
169 |
-
| 2.0 | 150 | 0.0003 | - |
|
170 |
-
| 2.6667 | 200 | 0.0003 | - |
|
171 |
-
| 3.3333 | 250 | 0.0002 | - |
|
172 |
-
| 4.0 | 300 | 0.0001 | - |
|
173 |
|
174 |
### Framework Versions
|
175 |
- Python: 3.10.13
|
|
|
9 |
metrics:
|
10 |
- accuracy
|
11 |
widget:
|
12 |
+
- text: ' I would do this too but then I run the risk of meeting someone I know lol'
|
13 |
+
- text: ' That''s kind of the nature of my volunteer work, but you could volunteer
|
14 |
+
with a food bank or boys and girls club, which would involve more social interaction
|
15 |
+
Just breaking that cycle by going for a short walk around the neighbourhood is
|
16 |
+
a good idea'
|
17 |
+
- text: ' Your body is trying to reduce weight by throwing up and having to go to
|
18 |
+
the bathroon, in case you need to run from the "enemy", so you''ll be lighter'
|
19 |
- text: ' But even then, I didn''t have any other problems outside school I still
|
20 |
had no friends at European school, I haven''t had any walks which I had constantly
|
21 |
with my friends back in Ukraine'
|
22 |
+
- text: ' I like art and nature but you can’t really talk about those for more than
|
23 |
+
a few seconds'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
pipeline_tag: text-classification
|
25 |
inference: true
|
26 |
model-index:
|
|
|
35 |
split: test
|
36 |
metrics:
|
37 |
- type: accuracy
|
38 |
+
value: 0.3416666666666667
|
39 |
name: Accuracy
|
40 |
---
|
41 |
|
|
|
67 |
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
|
68 |
|
69 |
### Model Labels
|
70 |
+
| Label | Examples |
|
71 |
+
|:------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
|
72 |
+
| 1.0 | <ul><li>' and i do biking, i bike for atleast 20mins and towards the end i feel anxious free, less socially anxious too and i do biking, i bike for atleast 20mins and towards the end i feel anxious free, less socially anxious too'</li><li>" That's kind of the nature of my volunteer work, but you could volunteer with a food bank or boys and girls club, which would involve more social interaction Just breaking that cycle by going for a short walk around the neighbourhood is a good idea"</li></ul> |
|
73 |
+
| 2.0 | <ul><li>" I didn't go outside too much"</li><li>' I like art and nature but you can’t really talk about those for more than a few seconds'</li></ul> |
|
74 |
+
| 3.0 | <ul><li>" I don't know anyone in the city, I don't think I have enough courage to go outside alone"</li><li>" But even then, I didn't have any other problems outside school I still had no friends at European school, I haven't had any walks which I had constantly with my friends back in Ukraine"</li></ul> |
|
75 |
+
| 0.0 | <ul><li>' Your body is trying to reduce weight by throwing up and having to go to the bathroon, in case you need to run from the "enemy", so you\'ll be lighter'</li><li>' I would do this too but then I run the risk of meeting someone I know lol'</li></ul> |
|
76 |
|
77 |
## Evaluation
|
78 |
|
79 |
### Metrics
|
80 |
| Label | Accuracy |
|
81 |
|:--------|:---------|
|
82 |
+
| **all** | 0.3417 |
|
83 |
|
84 |
## Uses
|
85 |
|
|
|
99 |
# Download from the 🤗 Hub
|
100 |
model = SetFitModel.from_pretrained("Omar-Nasr/setfitmodel")
|
101 |
# Run inference
|
102 |
+
preds = model(" I would do this too but then I run the risk of meeting someone I know lol")
|
103 |
```
|
104 |
|
105 |
<!--
|
|
|
131 |
### Training Set Metrics
|
132 |
| Training set | Min | Median | Max |
|
133 |
|:-------------|:----|:-------|:----|
|
134 |
+
| Word count | 7 | 27.375 | 47 |
|
135 |
|
136 |
| Label | Training Sample Count |
|
137 |
|:------|:----------------------|
|
138 |
+
| 0.0 | 2 |
|
139 |
+
| 1.0 | 2 |
|
140 |
+
| 2.0 | 2 |
|
141 |
+
| 3.0 | 2 |
|
142 |
|
143 |
### Training Hyperparameters
|
144 |
- batch_size: (16, 16)
|
|
|
160 |
### Training Results
|
161 |
| Epoch | Step | Training Loss | Validation Loss |
|
162 |
|:------:|:----:|:-------------:|:---------------:|
|
163 |
+
| 0.3333 | 1 | 0.0708 | - |
|
|
|
|
|
|
|
|
|
|
|
|
|
164 |
|
165 |
### Framework Versions
|
166 |
- Python: 3.10.13
|
config_setfit.json
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
{
|
2 |
-
"
|
3 |
-
"
|
4 |
}
|
|
|
1 |
{
|
2 |
+
"normalize_embeddings": false,
|
3 |
+
"labels": null
|
4 |
}
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1421483904
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:914b970e09e034c7be1d72ee9d1c69d66f11251d67f58c62db0972b5736d0e38
|
3 |
size 1421483904
|
model_head.pkl
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 33639
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:15bf413c69462d4ffdc39a30464d39f872fd6fdcca7666f5bdfe0e7fcfe51594
|
3 |
size 33639
|