Add SetFit model
Browse files- README.md +59 -81
- config_sentence_transformers.json +1 -1
- config_setfit.json +2 -2
- model.safetensors +1 -1
- model_head.pkl +1 -1
README.md
CHANGED
@@ -13,39 +13,45 @@ tags:
|
|
13 |
- text-classification
|
14 |
- generated_from_setfit_trainer
|
15 |
widget:
|
16 |
-
- text:
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
inference: false
|
50 |
model-index:
|
51 |
- name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2
|
@@ -59,16 +65,16 @@ model-index:
|
|
59 |
split: test
|
60 |
metrics:
|
61 |
- type: accuracy
|
62 |
-
value: 0.
|
63 |
name: Accuracy
|
64 |
- type: precision
|
65 |
-
value: 0.
|
66 |
name: Precision
|
67 |
- type: recall
|
68 |
-
value: 0.
|
69 |
name: Recall
|
70 |
- type: f1
|
71 |
-
value: 0.
|
72 |
name: F1
|
73 |
---
|
74 |
|
@@ -104,7 +110,7 @@ The model has been trained using an efficient few-shot learning technique that i
|
|
104 |
### Metrics
|
105 |
| Label | Accuracy | Precision | Recall | F1 |
|
106 |
|:--------|:---------|:----------|:-------|:-------|
|
107 |
-
| **all** | 0.
|
108 |
|
109 |
## Uses
|
110 |
|
@@ -124,7 +130,7 @@ from setfit import SetFitModel
|
|
124 |
# Download from the 🤗 Hub
|
125 |
model = SetFitModel.from_pretrained("lgd/setfit-multilabel")
|
126 |
# Run inference
|
127 |
-
preds = model("
|
128 |
```
|
129 |
|
130 |
<!--
|
@@ -156,11 +162,11 @@ preds = model("neighbourhood")
|
|
156 |
### Training Set Metrics
|
157 |
| Training set | Min | Median | Max |
|
158 |
|:-------------|:----|:-------|:----|
|
159 |
-
| Word count | 1 | 4.
|
160 |
|
161 |
### Training Hyperparameters
|
162 |
- batch_size: (16, 16)
|
163 |
-
- num_epochs: (
|
164 |
- max_steps: -1
|
165 |
- sampling_strategy: oversampling
|
166 |
- num_iterations: 20
|
@@ -179,52 +185,24 @@ preds = model("neighbourhood")
|
|
179 |
### Training Results
|
180 |
| Epoch | Step | Training Loss | Validation Loss |
|
181 |
|:-----:|:----:|:-------------:|:---------------:|
|
182 |
-
| 0.
|
183 |
-
| 0.
|
184 |
-
| 0.
|
185 |
-
| 0.
|
186 |
-
| 0.
|
187 |
-
|
|
188 |
-
| 0.
|
189 |
-
| 0.
|
190 |
-
| 0.
|
191 |
-
| 0.
|
192 |
-
| 0
|
193 |
-
| 1.0 | 250 | 0.0302 | - |
|
194 |
-
| 0.004 | 1 | 0.0151 | - |
|
195 |
-
| 0.2 | 50 | 0.0232 | - |
|
196 |
-
| 0.4 | 100 | 0.017 | - |
|
197 |
-
| 0.6 | 150 | 0.0133 | - |
|
198 |
-
| 0.8 | 200 | 0.0629 | - |
|
199 |
-
| 1.0 | 250 | 0.0349 | - |
|
200 |
-
| 1.2 | 300 | 0.0585 | - |
|
201 |
-
| 1.4 | 350 | 0.0658 | - |
|
202 |
-
| 1.6 | 400 | 0.0446 | - |
|
203 |
-
| 1.8 | 450 | 0.0073 | - |
|
204 |
-
| 2.0 | 500 | 0.0326 | - |
|
205 |
-
| 0.004 | 1 | 0.017 | - |
|
206 |
-
| 0.2 | 50 | 0.0038 | - |
|
207 |
-
| 0.4 | 100 | 0.0095 | - |
|
208 |
-
| 0.6 | 150 | 0.0154 | - |
|
209 |
-
| 0.8 | 200 | 0.0444 | - |
|
210 |
-
| 1.0 | 250 | 0.0221 | - |
|
211 |
-
| 1.2 | 300 | 0.0362 | - |
|
212 |
-
| 1.4 | 350 | 0.0565 | - |
|
213 |
-
| 1.6 | 400 | 0.0338 | - |
|
214 |
-
| 1.8 | 450 | 0.0081 | - |
|
215 |
-
| 2.0 | 500 | 0.0299 | - |
|
216 |
-
| 2.2 | 550 | 0.106 | - |
|
217 |
-
| 2.4 | 600 | 0.0191 | - |
|
218 |
-
| 2.6 | 650 | 0.0104 | - |
|
219 |
-
| 2.8 | 700 | 0.0369 | - |
|
220 |
-
| 3.0 | 750 | 0.024 | - |
|
221 |
|
222 |
### Framework Versions
|
223 |
- Python: 3.10.12
|
224 |
- SetFit: 1.0.3
|
225 |
- Sentence Transformers: 3.0.1
|
226 |
- Transformers: 4.39.0
|
227 |
-
- PyTorch: 2.3.
|
228 |
- Datasets: 2.20.0
|
229 |
- Tokenizers: 0.15.2
|
230 |
|
|
|
13 |
- text-classification
|
14 |
- generated_from_setfit_trainer
|
15 |
widget:
|
16 |
+
- text: weather satellite imagery update every 10 minute cloud top temperature colorized
|
17 |
+
reveal area intensity lower level transparent satellite imagery combine data noaa
|
18 |
+
go east west satellite jma himawari satellite providing full coverage weather
|
19 |
+
event world west coast africa west east coast india tile service update recent
|
20 |
+
image every 10 minute 15 km per pixel resolution infrared ir band detects radiation
|
21 |
+
emitted earth???s surface atmosphere cloud ??·infrared window??? portion spectrum
|
22 |
+
radiation wavelength near 103 micrometer term ??·window??? mean pass atmosphere
|
23 |
+
relatively little absorption gas water vapor useful estimating emitting temperature
|
24 |
+
earth???s surface cloud top major advantage ir band sense energy night imagery
|
25 |
+
available 24 hour day advanced baseline imager abi instrument sample radiance
|
26 |
+
earth sixteen spectral band using several array detector instrument???s focal
|
27 |
+
plane single reflective band abi level 1b radiance product channel 1 6 approximate
|
28 |
+
center wavelength 047 064 0865 1378 161 225 micron respectively digital map outgoing
|
29 |
+
radiance value top atmosphere visible nearinfrared ir band single emissive band
|
30 |
+
abi l1b radiance product channel 7 16 approximate center wavelength 39 6185 695
|
31 |
+
734 85 961 1035 112 123 133 micron respectively digital map outgoing radiance
|
32 |
+
value top atmosphere ir band detector sample compressed packetized downlinked
|
33 |
+
ground station level 0 data conversion calibrated geolocated pixel level 1b radiance
|
34 |
+
data detector sample decompressed radiometrically corrected navigated resampled
|
35 |
+
onto invariant output grid referred abi fixed grid
|
36 |
+
- text: pipeline operator conducting risk assessment use ecological usa conjunction
|
37 |
+
pipeline information data identify area may suffer longterm permanent environmentalresource
|
38 |
+
damage event hazardous liquid pipeline accident user data encouraged read carefully
|
39 |
+
technical report cited cross reference section understand limitation ecological
|
40 |
+
usa data dataset comprises unusually sensitive area usa data ecological resource
|
41 |
+
state wyoming accordance pipeline safety law 49 usc section 60109 phmsa required
|
42 |
+
identify area unusually sensitive environmental damage event hazardous liquid
|
43 |
+
pipeline accident interaction various regulatory agency pipeline operator private
|
44 |
+
contractor nonprofit conservation organization general public process developed
|
45 |
+
adopted phmsa identify usa ecological resource process consists identifying set
|
46 |
+
candidate ecological resource using approved data source subjecting candidate
|
47 |
+
set filter criterion determine usa identification usa conducted using standardized
|
48 |
+
data processing step automated gi model resultant usa data applicable current
|
49 |
+
future regulatory requirement specified phmsa including limited pipeline integrity
|
50 |
+
management spill response planning additional information concerning ecological
|
51 |
+
usa please refer document listed cross reference section report
|
52 |
+
- text: southern ontario land resource information system solris 20
|
53 |
+
- text: toronto employment survey summary table
|
54 |
+
- text: cordon data directional traffic count
|
55 |
inference: false
|
56 |
model-index:
|
57 |
- name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2
|
|
|
65 |
split: test
|
66 |
metrics:
|
67 |
- type: accuracy
|
68 |
+
value: 0.295
|
69 |
name: Accuracy
|
70 |
- type: precision
|
71 |
+
value: 0.41697416974169743
|
72 |
name: Precision
|
73 |
- type: recall
|
74 |
+
value: 0.5044642857142857
|
75 |
name: Recall
|
76 |
- type: f1
|
77 |
+
value: 0.45656565656565656
|
78 |
name: F1
|
79 |
---
|
80 |
|
|
|
110 |
### Metrics
|
111 |
| Label | Accuracy | Precision | Recall | F1 |
|
112 |
|:--------|:---------|:----------|:-------|:-------|
|
113 |
+
| **all** | 0.295 | 0.4170 | 0.5045 | 0.4566 |
|
114 |
|
115 |
## Uses
|
116 |
|
|
|
130 |
# Download from the 🤗 Hub
|
131 |
model = SetFitModel.from_pretrained("lgd/setfit-multilabel")
|
132 |
# Run inference
|
133 |
+
preds = model("cordon data directional traffic count")
|
134 |
```
|
135 |
|
136 |
<!--
|
|
|
162 |
### Training Set Metrics
|
163 |
| Training set | Min | Median | Max |
|
164 |
|:-------------|:----|:-------|:----|
|
165 |
+
| Word count | 1 | 4.55 | 11 |
|
166 |
|
167 |
### Training Hyperparameters
|
168 |
- batch_size: (16, 16)
|
169 |
+
- num_epochs: (1, 1)
|
170 |
- max_steps: -1
|
171 |
- sampling_strategy: oversampling
|
172 |
- num_iterations: 20
|
|
|
185 |
### Training Results
|
186 |
| Epoch | Step | Training Loss | Validation Loss |
|
187 |
|:-----:|:----:|:-------------:|:---------------:|
|
188 |
+
| 0.002 | 1 | 0.3892 | - |
|
189 |
+
| 0.1 | 50 | 0.2344 | - |
|
190 |
+
| 0.2 | 100 | 0.2476 | - |
|
191 |
+
| 0.3 | 150 | 0.0538 | - |
|
192 |
+
| 0.4 | 200 | 0.0805 | - |
|
193 |
+
| 0.5 | 250 | 0.0974 | - |
|
194 |
+
| 0.6 | 300 | 0.0238 | - |
|
195 |
+
| 0.7 | 350 | 0.025 | - |
|
196 |
+
| 0.8 | 400 | 0.0497 | - |
|
197 |
+
| 0.9 | 450 | 0.0227 | - |
|
198 |
+
| 1.0 | 500 | 0.1179 | - |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
199 |
|
200 |
### Framework Versions
|
201 |
- Python: 3.10.12
|
202 |
- SetFit: 1.0.3
|
203 |
- Sentence Transformers: 3.0.1
|
204 |
- Transformers: 4.39.0
|
205 |
+
- PyTorch: 2.3.1+cu121
|
206 |
- Datasets: 2.20.0
|
207 |
- Tokenizers: 0.15.2
|
208 |
|
config_sentence_transformers.json
CHANGED
@@ -2,7 +2,7 @@
|
|
2 |
"__version__": {
|
3 |
"sentence_transformers": "3.0.1",
|
4 |
"transformers": "4.39.0",
|
5 |
-
"pytorch": "2.3.
|
6 |
},
|
7 |
"prompts": {},
|
8 |
"default_prompt_name": null,
|
|
|
2 |
"__version__": {
|
3 |
"sentence_transformers": "3.0.1",
|
4 |
"transformers": "4.39.0",
|
5 |
+
"pytorch": "2.3.1+cu121"
|
6 |
},
|
7 |
"prompts": {},
|
8 |
"default_prompt_name": null,
|
config_setfit.json
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
{
|
2 |
-
"
|
3 |
-
"
|
4 |
}
|
|
|
1 |
{
|
2 |
+
"labels": null,
|
3 |
+
"normalize_embeddings": false
|
4 |
}
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 437967672
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e41900dea02c1c3a7334b67354e5fee53c628344ac880ade9dfb61bad97cea0a
|
3 |
size 437967672
|
model_head.pkl
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 26916
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a7977aed629ae22bcdd0b7c348817a6cf8ba615d4b0464851c05e304b392410d
|
3 |
size 26916
|