Add SetFit model

README.md

@@ -13,39 +13,45 @@ tags:
 - text-classification
 - generated_from_setfit_trainer
 widget:
-- text: planning development plan environment natural environment land use development
-    plan put forward frame reference aimed better knowing protecting promoting heritage
-    data available set mainly come mapping section 21 23 31 urban urban plan montreal
-    namely adaptation climate change territory ecological interest territory ecological
-    interest green blue frame green blue frame well constraint nuisance urban planning
-    development plan agglomeration montreal outline main parameter guide montreal
-    agglomeration council decision relating land use planning coming year perspective
-    sustainable development document guide decision shape territory order promote
-    compact greener neighborhood increase public active transportation support economic
-    dynamism agglomeration highlight area interest consult interactive map httpssmvtmapsarcgiscomappswebappviewerindexhtmlidd152aaa85b6f4e9086cecdf10c7456db
-    planning development plan visualize thematic datathis third party metadata element
-    translated using automated translation tool amazon translate formdescriptors natureandenvironment
-    scienceandtechnology wood forest corridor green space falaise pente floodplain
-    development planning diagram urbanism heat island government information
-- text: senior survey 2017
-- text: list permit exemption force law responsibility agency following merchant must
-    licensed agency operate travel agent debt collector itinerant merchant solicit
-    consumer order make sale make sale elsewhere business established ie doortodoor
-    kiosk street mall etc retailer additional guarantee relating car motorcycle adapted
-    transport public road operator health studio fitness center weight loss center
-    example road vehicle dealer road vehicle recyclers retailer enter highcost credit
-    contract retailer conclude highcost credit contract debt settlement service merchant
-    negotiate consumer creditor receive amount distribute lender silver obligation
-    trader must comply allow office ensure compliance legislative provision area activity
-    risk considered significant license category linked financial protection mechanism
-    consumer mechanism allow consumer compensated certain situation merchant valid
-    license received authorization president office carry activity renewed permit
-    scheduled date applicable certain category trader obtain exemption submitting
-    bond effect exempting legal obligation particular depositing trust account money
-    collected good whose delivery scheduled delivered two month purchase governmentandpolitics
-    law retailer deliverance exemption permit consumer protection
-- text: ambient groundwater geochemistry data southwestern ontario
-- text: neighbourhood
+- text: weather satellite imagery update every 10 minute cloud top temperature colorized
+    reveal area intensity lower level transparent satellite imagery combine data noaa
+    go east west satellite jma himawari satellite providing full coverage weather
+    event world west coast africa west east coast india tile service update recent
+    image every 10 minute 15 km per pixel resolution infrared ir band detects radiation
+    emitted earth???s surface atmosphere cloud ??·infrared window??? portion spectrum
+    radiation wavelength near 103 micrometer term ??·window??? mean pass atmosphere
+    relatively little absorption gas water vapor useful estimating emitting temperature
+    earth???s surface cloud top major advantage ir band sense energy night imagery
+    available 24 hour day advanced baseline imager abi instrument sample radiance
+    earth sixteen spectral band using several array detector instrument???s focal
+    plane single reflective band abi level 1b radiance product channel 1 6 approximate
+    center wavelength 047 064 0865 1378 161 225 micron respectively digital map outgoing
+    radiance value top atmosphere visible nearinfrared ir band single emissive band
+    abi l1b radiance product channel 7 16 approximate center wavelength 39 6185 695
+    734 85 961 1035 112 123 133 micron respectively digital map outgoing radiance
+    value top atmosphere ir band detector sample compressed packetized downlinked
+    ground station level 0 data conversion calibrated geolocated pixel level 1b radiance
+    data detector sample decompressed radiometrically corrected navigated resampled
+    onto invariant output grid referred abi fixed grid
+- text: pipeline operator conducting risk assessment use ecological usa conjunction
+    pipeline information data identify area may suffer longterm permanent environmentalresource
+    damage event hazardous liquid pipeline accident user data encouraged read carefully
+    technical report cited cross reference section understand limitation ecological
+    usa data dataset comprises unusually sensitive area usa data ecological resource
+    state wyoming accordance pipeline safety law 49 usc section 60109 phmsa required
+    identify area unusually sensitive environmental damage event hazardous liquid
+    pipeline accident interaction various regulatory agency pipeline operator private
+    contractor nonprofit conservation organization general public process developed
+    adopted phmsa identify usa ecological resource process consists identifying set
+    candidate ecological resource using approved data source subjecting candidate
+    set filter criterion determine usa identification usa conducted using standardized
+    data processing step automated gi model resultant usa data applicable current
+    future regulatory requirement specified phmsa including limited pipeline integrity
+    management spill response planning additional information concerning ecological
+    usa please refer document listed cross reference section report
+- text: southern ontario land resource information system solris 20
+- text: toronto employment survey summary table
+- text: cordon data directional traffic count
 inference: false
 model-index:
 - name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2
@@ -59,16 +65,16 @@ model-index:
       split: test
     metrics:
     - type: accuracy
-      value: 0.21
+      value: 0.295
       name: Accuracy
     - type: precision
-      value: 0.4350282485875706
+      value: 0.41697416974169743
       name: Precision
     - type: recall
-      value: 0.652542372881356
+      value: 0.5044642857142857
       name: Recall
     - type: f1
-      value: 0.5220338983050848
+      value: 0.45656565656565656
       name: F1
 ---
 
@@ -104,7 +110,7 @@ The model has been trained using an efficient few-shot learning technique that i
 ### Metrics
 | Label   | Accuracy | Precision | Recall | F1     |
 |:--------|:---------|:----------|:-------|:-------|
-| **all** | 0.21     | 0.4350    | 0.6525 | 0.5220 |
+| **all** | 0.295    | 0.4170    | 0.5045 | 0.4566 |
 
 ## Uses
 
@@ -124,7 +130,7 @@ from setfit import SetFitModel
 # Download from the 🤗 Hub
 model = SetFitModel.from_pretrained("lgd/setfit-multilabel")
 # Run inference
-preds = model("neighbourhood")
+preds = model("cordon data directional traffic count")
 ```
 
 <!--
@@ -156,11 +162,11 @@ preds = model("neighbourhood")
 ### Training Set Metrics
 | Training set | Min | Median | Max |
 |:-------------|:----|:-------|:----|
-| Word count   | 1   | 4.35   | 11  |
+| Word count   | 1   | 4.55   | 11  |
 
 ### Training Hyperparameters
 - batch_size: (16, 16)
-- num_epochs: (3, 3)
+- num_epochs: (1, 1)
 - max_steps: -1
 - sampling_strategy: oversampling
 - num_iterations: 20
@@ -179,52 +185,24 @@ preds = model("neighbourhood")
 ### Training Results
 | Epoch | Step | Training Loss | Validation Loss |
 |:-----:|:----:|:-------------:|:---------------:|
-| 0.004 | 1    | 0.3342        | -               |
-| 0.2   | 50   | 0.1221        | -               |
-| 0.4   | 100  | 0.0837        | -               |
-| 0.6   | 150  | 0.0403        | -               |
-| 0.8   | 200  | 0.0798        | -               |
-| 1.0   | 250  | 0.0282        | -               |
-| 0.004 | 1    | 0.0266        | -               |
-| 0.2   | 50   | 0.0102        | -               |
-| 0.4   | 100  | 0.0501        | -               |
-| 0.6   | 150  | 0.0297        | -               |
-| 0.8   | 200  | 0.066         | -               |
-| 1.0   | 250  | 0.0302        | -               |
-| 0.004 | 1    | 0.0151        | -               |
-| 0.2   | 50   | 0.0232        | -               |
-| 0.4   | 100  | 0.017         | -               |
-| 0.6   | 150  | 0.0133        | -               |
-| 0.8   | 200  | 0.0629        | -               |
-| 1.0   | 250  | 0.0349        | -               |
-| 1.2   | 300  | 0.0585        | -               |
-| 1.4   | 350  | 0.0658        | -               |
-| 1.6   | 400  | 0.0446        | -               |
-| 1.8   | 450  | 0.0073        | -               |
-| 2.0   | 500  | 0.0326        | -               |
-| 0.004 | 1    | 0.017         | -               |
-| 0.2   | 50   | 0.0038        | -               |
-| 0.4   | 100  | 0.0095        | -               |
-| 0.6   | 150  | 0.0154        | -               |
-| 0.8   | 200  | 0.0444        | -               |
-| 1.0   | 250  | 0.0221        | -               |
-| 1.2   | 300  | 0.0362        | -               |
-| 1.4   | 350  | 0.0565        | -               |
-| 1.6   | 400  | 0.0338        | -               |
-| 1.8   | 450  | 0.0081        | -               |
-| 2.0   | 500  | 0.0299        | -               |
-| 2.2   | 550  | 0.106         | -               |
-| 2.4   | 600  | 0.0191        | -               |
-| 2.6   | 650  | 0.0104        | -               |
-| 2.8   | 700  | 0.0369        | -               |
-| 3.0   | 750  | 0.024         | -               |
+| 0.002 | 1    | 0.3892        | -               |
+| 0.1   | 50   | 0.2344        | -               |
+| 0.2   | 100  | 0.2476        | -               |
+| 0.3   | 150  | 0.0538        | -               |
+| 0.4   | 200  | 0.0805        | -               |
+| 0.5   | 250  | 0.0974        | -               |
+| 0.6   | 300  | 0.0238        | -               |
+| 0.7   | 350  | 0.025         | -               |
+| 0.8   | 400  | 0.0497        | -               |
+| 0.9   | 450  | 0.0227        | -               |
+| 1.0   | 500  | 0.1179        | -               |
 
 ### Framework Versions
 - Python: 3.10.12
 - SetFit: 1.0.3
 - Sentence Transformers: 3.0.1
 - Transformers: 4.39.0
-- PyTorch: 2.3.0+cu121
+- PyTorch: 2.3.1+cu121
 - Datasets: 2.20.0
 - Tokenizers: 0.15.2
 

config_sentence_transformers.json

@@ -2,7 +2,7 @@
   "__version__": {
     "sentence_transformers": "3.0.1",
     "transformers": "4.39.0",
-    "pytorch": "2.3.0+cu121"
+    "pytorch": "2.3.1+cu121"
   },
   "prompts": {},
   "default_prompt_name": null,

config_setfit.json

@@ -1,4 +1,4 @@
 {
-  "normalize_embeddings": false,
-  "labels": null
+  "labels": null,
+  "normalize_embeddings": false
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b4b2e1409e85db112f45aebf4ee858201e09354898f68fd2b35a468a74d97a61
+oid sha256:e41900dea02c1c3a7334b67354e5fee53c628344ac880ade9dfb61bad97cea0a
 size 437967672

model_head.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b8dfba481443ce98ee926990619e2c7d0a8d3d380c0ecb23b758a28716945513
+oid sha256:a7977aed629ae22bcdd0b7c348817a6cf8ba615d4b0464851c05e304b392410d
 size 26916