mp-02
/

layoutlmv3-finetuned-funsd

Generated from Trainer

Model card Files Files and versions Metrics Training metrics Community

mp-02 commited on Aug 24

Commit

8142819

•

1 Parent(s): d3d7e7a

update model card README.md

Files changed (1) hide show

README.md +12 -16

README.md CHANGED Viewed

@@ -19,11 +19,11 @@ should probably proofread and complete it, then remove this comment. -->
 This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6416
-- Precision: 0.8820
-- Recall: 0.908
-- F1: 0.8948
-- Accuracy: 0.8486
 ## Model description
@@ -48,22 +48,18 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- training_steps: 500
 ### Training results
 | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
-| No log        | 3.33  | 50   | 0.7810          | 0.7547    | 0.8155 | 0.7839 | 0.7633   |
-| No log        | 6.67  | 100  | 0.5576          | 0.8015    | 0.8805 | 0.8392 | 0.8060   |
-| No log        | 10.0  | 150  | 0.5810          | 0.8452    | 0.887  | 0.8656 | 0.8223   |
-| No log        | 13.33 | 200  | 0.5634          | 0.8498    | 0.8965 | 0.8725 | 0.8393   |
-| No log        | 16.67 | 250  | 0.5419          | 0.8814    | 0.907  | 0.8940 | 0.8529   |
-| No log        | 20.0  | 300  | 0.5817          | 0.8760    | 0.9005 | 0.8881 | 0.8465   |
-| No log        | 23.33 | 350  | 0.6015          | 0.8744    | 0.9085 | 0.8911 | 0.8429   |
-| No log        | 26.67 | 400  | 0.5982          | 0.8830    | 0.917  | 0.8997 | 0.8536   |
-| No log        | 30.0  | 450  | 0.6316          | 0.8832    | 0.907  | 0.8949 | 0.8493   |
-| 0.2944        | 33.33 | 500  | 0.6416          | 0.8820    | 0.908  | 0.8948 | 0.8486   |
 ### Framework versions

 This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the None dataset.
 It achieves the following results on the evaluation set:
+- Loss: 0.5930
+- Precision: 0.7981
+- Recall: 0.8675
+- F1: 0.8313
+- Accuracy: 0.8104
 ## Model description
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
+- training_steps: 150
 ### Training results
 | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
+| No log        | 1.67  | 25   | 1.2209          | 0.4642    | 0.5155 | 0.4885 | 0.6559   |
+| No log        | 3.33  | 50   | 0.8172          | 0.7324    | 0.776  | 0.7536 | 0.7619   |
+| No log        | 5.0   | 75   | 0.6125          | 0.7876    | 0.8435 | 0.8146 | 0.8126   |
+| No log        | 6.67  | 100  | 0.5984          | 0.8053    | 0.8665 | 0.8348 | 0.8107   |
+| No log        | 8.33  | 125  | 0.5674          | 0.8040    | 0.8715 | 0.8364 | 0.8217   |
+| No log        | 10.0  | 150  | 0.5930          | 0.7981    | 0.8675 | 0.8313 | 0.8104   |
 ### Framework versions