End of training

README.md

@@ -20,30 +20,80 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/layoutlm-base-uncased](https://huggingface.co/microsoft/layoutlm-base-uncased) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 3.8093
-- F1: 0.032
-- Recall: 0.04
-- Precision: 0.0267
-- Pred Var-ausf. 4: 1
-- Pred Fadenfarbnr.: 3
-- Pred Menge3: 1
-- Pred Zusatz 1: 7
-- Pred Menge7: 2
-- Pred Holz 2: 2
-- Percentage Pred Act Holz 2: 0.6667
-- Pred Gestelltnr.: 5
-- Pred Modell 1: 3
-- Percentage Pred Act Modell 1: 1.0
-- Pred Menge6: 1
-- Act Bestellnummer: 3
-- Act Kundennr.: 5
-- Act Modell 2: 5
-- Act Bezug 2: 1
-- Act Menge2: 3
-- Act Holz 2: 3
-- Act Möbelhaus: 1
-- Act Holz 1: 1
-- Act Modell 1: 3
+- Loss: 0.9964
+- F1: 0.7547
+- Recall: 0.7148
+- Precision: 0.7992
+- Pred Kommission: 44
+- Percentage Pred Act Kommission: 1.0476
+- Pred Bestellnummer: 146
+- Percentage Pred Act Bestellnummer: 1.0069
+- Pred Möbelhaus: 69
+- Percentage Pred Act Möbelhaus: 1.0781
+- Pred Var-ausf 1: 7
+- Percentage Pred Act Var-ausf 1: 0.4375
+- Pred Modell 2: 46
+- Percentage Pred Act Modell 2: 0.8679
+- Pred Modell 1: 126
+- Percentage Pred Act Modell 1: 1.0413
+- Pred Menge1: 45
+- Percentage Pred Act Menge1: 1.5
+- Pred Kundennr.: 73
+- Percentage Pred Act Kundennr.: 1.0735
+- Pred Bezug 3: 9
+- Percentage Pred Act Bezug 3: 1.8
+- Pred Menge2: 4
+- Percentage Pred Act Menge2: 0.2222
+- Pred Modell 3: 70
+- Percentage Pred Act Modell 3: 1.4583
+- Pred Bezug 1: 20
+- Percentage Pred Act Bezug 1: 1.0526
+- Pred Termin kundenwunsch - kw: 23
+- Percentage Pred Act Termin kundenwunsch - kw: 0.8846
+- Pred Holz 1: 10
+- Percentage Pred Act Holz 1: 0.5556
+- Pred Holz 2: 42
+- Percentage Pred Act Holz 2: 2.0
+- Pred Zusatz 1: 12
+- Percentage Pred Act Zusatz 1: 2.4
+- Pred Menge3: 22
+- Percentage Pred Act Menge3: 2.0
+- Pred Bezug 2: 5
+- Percentage Pred Act Bezug 2: 0.625
+- Pred La-anschrift: 2
+- Percentage Pred Act La-anschrift: 2.0
+- Act Kommission: 42
+- Act Bestellnummer: 145
+- Act Möbelhaus: 64
+- Act Var-ausf. 2: 7
+- Act Modell 1: 121
+- Act Modell 2: 53
+- Act Menge2: 18
+- Act Kundennr.: 68
+- Act Var-ausf 1: 16
+- Act Bezug 2: 8
+- Act Menge1: 30
+- Act Bezug 1: 19
+- Act Termin kundenwunsch - kw: 26
+- Act Holz 1: 18
+- Act Holz 2: 21
+- Act Menge3: 11
+- Act Modell 3: 48
+- Act Holz 3: 9
+- Act Zusatz 3: 2
+- Act Modell 4: 7
+- Act Zusatz 2: 8
+- Act Bezug 3: 5
+- Act Zusatz 1: 5
+- Act Menge4: 7
+- Act Var-ausf. 3: 4
+- Act Var-ausf. 4: 4
+- Act Modell 5: 1
+- Act Gestelltnr.: 1
+- Act Bezug 4: 4
+- Act La-anschrift: 1
+- Act Holz 4: 1
+- Act Menge5: 1
 
 ## Model description
 
@@ -68,7 +118,7 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
-- num_epochs: 1
+- num_epochs: 5
 - mixed_precision_training: Native AMP
 
 ### Training results

logs/events.out.tfevents.1749058232.phi-ThinkPad-T490

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a96fc5a35749bfe0240b4df4cfe25bcf8f7044fb12a844ba59a8d12950d9ecb
+size 4713