End of training

README.md

@@ -20,11 +20,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [ml6team/keyphrase-extraction-distilbert-inspec](https://huggingface.co/ml6team/keyphrase-extraction-distilbert-inspec) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0761
-- Precision: 0.7378
-- Recall: 0.7970
-- F1: 0.7663
-- Accuracy: 0.7418
+- Loss: 0.3161
+- Precision: 0.9068
+- Recall: 0.9119
+- F1: 0.9094
+- Accuracy: 0.9015
 
 ## Model description
 
@@ -43,12 +43,10 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 2e-06
+- learning_rate: 0.0002
 - train_batch_size: 64
 - eval_batch_size: 4
 - seed: 42
-- gradient_accumulation_steps: 4
-- total_train_batch_size: 256
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - num_epochs: 10
@@ -58,10 +56,15 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
-| No log        | 2.35  | 10   | 1.1426          | 0.7223    | 0.7795 | 0.7498 | 0.7267   |
-| No log        | 4.71  | 20   | 1.1059          | 0.7328    | 0.7915 | 0.7610 | 0.7372   |
-| No log        | 7.06  | 30   | 1.0839          | 0.7363    | 0.7953 | 0.7647 | 0.7404   |
-| No log        | 9.41  | 40   | 1.0761          | 0.7378    | 0.7970 | 0.7663 | 0.7418   |
+| No log        | 0.59  | 10   | 0.4403          | 0.8234    | 0.8643 | 0.8433 | 0.8282   |
+| No log        | 1.18  | 20   | 0.3432          | 0.8806    | 0.8796 | 0.8801 | 0.8714   |
+| No log        | 1.76  | 30   | 0.3102          | 0.8818    | 0.9022 | 0.8919 | 0.8838   |
+| No log        | 2.35  | 40   | 0.3139          | 0.9009    | 0.8939 | 0.8974 | 0.8907   |
+| No log        | 2.94  | 50   | 0.2956          | 0.9063    | 0.8953 | 0.9008 | 0.8922   |
+| No log        | 3.53  | 60   | 0.3161          | 0.9068    | 0.9119 | 0.9094 | 0.9015   |
+| No log        | 4.12  | 70   | 0.3459          | 0.9019    | 0.9154 | 0.9086 | 0.9006   |
+| No log        | 4.71  | 80   | 0.3387          | 0.9078    | 0.9099 | 0.9089 | 0.9017   |
+| No log        | 5.29  | 90   | 0.3658          | 0.9081    | 0.8982 | 0.9032 | 0.8942   |
 
 
 ### Framework versions

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4a69b59dce1af50b79b88ca92088a691287a3d5627fbdbb441fe9f2dcfaa3046
+oid sha256:034bdc010e9653203930571b7f9e78a576d1344170ee359f8a4d9ff58a6c3012
 size 265476168

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e023469f54131d93932aa39d88e17806cef0a18695219ce13a3ce6fddc48e4d3
+oid sha256:11e4ec6176b240c4c2c6a90105bc9183cfaa56661c346a9115e753cb5dfb2b2d
 size 4984