Hugging Face's logo

thejosango
/
nuha-binary

Text Classification
Transformers
PyTorch
Arabic
bert
hate-speech
gender-based-violence
arabic
binary-classification
pilot
Eval Results (legacy)
text-embeddings-inference
Model card Files
xet
 Community
yazansh commited on
Commit
2dfdcee
·
verified ·
1 Parent(s): f54519f

binary-53

Browse files
Files changed (4) hide show
  1. README.md +17 -16
  2. config.toml +3 -3
  3. pytorch_model.bin +1 -1
  4. training_args.bin +1 -1
README.md CHANGED
@@ -24,13 +24,13 @@ model-index:
24
  metrics:
25
  - name: F1
26
  type: f1
27
- value: 0.6902334758303189
28
  - name: Precision
29
  type: precision
30
- value: 0.6218338023996445
31
  - name: Recall
32
  type: recall
33
- value: 0.7755403657860706
34
  ---
35
 
36
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -40,10 +40,10 @@ should probably proofread and complete it, then remove this comment. -->
40
 
41
  This model is a fine-tuned version of [thejosango/nuha-mlm](https://huggingface.co/thejosango/nuha-mlm) on the nuha-dataset dataset.
42
  It achieves the following results on the evaluation set:
43
- - Loss: 0.5773
44
- - F1: 0.6902
45
- - Precision: 0.6218
46
- - Recall: 0.7755
47
  - Support: None
48
 
49
  ## Model description
@@ -68,7 +68,7 @@ The following hyperparameters were used during training:
68
  - eval_batch_size: 64
69
  - seed: 42
70
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
71
- - lr_scheduler_type: constant
72
  - num_epochs: 5
73
  - label_smoothing_factor: 0.1
74
 
@@ -76,14 +76,15 @@ The following hyperparameters were used during training:
76
 
77
  | Training Loss | Epoch | Step | Validation Loss | F1 | Precision | Recall | Support |
78
  |:-------------:|:-----:|:----:|:---------------:|:------:|:---------:|:------:|:-------:|
79
- | 0.5531 | 0.37 | 500 | 0.5324 | 0.6866 | 0.6190 | 0.7707 | None |
80
- | 0.5346 | 0.74 | 1000 | 0.5675 | 0.6920 | 0.5862 | 0.8444 | None |
81
- | 0.5142 | 1.11 | 1500 | 0.5444 | 0.6980 | 0.6032 | 0.8280 | None |
82
- | 0.472 | 1.48 | 2000 | 0.5466 | 0.6993 | 0.6165 | 0.8079 | None |
83
- | 0.4704 | 1.85 | 2500 | 0.5346 | 0.7020 | 0.6274 | 0.7968 | None |
84
- | 0.3858 | 2.22 | 3000 | 0.5762 | 0.6989 | 0.6292 | 0.7861 | None |
85
- | 0.3912 | 2.58 | 3500 | 0.5676 | 0.6929 | 0.6372 | 0.7593 | None |
86
- | 0.3966 | 2.95 | 4000 | 0.5773 | 0.6902 | 0.6218 | 0.7755 | None |
 
87
 
88
 
89
  ### Framework versions
 
24
  metrics:
25
  - name: F1
26
  type: f1
27
+ value: 0.6926479178836685
28
  - name: Precision
29
  type: precision
30
+ value: 0.6316590563165906
31
  - name: Recall
32
  type: recall
33
+ value: 0.7666728246813227
34
  ---
35
 
36
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
40
 
41
  This model is a fine-tuned version of [thejosango/nuha-mlm](https://huggingface.co/thejosango/nuha-mlm) on the nuha-dataset dataset.
42
  It achieves the following results on the evaluation set:
43
+ - Loss: 0.6119
44
+ - F1: 0.6926
45
+ - Precision: 0.6317
46
+ - Recall: 0.7667
47
  - Support: None
48
 
49
  ## Model description
 
68
  - eval_batch_size: 64
69
  - seed: 42
70
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
71
+ - lr_scheduler_type: linear
72
  - num_epochs: 5
73
  - label_smoothing_factor: 0.1
74
 
 
76
 
77
  | Training Loss | Epoch | Step | Validation Loss | F1 | Precision | Recall | Support |
78
  |:-------------:|:-----:|:----:|:---------------:|:------:|:---------:|:------:|:-------:|
79
+ | 0.553 | 0.37 | 500 | 0.5378 | 0.6847 | 0.6064 | 0.7863 | None |
80
+ | 0.5345 | 0.74 | 1000 | 0.5667 | 0.6915 | 0.5856 | 0.8443 | None |
81
+ | 0.5144 | 1.11 | 1500 | 0.5550 | 0.6968 | 0.5950 | 0.8406 | None |
82
+ | 0.4725 | 1.48 | 2000 | 0.5309 | 0.7020 | 0.6443 | 0.7711 | None |
83
+ | 0.468 | 1.85 | 2500 | 0.5317 | 0.7020 | 0.6252 | 0.8003 | None |
84
+ | 0.3986 | 2.22 | 3000 | 0.5538 | 0.6995 | 0.6443 | 0.7650 | None |
85
+ | 0.3967 | 2.58 | 3500 | 0.5581 | 0.6948 | 0.6466 | 0.7508 | None |
86
+ | 0.3971 | 2.95 | 4000 | 0.5643 | 0.6975 | 0.6290 | 0.7826 | None |
87
+ | 0.342 | 3.32 | 4500 | 0.6119 | 0.6926 | 0.6317 | 0.7667 | None |
88
 
89
 
90
  ### Framework versions
config.toml CHANGED
@@ -1,5 +1,5 @@
1
  [experiment]
2
- name = "binary-52"
3
  type = "binary"
4
 
5
 
@@ -19,13 +19,13 @@ classifier_dropout = 0.25
19
  [training]
20
  num_train_epochs = 5
21
  warmup_steps = 0
22
- lr_scheduler_type = "constant"
23
  learning_rate = 5e-5
24
  per_device_train_batch_size = 64
25
  per_device_eval_batch_size = 64
26
  gradient_accumulation_steps = 1
27
  weight_decay = 1e-3
28
  label_smoothing_factor = 0.1
29
- weighted_loss = false
30
  early_stopping_patience = 5
31
  early_stopping_threshold = 0.005
 
1
  [experiment]
2
+ name = "binary-53"
3
  type = "binary"
4
 
5
 
 
19
  [training]
20
  num_train_epochs = 5
21
  warmup_steps = 0
22
+ lr_scheduler_type = "linear"
23
  learning_rate = 5e-5
24
  per_device_train_batch_size = 64
25
  per_device_eval_batch_size = 64
26
  gradient_accumulation_steps = 1
27
  weight_decay = 1e-3
28
  label_smoothing_factor = 0.1
29
+ weighted_loss = true
30
  early_stopping_patience = 5
31
  early_stopping_threshold = 0.005
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:77854de095531e646b72ad7d6a9d3323b85a5652cee4d8a587558eaf19f4b5f4
3
  size 370706033
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a6c22f6383637027556d78ffec12f58f1d678a4d7f46de11664073c36d1d86d
3
  size 370706033
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:21f5bef3aca0c3c13753416b32c9243167e3d497cac8c70cb0b7b3354d422cb3
3
  size 4091
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c05259d62d10144d73bd28aed57f9da0ce390e5520f5caf1e30a66e33baa68d2
3
  size 4091