CocoRoF commited on
Commit
e22a6d1
·
verified ·
1 Parent(s): b1fe9af
Files changed (1) hide show
  1. README.md +90 -0
README.md ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: apache-2.0
4
+ base_model: x2bee/KoModernBERT-base-mlm-v03-ckp00
5
+ tags:
6
+ - generated_from_trainer
7
+ model-index:
8
+ - name: KMB_SimCSE_test
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # KMB_SimCSE_test
16
+
17
+ This model is a fine-tuned version of [x2bee/KoModernBERT-base-mlm-v03-ckp00](https://huggingface.co/x2bee/KoModernBERT-base-mlm-v03-ckp00) on an unknown dataset.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 0.0466
20
+ - Pearson Cosine: 0.8138
21
+ - Spearman Cosine: 0.8155
22
+ - Pearson Manhattan: 0.8156
23
+ - Spearman Manhattan: 0.8244
24
+ - Pearson Euclidean: 0.8159
25
+ - Spearman Euclidean: 0.8248
26
+ - Pearson Dot: 0.7831
27
+ - Spearman Dot: 0.7815
28
+
29
+ ## Model description
30
+
31
+ More information needed
32
+
33
+ ## Intended uses & limitations
34
+
35
+ More information needed
36
+
37
+ ## Training and evaluation data
38
+
39
+ More information needed
40
+
41
+ ## Training procedure
42
+
43
+ ### Training hyperparameters
44
+
45
+ The following hyperparameters were used during training:
46
+ - learning_rate: 0.0001
47
+ - train_batch_size: 16
48
+ - eval_batch_size: 16
49
+ - seed: 42
50
+ - distributed_type: multi-GPU
51
+ - gradient_accumulation_steps: 8
52
+ - total_train_batch_size: 128
53
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
54
+ - lr_scheduler_type: linear
55
+ - lr_scheduler_warmup_ratio: 0.1
56
+ - num_epochs: 1.0
57
+
58
+ ### Training results
59
+
60
+ | Training Loss | Epoch | Step | Validation Loss | Pearson Cosine | Spearman Cosine | Pearson Manhattan | Spearman Manhattan | Pearson Euclidean | Spearman Euclidean | Pearson Dot | Spearman Dot |
61
+ |:-------------:|:------:|:----:|:---------------:|:--------------:|:---------------:|:-----------------:|:------------------:|:-----------------:|:------------------:|:-----------:|:------------:|
62
+ | 0.6155 | 0.0469 | 100 | 0.0872 | 0.7869 | 0.7901 | 0.7884 | 0.7903 | 0.7895 | 0.7914 | 0.7575 | 0.7545 |
63
+ | 0.4415 | 0.0937 | 200 | 0.0599 | 0.8037 | 0.8063 | 0.8064 | 0.8128 | 0.8065 | 0.8128 | 0.7711 | 0.7711 |
64
+ | 0.4497 | 0.1406 | 300 | 0.0626 | 0.8158 | 0.8165 | 0.8107 | 0.8179 | 0.8109 | 0.8181 | 0.7944 | 0.7933 |
65
+ | 0.4479 | 0.1874 | 400 | 0.0646 | 0.8130 | 0.8152 | 0.8134 | 0.8209 | 0.8133 | 0.8209 | 0.7848 | 0.7852 |
66
+ | 0.4284 | 0.2343 | 500 | 0.0651 | 0.8051 | 0.8089 | 0.8066 | 0.8134 | 0.8067 | 0.8138 | 0.7820 | 0.7820 |
67
+ | 0.3853 | 0.2812 | 600 | 0.0571 | 0.8038 | 0.8084 | 0.8036 | 0.8107 | 0.8034 | 0.8104 | 0.7922 | 0.7949 |
68
+ | 0.3798 | 0.3280 | 700 | 0.0627 | 0.7998 | 0.8041 | 0.7968 | 0.8043 | 0.7963 | 0.8040 | 0.7824 | 0.7847 |
69
+ | 0.3615 | 0.3749 | 800 | 0.0563 | 0.8038 | 0.8047 | 0.8074 | 0.8113 | 0.8070 | 0.8107 | 0.7799 | 0.7773 |
70
+ | 0.3333 | 0.4217 | 900 | 0.0584 | 0.8054 | 0.8067 | 0.8051 | 0.8101 | 0.8051 | 0.8098 | 0.7824 | 0.7821 |
71
+ | 0.3693 | 0.4686 | 1000 | 0.0583 | 0.7878 | 0.7864 | 0.7686 | 0.7809 | 0.7679 | 0.7804 | 0.7796 | 0.7773 |
72
+ | 0.3623 | 0.5155 | 1100 | 0.0574 | 0.8054 | 0.8090 | 0.8054 | 0.8116 | 0.8054 | 0.8116 | 0.7882 | 0.7907 |
73
+ | 0.3795 | 0.5623 | 1200 | 0.0592 | 0.8028 | 0.8065 | 0.8075 | 0.8133 | 0.8069 | 0.8122 | 0.7769 | 0.7771 |
74
+ | 0.3053 | 0.6092 | 1300 | 0.0460 | 0.8208 | 0.8220 | 0.8163 | 0.8238 | 0.8166 | 0.8240 | 0.8059 | 0.8056 |
75
+ | 0.3254 | 0.6560 | 1400 | 0.0519 | 0.8104 | 0.8132 | 0.8115 | 0.8189 | 0.8119 | 0.8194 | 0.7882 | 0.7892 |
76
+ | 0.3399 | 0.7029 | 1500 | 0.0482 | 0.8192 | 0.8200 | 0.8162 | 0.8242 | 0.8162 | 0.8238 | 0.8010 | 0.8003 |
77
+ | 0.3418 | 0.7498 | 1600 | 0.0507 | 0.8195 | 0.8227 | 0.8185 | 0.8248 | 0.8191 | 0.8252 | 0.7973 | 0.7978 |
78
+ | 0.329 | 0.7966 | 1700 | 0.0490 | 0.8062 | 0.8075 | 0.8066 | 0.8125 | 0.8070 | 0.8131 | 0.7809 | 0.7785 |
79
+ | 0.2774 | 0.8435 | 1800 | 0.0436 | 0.8131 | 0.8136 | 0.8124 | 0.8212 | 0.8124 | 0.8213 | 0.7812 | 0.7793 |
80
+ | 0.3113 | 0.8903 | 1900 | 0.0503 | 0.8192 | 0.8219 | 0.8187 | 0.8256 | 0.8191 | 0.8260 | 0.7944 | 0.7945 |
81
+ | 0.3495 | 0.9372 | 2000 | 0.0519 | 0.8079 | 0.8110 | 0.8123 | 0.8184 | 0.8121 | 0.8183 | 0.7810 | 0.7810 |
82
+ | 0.3015 | 0.9841 | 2100 | 0.0466 | 0.8138 | 0.8155 | 0.8156 | 0.8244 | 0.8159 | 0.8248 | 0.7831 | 0.7815 |
83
+
84
+
85
+ ### Framework versions
86
+
87
+ - Transformers 4.48.0.dev0
88
+ - Pytorch 2.5.1+cu121
89
+ - Datasets 3.1.0
90
+ - Tokenizers 0.21.0