Ahmet
commited on
Commit
•
7c4e56b
1
Parent(s):
27a1a08
update model
Browse files- README.md +13 -7
- config.json +1 -1
- pytorch_model.bin +2 -2
README.md
CHANGED
|
@@ -21,6 +21,12 @@ This model was adapted from [ytu-ce-cosmos/turkish-medium-bert-uncased](https://
|
|
| 21 |
- [nli_tr](https://huggingface.co/datasets/nli_tr)
|
| 22 |
- [emrecan/stsb-mt-turkish](https://huggingface.co/datasets/emrecan/stsb-mt-turkish)
|
| 23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
## Usage (Sentence-Transformers)
|
| 25 |
|
| 26 |
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
|
|
@@ -85,10 +91,10 @@ print(sentence_embeddings)
|
|
| 85 |
Achieved results on the [STS-b](https://huggingface.co/datasets/emrecan/stsb-mt-turkish) test split are given below:
|
| 86 |
|
| 87 |
```txt
|
| 88 |
-
Cosine-Similarity :
|
| 89 |
-
Manhattan-Distance:
|
| 90 |
-
Euclidean-Distance:
|
| 91 |
-
Dot-Product-Similarity:
|
| 92 |
```
|
| 93 |
|
| 94 |
|
|
@@ -109,8 +115,8 @@ The model was trained with the parameters:
|
|
| 109 |
Parameters of the fit()-Method:
|
| 110 |
```
|
| 111 |
{
|
| 112 |
-
"epochs":
|
| 113 |
-
"evaluation_steps":
|
| 114 |
"evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
|
| 115 |
"max_grad_norm": 1,
|
| 116 |
"optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
|
|
@@ -119,7 +125,7 @@ Parameters of the fit()-Method:
|
|
| 119 |
},
|
| 120 |
"scheduler": "WarmupLinear",
|
| 121 |
"steps_per_epoch": null,
|
| 122 |
-
"warmup_steps":
|
| 123 |
"weight_decay": 0.01
|
| 124 |
}
|
| 125 |
```
|
|
|
|
| 21 |
- [nli_tr](https://huggingface.co/datasets/nli_tr)
|
| 22 |
- [emrecan/stsb-mt-turkish](https://huggingface.co/datasets/emrecan/stsb-mt-turkish)
|
| 23 |
|
| 24 |
+
:warning: **All texts were manually lowercased,** [as stated](https://huggingface.co/ytu-ce-cosmos/turkish-medium-bert-uncased#%E2%9A%A0-uncased-use-requires-manual-lowercase-conversion) by the model's authors:
|
| 25 |
+
|
| 26 |
+
```python
|
| 27 |
+
text.replace("I", "ı").lower()
|
| 28 |
+
```
|
| 29 |
+
|
| 30 |
## Usage (Sentence-Transformers)
|
| 31 |
|
| 32 |
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
|
|
|
|
| 91 |
Achieved results on the [STS-b](https://huggingface.co/datasets/emrecan/stsb-mt-turkish) test split are given below:
|
| 92 |
|
| 93 |
```txt
|
| 94 |
+
Cosine-Similarity : Pearson: 0.8329 Spearman: 0.8336
|
| 95 |
+
Manhattan-Distance: Pearson: 0.8193 Spearman: 0.8188
|
| 96 |
+
Euclidean-Distance: Pearson: 0.8198 Spearman: 0.8195
|
| 97 |
+
Dot-Product-Similarity: Pearson: 0.7888 Spearman: 0.7822
|
| 98 |
```
|
| 99 |
|
| 100 |
|
|
|
|
| 115 |
Parameters of the fit()-Method:
|
| 116 |
```
|
| 117 |
{
|
| 118 |
+
"epochs": 4,
|
| 119 |
+
"evaluation_steps": 9,
|
| 120 |
"evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
|
| 121 |
"max_grad_norm": 1,
|
| 122 |
"optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
|
|
|
|
| 125 |
},
|
| 126 |
"scheduler": "WarmupLinear",
|
| 127 |
"steps_per_epoch": null,
|
| 128 |
+
"warmup_steps": 36,
|
| 129 |
"weight_decay": 0.01
|
| 130 |
}
|
| 131 |
```
|
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"architectures": [
|
| 4 |
"BertModel"
|
| 5 |
],
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "ytu_ce_cosmos-turkish_medium_bert_uncased-b64-e4-nli\\",
|
| 3 |
"architectures": [
|
| 4 |
"BertModel"
|
| 5 |
],
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb088867528093ee52e3f2c60b690dae20696531b7d55e744038bc7c78ebf565
|
| 3 |
+
size 168569197
|