bertoslav-limited-ner
Browse files- README.md +95 -0
- config.json +42 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
README.md
ADDED
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- generated_from_trainer
|
4 |
+
datasets:
|
5 |
+
- wikiann
|
6 |
+
metrics:
|
7 |
+
- precision
|
8 |
+
- recall
|
9 |
+
- f1
|
10 |
+
- accuracy
|
11 |
+
model-index:
|
12 |
+
- name: bertoslav
|
13 |
+
results:
|
14 |
+
- task:
|
15 |
+
name: Token Classification
|
16 |
+
type: token-classification
|
17 |
+
dataset:
|
18 |
+
name: wikiann sk
|
19 |
+
type: wikiann
|
20 |
+
args: sk
|
21 |
+
metrics:
|
22 |
+
- name: Precision
|
23 |
+
type: precision
|
24 |
+
value: 0.8985571260306242
|
25 |
+
- name: Recall
|
26 |
+
type: recall
|
27 |
+
value: 0.9173994738819993
|
28 |
+
- name: F1
|
29 |
+
type: f1
|
30 |
+
value: 0.9078805459481573
|
31 |
+
- name: Accuracy
|
32 |
+
type: accuracy
|
33 |
+
value: 0.9700235061239639
|
34 |
+
---
|
35 |
+
|
36 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
37 |
+
should probably proofread and complete it, then remove this comment. -->
|
38 |
+
|
39 |
+
# bertoslav
|
40 |
+
|
41 |
+
This model is a fine-tuned version of [crabz/bertoslav-limited](https://huggingface.co/crabz/bertoslav-limited) on the wikiann sk dataset.
|
42 |
+
It achieves the following results on the evaluation set:
|
43 |
+
- Loss: 0.2119
|
44 |
+
- Precision: 0.8986
|
45 |
+
- Recall: 0.9174
|
46 |
+
- F1: 0.9079
|
47 |
+
- Accuracy: 0.9700
|
48 |
+
|
49 |
+
## Model description
|
50 |
+
|
51 |
+
More information needed
|
52 |
+
|
53 |
+
## Intended uses & limitations
|
54 |
+
|
55 |
+
More information needed
|
56 |
+
|
57 |
+
## Training and evaluation data
|
58 |
+
|
59 |
+
More information needed
|
60 |
+
|
61 |
+
## Training procedure
|
62 |
+
|
63 |
+
### Training hyperparameters
|
64 |
+
|
65 |
+
The following hyperparameters were used during training:
|
66 |
+
- learning_rate: 5e-05
|
67 |
+
- train_batch_size: 24
|
68 |
+
- eval_batch_size: 24
|
69 |
+
- seed: 42
|
70 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
71 |
+
- lr_scheduler_type: linear
|
72 |
+
- num_epochs: 10.0
|
73 |
+
|
74 |
+
### Training results
|
75 |
+
|
76 |
+
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|
77 |
+
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
|
78 |
+
| 0.2953 | 1.0 | 834 | 0.1516 | 0.8413 | 0.8647 | 0.8529 | 0.9549 |
|
79 |
+
| 0.0975 | 2.0 | 1668 | 0.1304 | 0.8787 | 0.9056 | 0.8920 | 0.9658 |
|
80 |
+
| 0.0487 | 3.0 | 2502 | 0.1405 | 0.8916 | 0.8958 | 0.8937 | 0.9660 |
|
81 |
+
| 0.025 | 4.0 | 3336 | 0.1658 | 0.8850 | 0.9116 | 0.8981 | 0.9669 |
|
82 |
+
| 0.0161 | 5.0 | 4170 | 0.1739 | 0.8974 | 0.9127 | 0.9050 | 0.9693 |
|
83 |
+
| 0.0074 | 6.0 | 5004 | 0.1888 | 0.8900 | 0.9144 | 0.9020 | 0.9687 |
|
84 |
+
| 0.0051 | 7.0 | 5838 | 0.1996 | 0.8946 | 0.9145 | 0.9044 | 0.9693 |
|
85 |
+
| 0.0039 | 8.0 | 6672 | 0.2052 | 0.8993 | 0.9158 | 0.9075 | 0.9697 |
|
86 |
+
| 0.0024 | 9.0 | 7506 | 0.2112 | 0.8946 | 0.9171 | 0.9057 | 0.9696 |
|
87 |
+
| 0.0018 | 10.0 | 8340 | 0.2119 | 0.8986 | 0.9174 | 0.9079 | 0.9700 |
|
88 |
+
|
89 |
+
|
90 |
+
### Framework versions
|
91 |
+
|
92 |
+
- Transformers 4.14.0.dev0
|
93 |
+
- Pytorch 1.10.0
|
94 |
+
- Datasets 1.16.1
|
95 |
+
- Tokenizers 0.10.3
|
config.json
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "crabz/bertoslav-limited",
|
3 |
+
"activation": "gelu",
|
4 |
+
"architectures": [
|
5 |
+
"DistilBertForTokenClassification"
|
6 |
+
],
|
7 |
+
"attention_dropout": 0.1,
|
8 |
+
"dim": 768,
|
9 |
+
"dropout": 0.1,
|
10 |
+
"finetuning_task": "ner",
|
11 |
+
"hidden_dim": 3072,
|
12 |
+
"id2label": {
|
13 |
+
"0": 0,
|
14 |
+
"1": 1,
|
15 |
+
"2": 2,
|
16 |
+
"3": 3,
|
17 |
+
"4": 4,
|
18 |
+
"5": 5,
|
19 |
+
"6": 6
|
20 |
+
},
|
21 |
+
"initializer_range": 0.02,
|
22 |
+
"label2id": {
|
23 |
+
"0": 0,
|
24 |
+
"1": 1,
|
25 |
+
"2": 2,
|
26 |
+
"3": 3,
|
27 |
+
"4": 4,
|
28 |
+
"5": 5,
|
29 |
+
"6": 6
|
30 |
+
},
|
31 |
+
"max_position_embeddings": 512,
|
32 |
+
"model_type": "distilbert",
|
33 |
+
"n_heads": 12,
|
34 |
+
"n_layers": 6,
|
35 |
+
"pad_token_id": 0,
|
36 |
+
"qa_dropout": 0.1,
|
37 |
+
"seq_classif_dropout": 0.2,
|
38 |
+
"sinusoidal_pos_embds": false,
|
39 |
+
"torch_dtype": "float32",
|
40 |
+
"transformers_version": "4.14.0.dev0",
|
41 |
+
"vocab_size": 30000
|
42 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:625090146b1b2550cad57d542c1a05c11be11e600dbef240bb8fb752015a8b88
|
3 |
+
size 263908853
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"special_tokens_map_file": "/ext/ivan/c4/multilingual/tokenizer/trained_tokenizer/special_tokens_map.json", "name_or_path": "crabz/bertoslav-limited", "tokenizer_class": "PreTrainedTokenizerFast"}
|