Iiro/bert_reviews
Browse files- README.md +48 -11
- config.json +19 -4
- pytorch_model.bin +2 -2
- training_args.bin +1 -1
README.md
CHANGED
@@ -1,12 +1,28 @@
|
|
1 |
---
|
|
|
|
|
2 |
tags:
|
3 |
- generated_from_trainer
|
4 |
datasets:
|
5 |
- amazon_reviews_multi
|
|
|
|
|
6 |
model-index:
|
7 |
- name: bert_reviews
|
8 |
-
results:
|
9 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
---
|
11 |
|
12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -14,15 +30,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
14 |
|
15 |
# bert_reviews
|
16 |
|
17 |
-
This model
|
18 |
It achieves the following results on the evaluation set:
|
19 |
-
-
|
20 |
-
-
|
21 |
-
- eval_runtime: 39.7329
|
22 |
-
- eval_samples_per_second: 125.84
|
23 |
-
- eval_steps_per_second: 15.73
|
24 |
-
- epoch: 0.52
|
25 |
-
- step: 13000
|
26 |
|
27 |
## Model description
|
28 |
|
@@ -49,9 +60,35 @@ The following hyperparameters were used during training:
|
|
49 |
- lr_scheduler_type: linear
|
50 |
- training_steps: 20000
|
51 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
### Framework versions
|
53 |
|
54 |
- Transformers 4.34.1
|
55 |
- Pytorch 2.1.0+cu118
|
56 |
- Datasets 2.14.6
|
57 |
-
- Tokenizers 0.14.1
|
|
|
1 |
---
|
2 |
+
license: apache-2.0
|
3 |
+
base_model: distilbert-base-uncased
|
4 |
tags:
|
5 |
- generated_from_trainer
|
6 |
datasets:
|
7 |
- amazon_reviews_multi
|
8 |
+
metrics:
|
9 |
+
- accuracy
|
10 |
model-index:
|
11 |
- name: bert_reviews
|
12 |
+
results:
|
13 |
+
- task:
|
14 |
+
name: Text Classification
|
15 |
+
type: text-classification
|
16 |
+
dataset:
|
17 |
+
name: amazon_reviews_multi
|
18 |
+
type: amazon_reviews_multi
|
19 |
+
config: en
|
20 |
+
split: test
|
21 |
+
args: en
|
22 |
+
metrics:
|
23 |
+
- name: Accuracy
|
24 |
+
type: accuracy
|
25 |
+
value: 0.6062
|
26 |
---
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
30 |
|
31 |
# bert_reviews
|
32 |
|
33 |
+
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the amazon_reviews_multi dataset.
|
34 |
It achieves the following results on the evaluation set:
|
35 |
+
- Loss: 0.9204
|
36 |
+
- Accuracy: 0.6062
|
|
|
|
|
|
|
|
|
|
|
37 |
|
38 |
## Model description
|
39 |
|
|
|
60 |
- lr_scheduler_type: linear
|
61 |
- training_steps: 20000
|
62 |
|
63 |
+
### Training results
|
64 |
+
|
65 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
66 |
+
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
|
67 |
+
| 0.8812 | 0.04 | 1000 | 0.9970 | 0.5738 |
|
68 |
+
| 0.8495 | 0.08 | 2000 | 1.0120 | 0.569 |
|
69 |
+
| 0.8067 | 0.12 | 3000 | 1.0442 | 0.5766 |
|
70 |
+
| 0.7934 | 0.16 | 4000 | 1.0629 | 0.5772 |
|
71 |
+
| 0.7845 | 0.2 | 5000 | 1.0236 | 0.5876 |
|
72 |
+
| 0.9033 | 0.24 | 6000 | 0.9822 | 0.5774 |
|
73 |
+
| 0.8993 | 0.28 | 7000 | 0.9693 | 0.5816 |
|
74 |
+
| 0.9012 | 0.32 | 8000 | 1.0075 | 0.5738 |
|
75 |
+
| 0.873 | 0.36 | 9000 | 0.9663 | 0.5886 |
|
76 |
+
| 0.9376 | 0.4 | 10000 | 0.9447 | 0.5816 |
|
77 |
+
| 0.9398 | 0.44 | 11000 | 0.9509 | 0.5802 |
|
78 |
+
| 0.9402 | 0.48 | 12000 | 0.9561 | 0.5916 |
|
79 |
+
| 0.9247 | 0.52 | 13000 | 0.9303 | 0.6008 |
|
80 |
+
| 0.9247 | 0.56 | 14000 | 0.9241 | 0.5998 |
|
81 |
+
| 0.9192 | 0.6 | 15000 | 0.9276 | 0.6104 |
|
82 |
+
| 0.907 | 0.64 | 16000 | 0.9251 | 0.603 |
|
83 |
+
| 0.9177 | 0.68 | 17000 | 0.9198 | 0.6056 |
|
84 |
+
| 0.9129 | 0.72 | 18000 | 0.9167 | 0.6078 |
|
85 |
+
| 0.8948 | 0.76 | 19000 | 0.9213 | 0.604 |
|
86 |
+
| 0.906 | 0.8 | 20000 | 0.9204 | 0.6062 |
|
87 |
+
|
88 |
+
|
89 |
### Framework versions
|
90 |
|
91 |
- Transformers 4.34.1
|
92 |
- Pytorch 2.1.0+cu118
|
93 |
- Datasets 2.14.6
|
94 |
+
- Tokenizers 0.14.1
|
config.json
CHANGED
@@ -1,9 +1,13 @@
|
|
1 |
{
|
|
|
|
|
2 |
"architectures": [
|
3 |
-
"
|
4 |
],
|
5 |
-
"
|
6 |
-
"
|
|
|
|
|
7 |
"id2label": {
|
8 |
"0": "LABEL_0",
|
9 |
"1": "LABEL_1",
|
@@ -11,6 +15,7 @@
|
|
11 |
"3": "LABEL_3",
|
12 |
"4": "LABEL_4"
|
13 |
},
|
|
|
14 |
"label2id": {
|
15 |
"LABEL_0": 0,
|
16 |
"LABEL_1": 1,
|
@@ -18,7 +23,17 @@
|
|
18 |
"LABEL_3": 3,
|
19 |
"LABEL_4": 4
|
20 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
"torch_dtype": "float32",
|
22 |
"transformers_version": "4.34.1",
|
23 |
-
"vocab_size":
|
24 |
}
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "distilbert-base-uncased",
|
3 |
+
"activation": "gelu",
|
4 |
"architectures": [
|
5 |
+
"DistilBertForSequenceClassification"
|
6 |
],
|
7 |
+
"attention_dropout": 0.1,
|
8 |
+
"dim": 768,
|
9 |
+
"dropout": 0.1,
|
10 |
+
"hidden_dim": 3072,
|
11 |
"id2label": {
|
12 |
"0": "LABEL_0",
|
13 |
"1": "LABEL_1",
|
|
|
15 |
"3": "LABEL_3",
|
16 |
"4": "LABEL_4"
|
17 |
},
|
18 |
+
"initializer_range": 0.02,
|
19 |
"label2id": {
|
20 |
"LABEL_0": 0,
|
21 |
"LABEL_1": 1,
|
|
|
23 |
"LABEL_3": 3,
|
24 |
"LABEL_4": 4
|
25 |
},
|
26 |
+
"max_position_embeddings": 512,
|
27 |
+
"model_type": "distilbert",
|
28 |
+
"n_heads": 12,
|
29 |
+
"n_layers": 6,
|
30 |
+
"pad_token_id": 0,
|
31 |
+
"problem_type": "single_label_classification",
|
32 |
+
"qa_dropout": 0.1,
|
33 |
+
"seq_classif_dropout": 0.2,
|
34 |
+
"sinusoidal_pos_embds": false,
|
35 |
+
"tie_weights_": true,
|
36 |
"torch_dtype": "float32",
|
37 |
"transformers_version": "4.34.1",
|
38 |
+
"vocab_size": 30522
|
39 |
}
|
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 sha256:86ee65e834160aace6c12aa0d503cd6099f9b07664b12a565dbb52d046636268
|
3 |
+
size 267865194
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4472
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d9a7349f61988d6e6a3c339b566c8b29407e9359421ec483dcd69ca1f45a41f5
|
3 |
size 4472
|