Finetuned mt5_small_lg_inf_en model
Browse files- README.md +90 -0
- generation_config.json +6 -0
README.md
ADDED
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: transformers
|
3 |
+
base_model: mt5_small/lg_inf_en
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
metrics:
|
7 |
+
- bleu
|
8 |
+
model-index:
|
9 |
+
- name: mt5_small_lg_inf_en
|
10 |
+
results: []
|
11 |
+
---
|
12 |
+
|
13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
14 |
+
should probably proofread and complete it, then remove this comment. -->
|
15 |
+
|
16 |
+
# mt5_small_lg_inf_en
|
17 |
+
|
18 |
+
This model is a fine-tuned version of [mt5_small/lg_inf_en](https://huggingface.co/mt5_small/lg_inf_en) on the Luganda Informal Data dataset.
|
19 |
+
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 0.4301
|
21 |
+
- Bleu: 0.3034
|
22 |
+
- Gen Len: 8.1551
|
23 |
+
|
24 |
+
## Model description
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Intended uses & limitations
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training and evaluation data
|
33 |
+
|
34 |
+
More information needed
|
35 |
+
|
36 |
+
## Training procedure
|
37 |
+
|
38 |
+
### Training hyperparameters
|
39 |
+
|
40 |
+
The following hyperparameters were used during training:
|
41 |
+
- learning_rate: 2e-05
|
42 |
+
- train_batch_size: 16
|
43 |
+
- eval_batch_size: 16
|
44 |
+
- seed: 42
|
45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
+
- lr_scheduler_type: linear
|
47 |
+
- num_epochs: 30
|
48 |
+
|
49 |
+
### Training results
|
50 |
+
|
51 |
+
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|
52 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|:-------:|
|
53 |
+
| No log | 1.0 | 138 | 0.4671 | 0.0646 | 9.4449 |
|
54 |
+
| No log | 2.0 | 276 | 0.4562 | 0.1318 | 7.8898 |
|
55 |
+
| No log | 3.0 | 414 | 0.4511 | 0.2119 | 7.9878 |
|
56 |
+
| 0.4729 | 4.0 | 552 | 0.4476 | 0.2133 | 8.1184 |
|
57 |
+
| 0.4729 | 5.0 | 690 | 0.4451 | 0.2128 | 8.0816 |
|
58 |
+
| 0.4729 | 6.0 | 828 | 0.4433 | 0.3272 | 7.9224 |
|
59 |
+
| 0.4729 | 7.0 | 966 | 0.4415 | 0.3383 | 7.6571 |
|
60 |
+
| 0.4479 | 8.0 | 1104 | 0.4401 | 0.3281 | 7.5347 |
|
61 |
+
| 0.4479 | 9.0 | 1242 | 0.4390 | 0.3296 | 7.4286 |
|
62 |
+
| 0.4479 | 10.0 | 1380 | 0.4378 | 0.3157 | 7.6 |
|
63 |
+
| 0.4418 | 11.0 | 1518 | 0.4367 | 0.3288 | 7.4327 |
|
64 |
+
| 0.4418 | 12.0 | 1656 | 0.4360 | 0.316 | 7.4857 |
|
65 |
+
| 0.4418 | 13.0 | 1794 | 0.4350 | 0.3167 | 7.4898 |
|
66 |
+
| 0.4418 | 14.0 | 1932 | 0.4342 | 0.3161 | 7.698 |
|
67 |
+
| 0.4347 | 15.0 | 2070 | 0.4337 | 0.316 | 7.849 |
|
68 |
+
| 0.4347 | 16.0 | 2208 | 0.4333 | 0.3177 | 7.6735 |
|
69 |
+
| 0.4347 | 17.0 | 2346 | 0.4326 | 0.3174 | 7.8082 |
|
70 |
+
| 0.4347 | 18.0 | 2484 | 0.4324 | 0.3167 | 7.8531 |
|
71 |
+
| 0.4315 | 19.0 | 2622 | 0.4319 | 0.3185 | 8.0163 |
|
72 |
+
| 0.4315 | 20.0 | 2760 | 0.4316 | 0.318 | 8.0449 |
|
73 |
+
| 0.4315 | 21.0 | 2898 | 0.4313 | 0.3171 | 8.0571 |
|
74 |
+
| 0.4289 | 22.0 | 3036 | 0.4311 | 0.3195 | 7.9837 |
|
75 |
+
| 0.4289 | 23.0 | 3174 | 0.4308 | 0.3188 | 8.049 |
|
76 |
+
| 0.4289 | 24.0 | 3312 | 0.4307 | 0.3048 | 8.0694 |
|
77 |
+
| 0.4289 | 25.0 | 3450 | 0.4304 | 0.3046 | 8.1306 |
|
78 |
+
| 0.4264 | 26.0 | 3588 | 0.4303 | 0.3041 | 8.1224 |
|
79 |
+
| 0.4264 | 27.0 | 3726 | 0.4302 | 0.3044 | 8.1592 |
|
80 |
+
| 0.4264 | 28.0 | 3864 | 0.4301 | 0.3046 | 8.1306 |
|
81 |
+
| 0.4256 | 29.0 | 4002 | 0.4301 | 0.3039 | 8.1429 |
|
82 |
+
| 0.4256 | 30.0 | 4140 | 0.4301 | 0.3034 | 8.1551 |
|
83 |
+
|
84 |
+
|
85 |
+
### Framework versions
|
86 |
+
|
87 |
+
- Transformers 4.45.1
|
88 |
+
- Pytorch 2.4.0
|
89 |
+
- Datasets 3.0.1
|
90 |
+
- Tokenizers 0.20.0
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"decoder_start_token_id": 0,
|
3 |
+
"eos_token_id": 1,
|
4 |
+
"pad_token_id": 0,
|
5 |
+
"transformers_version": "4.45.1"
|
6 |
+
}
|