End of training
Browse files- README.md +19 -85
- adapter.bew.safetensors +1 -1
- model.safetensors +1 -1
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -1,7 +1,5 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
-
license: cc-by-nc-4.0
|
| 4 |
-
base_model: facebook/mms-1b-all
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
| 7 |
metrics:
|
|
@@ -16,11 +14,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# ssc-bew-mms-model-mix-adapt-max3-devtrain
|
| 18 |
|
| 19 |
-
This model
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
-
- Cer: 0.
|
| 23 |
-
- Wer: 0.
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
@@ -40,95 +38,31 @@ More information needed
|
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
- learning_rate: 0.0005
|
| 43 |
-
- train_batch_size:
|
| 44 |
- eval_batch_size: 6
|
| 45 |
- seed: 42
|
| 46 |
- gradient_accumulation_steps: 2
|
| 47 |
-
- total_train_batch_size:
|
| 48 |
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 49 |
- lr_scheduler_type: linear
|
| 50 |
- lr_scheduler_warmup_steps: 100
|
| 51 |
-
- num_epochs:
|
| 52 |
- mixed_precision_training: Native AMP
|
| 53 |
|
| 54 |
### Training results
|
| 55 |
|
| 56 |
-
| Training Loss | Epoch
|
| 57 |
-
|:-------------:|:------
|
| 58 |
-
|
|
| 59 |
-
|
|
| 60 |
-
| 0.
|
| 61 |
-
| 0.
|
| 62 |
-
| 0.
|
| 63 |
-
| 0.
|
| 64 |
-
| 0.
|
| 65 |
-
| 0.
|
| 66 |
-
| 0.
|
| 67 |
-
| 0.
|
| 68 |
-
| 0.878 | 2.9497 | 2200 | 0.6590 | 0.1705 | 0.5016 |
|
| 69 |
-
| 0.87 | 3.2173 | 2400 | 0.6451 | 0.1692 | 0.5008 |
|
| 70 |
-
| 0.8431 | 3.4856 | 2600 | 0.6385 | 0.1681 | 0.4989 |
|
| 71 |
-
| 0.8301 | 3.7539 | 2800 | 0.6358 | 0.1671 | 0.4966 |
|
| 72 |
-
| 0.8446 | 4.0215 | 3000 | 0.6344 | 0.1664 | 0.4939 |
|
| 73 |
-
| 0.8636 | 4.2897 | 3200 | 0.6380 | 0.1685 | 0.5040 |
|
| 74 |
-
| 0.8475 | 4.5580 | 3400 | 0.6249 | 0.1648 | 0.4921 |
|
| 75 |
-
| 0.855 | 4.8263 | 3600 | 0.6209 | 0.1636 | 0.4869 |
|
| 76 |
-
| 0.8315 | 5.0939 | 3800 | 0.6188 | 0.1625 | 0.4858 |
|
| 77 |
-
| 0.7907 | 5.3622 | 4000 | 0.6157 | 0.1623 | 0.4836 |
|
| 78 |
-
| 0.8096 | 5.6304 | 4200 | 0.6171 | 0.1631 | 0.4912 |
|
| 79 |
-
| 0.8157 | 5.8987 | 4400 | 0.6056 | 0.1611 | 0.4785 |
|
| 80 |
-
| 0.8043 | 6.1663 | 4600 | 0.6086 | 0.1615 | 0.4877 |
|
| 81 |
-
| 0.79 | 6.4346 | 4800 | 0.6015 | 0.1599 | 0.4817 |
|
| 82 |
-
| 0.8058 | 6.7029 | 5000 | 0.6016 | 0.1597 | 0.4800 |
|
| 83 |
-
| 0.7965 | 6.9712 | 5200 | 0.5985 | 0.1591 | 0.4751 |
|
| 84 |
-
| 0.7811 | 7.2388 | 5400 | 0.5994 | 0.1582 | 0.4731 |
|
| 85 |
-
| 0.7642 | 7.5070 | 5600 | 0.5963 | 0.1571 | 0.4695 |
|
| 86 |
-
| 0.7975 | 7.7753 | 5800 | 0.6006 | 0.1580 | 0.4732 |
|
| 87 |
-
| 0.8164 | 8.0429 | 6000 | 0.5923 | 0.1573 | 0.4720 |
|
| 88 |
-
| 0.7464 | 8.3112 | 6200 | 0.5944 | 0.1556 | 0.4656 |
|
| 89 |
-
| 0.7582 | 8.5795 | 6400 | 0.5879 | 0.1546 | 0.4632 |
|
| 90 |
-
| 0.8074 | 8.8478 | 6600 | 0.5939 | 0.1557 | 0.4631 |
|
| 91 |
-
| 0.7227 | 9.1154 | 6800 | 0.5839 | 0.1548 | 0.4624 |
|
| 92 |
-
| 0.7416 | 9.3836 | 7000 | 0.5762 | 0.1539 | 0.4624 |
|
| 93 |
-
| 0.7886 | 9.6519 | 7200 | 0.5712 | 0.1531 | 0.4603 |
|
| 94 |
-
| 0.7425 | 9.9202 | 7400 | 0.5703 | 0.1523 | 0.4590 |
|
| 95 |
-
| 0.7471 | 10.1878 | 7600 | 0.5690 | 0.1523 | 0.4596 |
|
| 96 |
-
| 0.7391 | 10.4561 | 7800 | 0.5622 | 0.1525 | 0.4613 |
|
| 97 |
-
| 0.7354 | 10.7243 | 8000 | 0.5863 | 0.1533 | 0.4573 |
|
| 98 |
-
| 0.768 | 10.9926 | 8200 | 0.5637 | 0.1513 | 0.4566 |
|
| 99 |
-
| 0.7394 | 11.2602 | 8400 | 0.5553 | 0.1512 | 0.4569 |
|
| 100 |
-
| 0.7257 | 11.5285 | 8600 | 0.5611 | 0.1500 | 0.4504 |
|
| 101 |
-
| 0.7402 | 11.7968 | 8800 | 0.5561 | 0.1513 | 0.4594 |
|
| 102 |
-
| 0.7179 | 12.0644 | 9000 | 0.5560 | 0.1526 | 0.4657 |
|
| 103 |
-
| 0.7 | 12.3327 | 9200 | 0.5605 | 0.1495 | 0.4496 |
|
| 104 |
-
| 0.7128 | 12.6009 | 9400 | 0.5576 | 0.1484 | 0.4441 |
|
| 105 |
-
| 0.7067 | 12.8692 | 9600 | 0.5525 | 0.1475 | 0.4454 |
|
| 106 |
-
| 0.7257 | 13.1368 | 9800 | 0.5496 | 0.1474 | 0.4450 |
|
| 107 |
-
| 0.719 | 13.4051 | 10000 | 0.5442 | 0.1473 | 0.4462 |
|
| 108 |
-
| 0.72 | 13.6734 | 10200 | 0.5420 | 0.1468 | 0.4423 |
|
| 109 |
-
| 0.6803 | 13.9416 | 10400 | 0.5526 | 0.1473 | 0.4443 |
|
| 110 |
-
| 0.6859 | 14.2093 | 10600 | 0.5478 | 0.1469 | 0.4441 |
|
| 111 |
-
| 0.7255 | 14.4775 | 10800 | 0.5392 | 0.1466 | 0.4410 |
|
| 112 |
-
| 0.6941 | 14.7458 | 11000 | 0.5413 | 0.1460 | 0.4375 |
|
| 113 |
-
| 0.6902 | 15.0134 | 11200 | 0.5367 | 0.1443 | 0.4351 |
|
| 114 |
-
| 0.7062 | 15.2817 | 11400 | 0.5349 | 0.1443 | 0.4371 |
|
| 115 |
-
| 0.6891 | 15.5500 | 11600 | 0.5309 | 0.1437 | 0.4330 |
|
| 116 |
-
| 0.6555 | 15.8182 | 11800 | 0.5326 | 0.1443 | 0.4388 |
|
| 117 |
-
| 0.6816 | 16.0858 | 12000 | 0.5262 | 0.1432 | 0.4337 |
|
| 118 |
-
| 0.7046 | 16.3541 | 12200 | 0.5294 | 0.1433 | 0.4328 |
|
| 119 |
-
| 0.6984 | 16.6224 | 12400 | 0.5274 | 0.1432 | 0.4316 |
|
| 120 |
-
| 0.6672 | 16.8907 | 12600 | 0.5241 | 0.1422 | 0.4326 |
|
| 121 |
-
| 0.6953 | 17.1583 | 12800 | 0.5241 | 0.1418 | 0.4319 |
|
| 122 |
-
| 0.6736 | 17.4266 | 13000 | 0.5228 | 0.1420 | 0.4327 |
|
| 123 |
-
| 0.6858 | 17.6948 | 13200 | 0.5268 | 0.1425 | 0.4315 |
|
| 124 |
-
| 0.6969 | 17.9631 | 13400 | 0.5215 | 0.1417 | 0.4308 |
|
| 125 |
-
| 0.667 | 18.2307 | 13600 | 0.5216 | 0.1411 | 0.4302 |
|
| 126 |
-
| 0.6664 | 18.4990 | 13800 | 0.5196 | 0.1410 | 0.4276 |
|
| 127 |
-
| 0.667 | 18.7673 | 14000 | 0.5184 | 0.1407 | 0.4260 |
|
| 128 |
-
| 0.6609 | 19.0349 | 14200 | 0.5197 | 0.1407 | 0.4263 |
|
| 129 |
-
| 0.664 | 19.3032 | 14400 | 0.5202 | 0.1408 | 0.4259 |
|
| 130 |
-
| 0.675 | 19.5714 | 14600 | 0.5183 | 0.1406 | 0.4269 |
|
| 131 |
-
| 0.6707 | 19.8397 | 14800 | 0.5184 | 0.1406 | 0.4260 |
|
| 132 |
|
| 133 |
|
| 134 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
|
|
|
|
|
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
| 5 |
metrics:
|
|
|
|
| 14 |
|
| 15 |
# ssc-bew-mms-model-mix-adapt-max3-devtrain
|
| 16 |
|
| 17 |
+
This model was trained from scratch on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.5232
|
| 20 |
+
- Cer: 0.1493
|
| 21 |
+
- Wer: 0.4536
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 38 |
|
| 39 |
The following hyperparameters were used during training:
|
| 40 |
- learning_rate: 0.0005
|
| 41 |
+
- train_batch_size: 1
|
| 42 |
- eval_batch_size: 6
|
| 43 |
- seed: 42
|
| 44 |
- gradient_accumulation_steps: 2
|
| 45 |
+
- total_train_batch_size: 2
|
| 46 |
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
- lr_scheduler_warmup_steps: 100
|
| 49 |
+
- num_epochs: 5
|
| 50 |
- mixed_precision_training: Native AMP
|
| 51 |
|
| 52 |
### Training results
|
| 53 |
|
| 54 |
+
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
|
| 55 |
+
|:-------------:|:------:|:----:|:---------------:|:------:|:------:|
|
| 56 |
+
| 0.5415 | 0.4920 | 200 | 0.5600 | 0.1618 | 0.4943 |
|
| 57 |
+
| 0.5541 | 0.9840 | 400 | 0.5432 | 0.1611 | 0.4941 |
|
| 58 |
+
| 0.5282 | 1.4748 | 600 | 0.5389 | 0.1541 | 0.4669 |
|
| 59 |
+
| 0.5117 | 1.9668 | 800 | 0.5340 | 0.1549 | 0.4759 |
|
| 60 |
+
| 0.4838 | 2.4576 | 1000 | 0.5371 | 0.1541 | 0.4689 |
|
| 61 |
+
| 0.4974 | 2.9496 | 1200 | 0.5348 | 0.1527 | 0.4646 |
|
| 62 |
+
| 0.4507 | 3.4403 | 1400 | 0.5267 | 0.1503 | 0.4536 |
|
| 63 |
+
| 0.4644 | 3.9323 | 1600 | 0.5283 | 0.1508 | 0.4592 |
|
| 64 |
+
| 0.4739 | 4.4231 | 1800 | 0.5228 | 0.1496 | 0.4553 |
|
| 65 |
+
| 0.4471 | 4.9151 | 2000 | 0.5232 | 0.1493 | 0.4536 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
|
| 67 |
|
| 68 |
### Framework versions
|
adapter.bew.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 9049624
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0bb1cf45b0656ec05fc0f19b6e220eab3d4d2ea72c05f5667989b5cb681118ff
|
| 3 |
size 9049624
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3859142000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e3c8ce9c7bb8d568749c6dba2981fe8b1c4eda94937fee895e2eeabd67642ad
|
| 3 |
size 3859142000
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5969
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:91e277c72286b8423942c1cbc2065476ea0dc5d79a29d9ef51b124e20560d3a1
|
| 3 |
size 5969
|