qwen-rep-nampdn-ai/tiny-textbooks
Browse files
README.md
CHANGED
|
@@ -14,7 +14,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
|
| 15 |
This model is a fine-tuned version of [Qwen/Qwen-14B](https://huggingface.co/Qwen/Qwen-14B) on an unknown dataset.
|
| 16 |
It achieves the following results on the evaluation set:
|
| 17 |
-
- Loss: 2.
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
@@ -33,12 +33,12 @@ More information needed
|
|
| 33 |
### Training hyperparameters
|
| 34 |
|
| 35 |
The following hyperparameters were used during training:
|
| 36 |
-
- learning_rate:
|
| 37 |
-
- train_batch_size:
|
| 38 |
-
- eval_batch_size:
|
| 39 |
- seed: 42
|
| 40 |
- gradient_accumulation_steps: 8
|
| 41 |
-
- total_train_batch_size:
|
| 42 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
- lr_scheduler_type: cosine
|
| 44 |
- lr_scheduler_warmup_steps: 0.01
|
|
@@ -46,62 +46,63 @@ The following hyperparameters were used during training:
|
|
| 46 |
|
| 47 |
### Training results
|
| 48 |
|
| 49 |
-
| Training Loss | Epoch | Step
|
| 50 |
-
|
| 51 |
-
| 2.
|
| 52 |
-
| 2.
|
| 53 |
-
| 2.
|
| 54 |
-
| 2.
|
| 55 |
-
| 2.
|
| 56 |
-
| 2.
|
| 57 |
-
| 2.
|
| 58 |
-
| 2.
|
| 59 |
-
| 2.
|
| 60 |
-
| 2.
|
| 61 |
-
| 2.
|
| 62 |
-
| 2.
|
| 63 |
-
| 2.
|
| 64 |
-
| 2.
|
| 65 |
-
| 2.
|
| 66 |
-
| 2.
|
| 67 |
-
| 2.
|
| 68 |
-
| 2.
|
| 69 |
-
| 2.
|
| 70 |
-
| 2.
|
| 71 |
-
| 2.
|
| 72 |
-
| 2.
|
| 73 |
-
| 2.
|
| 74 |
-
| 2.
|
| 75 |
-
| 2.
|
| 76 |
-
| 2.
|
| 77 |
-
| 2.
|
| 78 |
-
| 2.
|
| 79 |
-
| 2.
|
| 80 |
-
| 2.
|
| 81 |
-
| 2.
|
| 82 |
-
| 2.
|
| 83 |
-
| 2.
|
| 84 |
-
| 2.
|
| 85 |
-
| 2.
|
| 86 |
-
| 2.
|
| 87 |
-
| 2.
|
| 88 |
-
| 2.
|
| 89 |
-
| 2.
|
| 90 |
-
| 2.
|
| 91 |
-
| 2.
|
| 92 |
-
| 2.
|
| 93 |
-
| 2.
|
| 94 |
-
| 2.
|
| 95 |
-
| 2.
|
| 96 |
-
| 2.
|
| 97 |
-
| 2.
|
| 98 |
-
| 2.
|
| 99 |
-
| 2.
|
|
|
|
| 100 |
|
| 101 |
|
| 102 |
### Framework versions
|
| 103 |
|
| 104 |
-
- Transformers 4.
|
| 105 |
-
- Pytorch 2.1.0
|
| 106 |
-
- Datasets 2.5
|
| 107 |
-
- Tokenizers 0.14.
|
|
|
|
| 14 |
|
| 15 |
This model is a fine-tuned version of [Qwen/Qwen-14B](https://huggingface.co/Qwen/Qwen-14B) on an unknown dataset.
|
| 16 |
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 2.3572
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
|
|
| 33 |
### Training hyperparameters
|
| 34 |
|
| 35 |
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 1e-05
|
| 37 |
+
- train_batch_size: 1
|
| 38 |
+
- eval_batch_size: 1
|
| 39 |
- seed: 42
|
| 40 |
- gradient_accumulation_steps: 8
|
| 41 |
+
- total_train_batch_size: 8
|
| 42 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
- lr_scheduler_type: cosine
|
| 44 |
- lr_scheduler_warmup_steps: 0.01
|
|
|
|
| 46 |
|
| 47 |
### Training results
|
| 48 |
|
| 49 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 50 |
+
|:-------------:|:-----:|:-----:|:---------------:|
|
| 51 |
+
| 2.4651 | 0.02 | 200 | 2.3996 |
|
| 52 |
+
| 2.4335 | 0.04 | 400 | 2.3799 |
|
| 53 |
+
| 2.3848 | 0.06 | 600 | 2.3746 |
|
| 54 |
+
| 2.4037 | 0.08 | 800 | 2.3714 |
|
| 55 |
+
| 2.3985 | 0.1 | 1000 | 2.3693 |
|
| 56 |
+
| 2.4072 | 0.12 | 1200 | 2.3673 |
|
| 57 |
+
| 2.4028 | 0.14 | 1400 | 2.3665 |
|
| 58 |
+
| 2.3748 | 0.16 | 1600 | 2.3643 |
|
| 59 |
+
| 2.4119 | 0.18 | 1800 | 2.3635 |
|
| 60 |
+
| 2.4002 | 0.2 | 2000 | 2.3640 |
|
| 61 |
+
| 2.3865 | 0.22 | 2200 | 2.3635 |
|
| 62 |
+
| 2.4 | 0.24 | 2400 | 2.3628 |
|
| 63 |
+
| 2.4096 | 0.26 | 2600 | 2.3625 |
|
| 64 |
+
| 2.3976 | 0.28 | 2800 | 2.3614 |
|
| 65 |
+
| 2.3767 | 0.3 | 3000 | 2.3618 |
|
| 66 |
+
| 2.4151 | 0.32 | 3200 | 2.3616 |
|
| 67 |
+
| 2.3835 | 0.34 | 3400 | 2.3605 |
|
| 68 |
+
| 2.3995 | 0.36 | 3600 | 2.3608 |
|
| 69 |
+
| 2.4121 | 0.38 | 3800 | 2.3602 |
|
| 70 |
+
| 2.4262 | 0.4 | 4000 | 2.3591 |
|
| 71 |
+
| 2.3604 | 0.42 | 4200 | 2.3594 |
|
| 72 |
+
| 2.3954 | 0.44 | 4400 | 2.3594 |
|
| 73 |
+
| 2.3743 | 0.46 | 4600 | 2.3587 |
|
| 74 |
+
| 2.4069 | 0.48 | 4800 | 2.3591 |
|
| 75 |
+
| 2.4103 | 0.5 | 5000 | 2.3585 |
|
| 76 |
+
| 2.4133 | 0.52 | 5200 | 2.3585 |
|
| 77 |
+
| 2.4229 | 0.54 | 5400 | 2.3578 |
|
| 78 |
+
| 2.4397 | 0.56 | 5600 | 2.3581 |
|
| 79 |
+
| 2.4237 | 0.58 | 5800 | 2.3581 |
|
| 80 |
+
| 2.4109 | 0.6 | 6000 | 2.3577 |
|
| 81 |
+
| 2.43 | 0.62 | 6200 | 2.3575 |
|
| 82 |
+
| 2.3999 | 0.64 | 6400 | 2.3572 |
|
| 83 |
+
| 2.3771 | 0.66 | 6600 | 2.3577 |
|
| 84 |
+
| 2.4119 | 0.68 | 6800 | 2.3576 |
|
| 85 |
+
| 2.3877 | 0.7 | 7000 | 2.3576 |
|
| 86 |
+
| 2.411 | 0.72 | 7200 | 2.3569 |
|
| 87 |
+
| 2.3808 | 0.74 | 7400 | 2.3570 |
|
| 88 |
+
| 2.3989 | 0.76 | 7600 | 2.3571 |
|
| 89 |
+
| 2.422 | 0.78 | 7800 | 2.3569 |
|
| 90 |
+
| 2.3768 | 0.8 | 8000 | 2.3569 |
|
| 91 |
+
| 2.3988 | 0.82 | 8200 | 2.3572 |
|
| 92 |
+
| 2.3927 | 0.84 | 8400 | 2.3572 |
|
| 93 |
+
| 2.3961 | 0.86 | 8600 | 2.3573 |
|
| 94 |
+
| 2.4021 | 0.88 | 8800 | 2.3570 |
|
| 95 |
+
| 2.3889 | 0.9 | 9000 | 2.3570 |
|
| 96 |
+
| 2.404 | 0.92 | 9200 | 2.3570 |
|
| 97 |
+
| 2.3982 | 0.94 | 9400 | 2.3572 |
|
| 98 |
+
| 2.4018 | 0.96 | 9600 | 2.3573 |
|
| 99 |
+
| 2.3717 | 0.98 | 9800 | 2.3572 |
|
| 100 |
+
| 2.4076 | 1.0 | 10000 | 2.3572 |
|
| 101 |
|
| 102 |
|
| 103 |
### Framework versions
|
| 104 |
|
| 105 |
+
- Transformers 4.34.1
|
| 106 |
+
- Pytorch 2.1.0
|
| 107 |
+
- Datasets 2.14.5
|
| 108 |
+
- Tokenizers 0.14.1
|