Hugging Face's logo

lapp0
/
distily_experiments_loss_kl

TensorBoard
Safetensors
distily
qwen2
Generated from Trainer
8-bit precision
bitsandbytes
Model card Files Metrics Community
lapp0 commited on
Commit
6d41d04
•
1 Parent(s): 57f636b

End of training

Browse files
Files changed (2) hide show
  1. README.md +24 -24
  2. runs/Aug07_03-03-24_232a0f8c3879/events.out.tfevents.1723024479.232a0f8c3879 +3 -0
README.md CHANGED
@@ -16,13 +16,13 @@ This student model is distilled from the teacher model [Qwen/Qwen2-0.5B-Instruct
16
  The [Distily](https://github.com/lapp0/distily) library was used for this distillation.
17
 
18
  It achieves the following results on the evaluation set:
19
- - eval_enwikippl: nan
20
- - eval_frwikippl: nan
21
- - eval_zhwikippl: nan
22
- - eval_loss: nan
23
- - eval_runtime: 90.8319
24
- - eval_samples_per_second: 11.009
25
- - eval_steps_per_second: 2.752
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment.
@@ -49,36 +49,36 @@ The following hyperparameters were used during training:
49
  - loss_fn: kl
50
  - train_embeddings: True
51
  - learning_rate: 4e-05
52
- - train_batch_size: 4
53
  - eval_batch_size: 4
54
  - seed: 42
55
- - gradient_accumulation_steps: 4
56
  - total_train_batch_size: 16
57
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
58
  - lr_scheduler_type: constant
59
  - num_epochs: 1.0
60
 
61
  ### Resource Usage
62
- Peak GPU Memory: 19.8804 GB
63
 
64
  ### Eval-Phase Metrics
65
  | step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | zhwikippl |
66
  | --- | --- | --- | --- | --- | --- | --- | --- | --- |
67
  | **teacher eval** | | 13.0697 | 11.6518 | | | | | 21.6262 |
68
- | 0 | 0 | 179854.0938 | 180181.3125 | 782.5541 | 90.5683 | 11.041 | 2.76 | 180786.8438 |
69
- | 500 | 0.0808 | nan | nan | nan | 90.7978 | 11.013 | 2.753 | nan |
70
- | 1000 | 0.1616 | nan | nan | nan | 91.1195 | 10.975 | 2.744 | nan |
71
- | 1500 | 0.2424 | nan | nan | nan | 90.9106 | 11.0 | 2.75 | nan |
72
- | 2000 | 0.3232 | nan | nan | nan | 90.7616 | 11.018 | 2.754 | nan |
73
- | 2500 | 0.4040 | nan | nan | nan | 90.73 | 11.022 | 2.755 | nan |
74
- | 3000 | 0.4848 | nan | nan | nan | 90.7247 | 11.022 | 2.756 | nan |
75
- | 3500 | 0.5657 | nan | nan | nan | 90.5701 | 11.041 | 2.76 | nan |
76
- | 4000 | 0.6465 | nan | nan | nan | 90.768 | 11.017 | 2.754 | nan |
77
- | 4500 | 0.7273 | nan | nan | nan | 90.8206 | 11.011 | 2.753 | nan |
78
- | 5000 | 0.8081 | nan | nan | nan | 90.7225 | 11.023 | 2.756 | nan |
79
- | 5500 | 0.8889 | nan | nan | nan | 90.7674 | 11.017 | 2.754 | nan |
80
- | 6000 | 0.9697 | nan | nan | nan | 90.7133 | 11.024 | 2.756 | nan |
81
- | 6187 | 0.9999 | nan | nan | nan | 90.8319 | 11.009 | 2.752 | nan |
82
 
83
  ### Framework versions
84
  - Distily 0.1.0
 
16
  The [Distily](https://github.com/lapp0/distily) library was used for this distillation.
17
 
18
  It achieves the following results on the evaluation set:
19
+ - eval_enwikippl: 44644.5977
20
+ - eval_frwikippl: 57738.3242
21
+ - eval_zhwikippl: 350857.0312
22
+ - eval_loss: 34.5918
23
+ - eval_runtime: 91.6837
24
+ - eval_samples_per_second: 10.907
25
+ - eval_steps_per_second: 2.727
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment.
 
49
  - loss_fn: kl
50
  - train_embeddings: True
51
  - learning_rate: 4e-05
52
+ - train_batch_size: 2
53
  - eval_batch_size: 4
54
  - seed: 42
55
+ - gradient_accumulation_steps: 8
56
  - total_train_batch_size: 16
57
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
58
  - lr_scheduler_type: constant
59
  - num_epochs: 1.0
60
 
61
  ### Resource Usage
62
+ Peak GPU Memory: 18.5713 GB
63
 
64
  ### Eval-Phase Metrics
65
  | step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | zhwikippl |
66
  | --- | --- | --- | --- | --- | --- | --- | --- | --- |
67
  | **teacher eval** | | 13.0697 | 11.6518 | | | | | 21.6262 |
68
+ | 0 | 0 | 182369.3906 | 181455.125 | 780.3353 | 91.2363 | 10.961 | 2.74 | 180761.6719 |
69
+ | 500 | 0.0808 | 86362.2812 | 101159.6328 | 54.5867 | 91.4066 | 10.94 | 2.735 | 232064.5 |
70
+ | 1000 | 0.1616 | 69487.6797 | 86111.1172 | 53.6604 | 91.4579 | 10.934 | 2.733 | 345169.5312 |
71
+ | 1500 | 0.2424 | 82681.5703 | 89315.8984 | 35.7168 | 91.5034 | 10.929 | 2.732 | 445378.8125 |
72
+ | 2000 | 0.3232 | 57160.3164 | 68079.1016 | 35.2824 | 91.7023 | 10.905 | 2.726 | 518689.375 |
73
+ | 2500 | 0.4040 | 51816.1094 | 64056.9492 | 35.1557 | 91.637 | 10.913 | 2.728 | 541196.6875 |
74
+ | 3000 | 0.4848 | 48757.2461 | 61247.7969 | 34.9901 | 91.7102 | 10.904 | 2.726 | 490950.375 |
75
+ | 3500 | 0.5657 | 46986.8359 | 60241.9648 | 34.8887 | 91.7509 | 10.899 | 2.725 | 494001.2812 |
76
+ | 4000 | 0.6465 | 45312.1992 | 58663.0117 | 34.8189 | 91.5035 | 10.929 | 2.732 | 481694.25 |
77
+ | 4500 | 0.7273 | 44666.1914 | 58929.5781 | 34.8044 | 91.6675 | 10.909 | 2.727 | 422349.9062 |
78
+ | 5000 | 0.8081 | 44634.5938 | 59004.7656 | 34.6895 | 91.5871 | 10.919 | 2.73 | 366104.7188 |
79
+ | 5500 | 0.8889 | 43732.4805 | 57270.8125 | 34.6591 | 91.6394 | 10.912 | 2.728 | 369470.4688 |
80
+ | 6000 | 0.9697 | 43799.5938 | 57096.7930 | 34.6125 | 91.6634 | 10.909 | 2.727 | 372021.5 |
81
+ | 6187 | 0.9999 | 44644.5977 | 57738.3242 | 34.5918 | 91.6837 | 10.907 | 2.727 | 350857.0312 |
82
 
83
  ### Framework versions
84
  - Distily 0.1.0
runs/Aug07_03-03-24_232a0f8c3879/events.out.tfevents.1723024479.232a0f8c3879 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a0e8669f12d57996a6f9eab8750c4dd4a27cc0172b70f94ab5c943101f5e1bc
3
+ size 520