Hugging Face's logo

distily
/
distily_multi_experiment

TensorBoard
Safetensors
Distily
gpt2
bitnet
1.58b
Generated from Trainer
Model card Files Metrics Community
lapp0 commited on
Commit
59d7179
·
verified ·
1 Parent(s): 92100a3

End of training

Browse files
Files changed (2) hide show
  1. README.md +29 -29
  2. logs/attn_loss_fn=cos, attn_weight=5, layer_mapper=last_k_2, projector=linear/events.out.tfevents.1724352430.e3f806ea38c9 +3 -0
README.md CHANGED
@@ -44,32 +44,32 @@ More information needed
44
  | step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | tinystoriesppl | zhwikippl |
45
  | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
46
  | **teacher eval** | | 43.25 | 61.25 | | | | | 11.6875 | 19.125 |
47
- | 0 | 0 | 2473901162496.0 | 170424302305280.0 | 45.7764 | 25.2689 | 98.936 | 12.387 | 4060086272.0 | 71468255805440.0 |
48
- | 2500 | 0.0404 | 2256.0 | 19712.0 | 20.8732 | 25.2001 | 99.206 | 12.421 | 1480.0 | 77312.0 |
49
- | 5000 | 0.0808 | 486.0 | 3104.0 | 18.4096 | 25.2957 | 98.831 | 12.374 | 338.0 | 1128.0 |
50
- | 7500 | 0.1212 | 276.0 | 1296.0 | 17.0014 | 25.2794 | 98.895 | 12.382 | 255.0 | 249.0 |
51
- | 10000 | 0.1616 | 202.0 | 756.0 | 16.2107 | 25.3199 | 98.736 | 12.362 | 187.0 | 296.0 |
52
- | 12500 | 0.2020 | 145.0 | 540.0 | 15.1004 | 25.2816 | 98.886 | 12.381 | 131.0 | 176.0 |
53
- | 15000 | 0.2424 | 123.5 | 488.0 | 14.5287 | 25.2154 | 99.146 | 12.413 | 93.5 | 146.0 |
54
- | 17500 | 0.2828 | 94.5 | 374.0 | 14.1501 | 25.2965 | 98.828 | 12.373 | 76.0 | 134.0 |
55
- | 20000 | 0.3232 | 78.5 | 306.0 | 13.6428 | 25.2525 | 99.0 | 12.395 | 63.25 | 149.0 |
56
- | 22500 | 0.3636 | 66.0 | 218.0 | 13.1496 | 25.2593 | 98.973 | 12.391 | 50.0 | 82.0 |
57
- | 25000 | 0.4040 | 63.5 | 202.0 | 12.9562 | 25.3101 | 98.775 | 12.367 | 44.0 | 74.0 |
58
- | 27500 | 0.4444 | 59.0 | 200.0 | 12.8429 | 25.2626 | 98.96 | 12.39 | 40.25 | 65.0 |
59
- | 30000 | 0.4848 | 58.25 | 198.0 | 12.8220 | 25.2928 | 98.843 | 12.375 | 39.75 | 61.25 |
60
- | 32500 | 0.5253 | 59.25 | 171.0 | 12.7662 | 25.3006 | 98.812 | 12.371 | 39.25 | 55.25 |
61
- | 35000 | 0.5657 | 58.25 | 171.0 | 12.6525 | 25.3008 | 98.811 | 12.371 | 36.5 | 47.25 |
62
- | 37500 | 0.6061 | 57.0 | 158.0 | 12.6054 | 25.3065 | 98.789 | 12.368 | 36.5 | 49.5 |
63
- | 40000 | 0.6465 | 55.5 | 159.0 | 12.5787 | 25.1987 | 99.212 | 12.421 | 33.5 | 64.5 |
64
- | 42500 | 0.6869 | 55.0 | 151.0 | 12.5195 | 25.2996 | 98.816 | 12.372 | 34.75 | 48.25 |
65
- | 45000 | 0.7273 | 50.75 | 135.0 | 12.2859 | 25.2872 | 98.864 | 12.378 | 29.25 | 43.5 |
66
- | 47500 | 0.7677 | 50.75 | 125.5 | 12.2428 | 25.1481 | 99.411 | 12.446 | 28.75 | 38.25 |
67
- | 50000 | 0.8081 | 50.25 | 123.5 | 12.2168 | 25.2928 | 98.842 | 12.375 | 28.375 | 37.25 |
68
- | 52500 | 0.8485 | 49.0 | 121.0 | 12.1935 | 25.3029 | 98.803 | 12.37 | 28.125 | 36.0 |
69
- | 55000 | 0.8889 | 49.0 | 121.0 | 12.1614 | 25.271 | 98.927 | 12.386 | 27.5 | 35.75 |
70
- | 57500 | 0.9293 | 48.75 | 119.5 | 12.1483 | 25.2276 | 99.098 | 12.407 | 27.5 | 34.25 |
71
- | 60000 | 0.9697 | 48.75 | 118.5 | 12.1410 | 25.3014 | 98.809 | 12.371 | 27.375 | 33.75 |
72
- | 61875 | 1.0 | 48.5 | 119.0 | 12.1404 | 25.2551 | 98.99 | 12.394 | 27.25 | 33.75 |
73
 
74
  # Resource Usage Comparison
75
 
@@ -103,7 +103,7 @@ Trained on 145,744,973 tokens from the [wikimedia/wikipedia](https://huggingface
103
  # Training Objective
104
 
105
  ```
106
- DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=25.0, loss_fn=cos, layer_mapper=layer-2))
107
  ```
108
 
109
  # Hyperparameters
@@ -120,9 +120,9 @@ The following hyperparameters were used during training:
120
  - lr_scheduler_type: `linear`
121
  - lr_scheduler_warmup_ratio: `0.5`
122
  - num_epochs: `1.0`
123
- - distillation_objective: `DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=25.0, loss_fn=cos, layer_mapper=layer-2))`
124
  - train_embeddings: `True`
125
- - lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7f148c79de10>`
126
  - student_model_name_or_path: `None`
127
  - student_config_name_or_path: `None`
128
  - student_model_config: `None`
 
44
  | step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | tinystoriesppl | zhwikippl |
45
  | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
46
  | **teacher eval** | | 43.25 | 61.25 | | | | | 11.6875 | 19.125 |
47
+ | 0 | 0 | 2473901162496.0 | 170424302305280.0 | 25.7744 | 25.1565 | 99.378 | 12.442 | 4060086272.0 | 71468255805440.0 |
48
+ | 2500 | 0.0404 | 956.0 | 7968.0 | 6.1178 | 25.2813 | 98.887 | 12.381 | 668.0 | 6432.0 |
49
+ | 5000 | 0.0808 | 380.0 | 1896.0 | 5.0307 | 25.2529 | 98.999 | 12.395 | 270.0 | 286.0 |
50
+ | 7500 | 0.1212 | 230.0 | 824.0 | 4.5130 | 25.1994 | 99.209 | 12.421 | 202.0 | 174.0 |
51
+ | 10000 | 0.1616 | 171.0 | 628.0 | 4.2264 | 25.2715 | 98.926 | 12.386 | 150.0 | 173.0 |
52
+ | 12500 | 0.2020 | 127.0 | 482.0 | 3.8535 | 25.2552 | 98.99 | 12.394 | 106.0 | 156.0 |
53
+ | 15000 | 0.2424 | 109.5 | 432.0 | 3.6651 | 25.2716 | 98.925 | 12.385 | 88.0 | 154.0 |
54
+ | 17500 | 0.2828 | 93.0 | 350.0 | 3.5204 | 25.3046 | 98.796 | 12.369 | 73.5 | 121.0 |
55
+ | 20000 | 0.3232 | 76.5 | 280.0 | 3.3349 | 25.2856 | 98.87 | 12.379 | 64.0 | 115.0 |
56
+ | 22500 | 0.3636 | 67.5 | 217.0 | 3.1514 | 25.2809 | 98.889 | 12.381 | 52.75 | 79.5 |
57
+ | 25000 | 0.4040 | 64.0 | 191.0 | 3.0794 | 25.2078 | 99.176 | 12.417 | 45.25 | 79.0 |
58
+ | 27500 | 0.4444 | 59.5 | 208.0 | 3.0351 | 25.2765 | 98.906 | 12.383 | 41.5 | 77.0 |
59
+ | 30000 | 0.4848 | 60.25 | 200.0 | 3.0180 | 25.2853 | 98.872 | 12.379 | 43.5 | 69.5 |
60
+ | 32500 | 0.5253 | 58.75 | 174.0 | 2.9977 | 25.2495 | 99.012 | 12.396 | 40.5 | 63.75 |
61
+ | 35000 | 0.5657 | 58.25 | 172.0 | 2.9422 | 25.2895 | 98.855 | 12.377 | 37.75 | 51.0 |
62
+ | 37500 | 0.6061 | 56.5 | 156.0 | 2.9150 | 25.2595 | 98.973 | 12.391 | 38.25 | 59.5 |
63
+ | 40000 | 0.6465 | 54.75 | 164.0 | 2.8974 | 25.2591 | 98.974 | 12.392 | 34.75 | 74.5 |
64
+ | 42500 | 0.6869 | 54.5 | 155.0 | 2.8849 | 25.2872 | 98.864 | 12.378 | 34.5 | 66.5 |
65
+ | 45000 | 0.7273 | 50.75 | 137.0 | 2.7764 | 25.258 | 98.978 | 12.392 | 30.875 | 39.75 |
66
+ | 47500 | 0.7677 | 50.25 | 126.0 | 2.7507 | 25.2765 | 98.906 | 12.383 | 29.5 | 39.0 |
67
+ | 50000 | 0.8081 | 49.5 | 124.5 | 2.7361 | 25.2105 | 99.165 | 12.415 | 28.75 | 39.0 |
68
+ | 52500 | 0.8485 | 48.5 | 120.0 | 2.7265 | 25.2508 | 99.007 | 12.396 | 29.125 | 35.5 |
69
+ | 55000 | 0.8889 | 48.0 | 117.5 | 2.7110 | 25.2721 | 98.923 | 12.385 | 28.375 | 33.75 |
70
+ | 57500 | 0.9293 | 47.5 | 118.0 | 2.7049 | 25.2774 | 98.903 | 12.383 | 28.0 | 32.25 |
71
+ | 60000 | 0.9697 | 47.5 | 117.0 | 2.7018 | 25.2647 | 98.952 | 12.389 | 27.875 | 32.0 |
72
+ | 61875 | 1.0 | 47.5 | 117.0 | 2.7011 | 25.2892 | 98.856 | 12.377 | 28.0 | 32.25 |
73
 
74
  # Resource Usage Comparison
75
 
 
103
  # Training Objective
104
 
105
  ```
106
+ DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=5, loss_fn=cos, layer_mapper=layer-2))
107
  ```
108
 
109
  # Hyperparameters
 
120
  - lr_scheduler_type: `linear`
121
  - lr_scheduler_warmup_ratio: `0.5`
122
  - num_epochs: `1.0`
123
+ - distillation_objective: `DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=5, loss_fn=cos, layer_mapper=layer-2))`
124
  - train_embeddings: `True`
125
+ - lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7f146e3bce80>`
126
  - student_model_name_or_path: `None`
127
  - student_config_name_or_path: `None`
128
  - student_model_config: `None`
logs/attn_loss_fn=cos, attn_weight=5, layer_mapper=last_k_2, projector=linear/events.out.tfevents.1724352430.e3f806ea38c9 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b19e4ae7471bab986319bee65ba2b3671fab2459c39b39ac3348a5053746ec38
3
+ size 588