End of training
Browse files
README.md
CHANGED
@@ -16,14 +16,14 @@ This student model is distilled from the teacher model [gpt2](https://huggingfac
|
|
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:
|
20 |
-
- eval_frwikippl:
|
21 |
-
- eval_zhwikippl:
|
22 |
-
- eval_tinystoriesppl:
|
23 |
-
- eval_loss:
|
24 |
-
- eval_runtime: 16.
|
25 |
-
- eval_samples_per_second: 59.
|
26 |
-
- eval_steps_per_second: 7.
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
29 |
should probably proofread and complete it, then remove this comment.
|
@@ -48,8 +48,8 @@ More information needed
|
|
48 |
The following hyperparameters were used during training:
|
49 |
- distillation_objective: DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl, layer_mapper=None, projector=None), hs_loss_component=LossComponent(label=hs, weight=0, loss_fn=None, layer_mapper=None, projector=None), attn_loss_component=LossComponent(label=attn, weight=0, loss_fn=None, layer_mapper=None, projector=None))
|
50 |
- train_embeddings: True
|
51 |
-
- learning_rate: 0.
|
52 |
-
- train_batch_size:
|
53 |
- eval_batch_size: 8
|
54 |
- seed: 42
|
55 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
@@ -58,26 +58,38 @@ The following hyperparameters were used during training:
|
|
58 |
- num_epochs: 1.0
|
59 |
|
60 |
### Resource Usage
|
61 |
-
Peak GPU Memory: 7.
|
62 |
|
63 |
### Eval-Phase Metrics
|
64 |
| step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | tinystoriesppl | zhwikippl |
|
65 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
|
66 |
| **teacher eval** | | 43.75 | 61.75 | | | | | 11.8125 | 19.125 |
|
67 |
-
| 0 | 0 |
|
68 |
-
| 1000 | 0.
|
69 |
-
| 2000 | 0.
|
70 |
-
| 3000 | 0.
|
71 |
-
| 4000 | 0.
|
72 |
-
| 5000 | 0.
|
73 |
-
| 6000 | 0.
|
74 |
-
| 7000 | 0.
|
75 |
-
| 8000 | 0.
|
76 |
-
| 9000 | 0.
|
77 |
-
| 10000 | 0.
|
78 |
-
| 11000 | 0.
|
79 |
-
| 12000 | 0.
|
80 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
81 |
|
82 |
### Framework versions
|
83 |
- Distily 0.2.0
|
|
|
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: 560.0
|
20 |
+
- eval_frwikippl: 644.0
|
21 |
+
- eval_zhwikippl: 488.0
|
22 |
+
- eval_tinystoriesppl: 284.0
|
23 |
+
- eval_loss: 0.6086
|
24 |
+
- eval_runtime: 16.7587
|
25 |
+
- eval_samples_per_second: 59.67
|
26 |
+
- eval_steps_per_second: 7.459
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
29 |
should probably proofread and complete it, then remove this comment.
|
|
|
48 |
The following hyperparameters were used during training:
|
49 |
- distillation_objective: DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl, layer_mapper=None, projector=None), hs_loss_component=LossComponent(label=hs, weight=0, loss_fn=None, layer_mapper=None, projector=None), attn_loss_component=LossComponent(label=attn, weight=0, loss_fn=None, layer_mapper=None, projector=None))
|
50 |
- train_embeddings: True
|
51 |
+
- learning_rate: 0.0001
|
52 |
+
- train_batch_size: 4
|
53 |
- eval_batch_size: 8
|
54 |
- seed: 42
|
55 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
|
|
58 |
- num_epochs: 1.0
|
59 |
|
60 |
### Resource Usage
|
61 |
+
Peak GPU Memory: 7.4226 GB
|
62 |
|
63 |
### Eval-Phase Metrics
|
64 |
| step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | tinystoriesppl | zhwikippl |
|
65 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
|
66 |
| **teacher eval** | | 43.75 | 61.75 | | | | | 11.8125 | 19.125 |
|
67 |
+
| 0 | 0 | 1408749273088.0 | 96207267430400.0 | 20.4380 | 16.6447 | 60.079 | 7.51 | 7482638336.0 | 43430709297152.0 |
|
68 |
+
| 1000 | 0.0404 | 1408.0 | 1432.0 | 0.8546 | 16.7128 | 59.835 | 7.479 | 788.0 | 1056.0 |
|
69 |
+
| 2000 | 0.0808 | 988.0 | 928.0 | 0.7631 | 16.6827 | 59.942 | 7.493 | 520.0 | 302.0 |
|
70 |
+
| 3000 | 0.1212 | 836.0 | 760.0 | 0.7155 | 16.633 | 60.121 | 7.515 | 402.0 | 196.0 |
|
71 |
+
| 4000 | 0.1616 | 732.0 | 676.0 | 0.6800 | 16.6836 | 59.939 | 7.492 | 378.0 | 157.0 |
|
72 |
+
| 5000 | 0.2020 | 676.0 | 668.0 | 0.6574 | 16.6514 | 60.055 | 7.507 | 322.0 | 227.0 |
|
73 |
+
| 6000 | 0.2424 | 648.0 | 732.0 | 0.6383 | 16.6833 | 59.94 | 7.493 | 286.0 | 190.0 |
|
74 |
+
| 7000 | 0.2828 | 612.0 | 632.0 | 0.6373 | 16.8106 | 59.486 | 7.436 | 286.0 | 169.0 |
|
75 |
+
| 8000 | 0.3232 | 588.0 | 704.0 | 0.6243 | 16.6588 | 60.028 | 7.504 | 266.0 | 596.0 |
|
76 |
+
| 9000 | 0.3636 | 560.0 | 644.0 | 0.6086 | 16.7587 | 59.67 | 7.459 | 284.0 | 488.0 |
|
77 |
+
| 10000 | 0.4040 | 532.0 | 564.0 | 0.5994 | 16.6696 | 59.989 | 7.499 | 256.0 | 142.0 |
|
78 |
+
| 11000 | 0.4444 | 544.0 | 628.0 | 0.5916 | 16.7004 | 59.879 | 7.485 | 252.0 | 153.0 |
|
79 |
+
| 12000 | 0.4848 | 540.0 | 612.0 | 0.5828 | 16.7602 | 59.665 | 7.458 | 252.0 | 568.0 |
|
80 |
+
| 13000 | 0.5253 | 528.0 | 612.0 | 0.5735 | 16.6596 | 60.025 | 7.503 | 260.0 | 160.0 |
|
81 |
+
| 14000 | 0.5657 | 528.0 | 576.0 | 0.5628 | 16.7207 | 59.806 | 7.476 | 246.0 | 250.0 |
|
82 |
+
| 15000 | 0.6061 | 478.0 | 524.0 | 0.5511 | 16.736 | 59.752 | 7.469 | 232.0 | 170.0 |
|
83 |
+
| 16000 | 0.6465 | 442.0 | 552.0 | 0.5270 | 16.7225 | 59.8 | 7.475 | 228.0 | 214.0 |
|
84 |
+
| 17000 | 0.6869 | 420.0 | 524.0 | 0.4692 | 16.6506 | 60.058 | 7.507 | 212.0 | 174.0 |
|
85 |
+
| 18000 | 0.7273 | 384.0 | 478.0 | 0.4115 | 16.7225 | 59.8 | 7.475 | 208.0 | 144.0 |
|
86 |
+
| 19000 | 0.7677 | 362.0 | 400.0 | 0.3610 | 16.6691 | 59.991 | 7.499 | 195.0 | 128.0 |
|
87 |
+
| 20000 | 0.8081 | 344.0 | 346.0 | 0.3370 | 16.6695 | 59.99 | 7.499 | 184.0 | 107.5 |
|
88 |
+
| 21000 | 0.8485 | 306.0 | 302.0 | 0.3061 | 16.7054 | 59.861 | 7.483 | 161.0 | 110.5 |
|
89 |
+
| 22000 | 0.8889 | 300.0 | 318.0 | 0.2974 | 16.6709 | 59.985 | 7.498 | 160.0 | 84.0 |
|
90 |
+
| 23000 | 0.9293 | 290.0 | 298.0 | 0.2890 | 16.7049 | 59.863 | 7.483 | 162.0 | 103.0 |
|
91 |
+
| 24000 | 0.9697 | 300.0 | 290.0 | 0.2970 | 16.6771 | 59.963 | 7.495 | 164.0 | 85.5 |
|
92 |
+
| 24750 | 1.0 | 280.0 | 290.0 | 0.2782 | 16.74 | 59.737 | 7.467 | 162.0 | 91.5 |
|
93 |
|
94 |
### Framework versions
|
95 |
- Distily 0.2.0
|
logs/dataset_subset=default, dataset_uri=distily_c4_multilingual_1M, learning_rate=0.0001, per_device_train_batch_size=4/events.out.tfevents.1724143970.02dbb11e2dcc
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4c5f2d8b25dfeef10e710f50dd3b67c11b795b907fed8a1442f9bc2d6103ea7c
|
3 |
+
size 312
|