End of training
Browse files
README.md
CHANGED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: gpt2
|
| 3 |
+
library_name: distily
|
| 4 |
+
license: mit
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
model-index:
|
| 8 |
+
- name: gpt2_model_card_distily_test
|
| 9 |
+
results: []
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# gpt2_model_card_distily_test
|
| 13 |
+
|
| 14 |
+
This student model is distilled from the teacher model [gpt2](https://huggingface.co/gpt2) using the dataset (unspecified).
|
| 15 |
+
|
| 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: 3305.2227
|
| 20 |
+
- eval_frwikippl: 13155.6016
|
| 21 |
+
- eval_zhwikippl: 73644.4141
|
| 22 |
+
- eval_loss: 2480.0
|
| 23 |
+
- eval_runtime: 0.0549
|
| 24 |
+
- eval_samples_per_second: 18.218
|
| 25 |
+
- eval_steps_per_second: 18.218
|
| 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.
|
| 29 |
+
|
| 30 |
+
## Model description
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Intended uses & limitations
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training and evaluation data
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
-->
|
| 42 |
+
|
| 43 |
+
## Training procedure
|
| 44 |
+
|
| 45 |
+
### Training hyperparameters
|
| 46 |
+
|
| 47 |
+
The following hyperparameters were used during training:
|
| 48 |
+
- distillation_strategy: logits_activations
|
| 49 |
+
- loss_fn: reverse_kl
|
| 50 |
+
- train_embeddings: True
|
| 51 |
+
- learning_rate: 0.0001
|
| 52 |
+
- train_batch_size: 1
|
| 53 |
+
- eval_batch_size: 2
|
| 54 |
+
- seed: 42
|
| 55 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 56 |
+
- lr_scheduler_type: cosine
|
| 57 |
+
- num_epochs: 1.0
|
| 58 |
+
|
| 59 |
+
### Resource Usage
|
| 60 |
+
Peak GPU Memory: 1.245255947113037GB
|
| 61 |
+
|
| 62 |
+
### Model Results
|
| 63 |
+
| epoch | eval_enwikippl | eval_frwikippl | eval_loss | eval_runtime | eval_samples_per_second | eval_steps_per_second | eval_zhwikippl | step |
|
| 64 |
+
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
|
| 65 |
+
| 0 | 58716.1836 | 59308.4531 | 6848.0 | 0.078 | 12.815 | 12.815 | 56780.0039 | 0 |
|
| 66 |
+
| 0.5025 | 2729.5120 | 12198.3467 | 2288.0 | 0.0559 | 17.885 | 17.885 | 86770.4375 | 100 |
|
| 67 |
+
| 0.7538 | 2523.9104 | 11865.6045 | 2240.0 | 0.0557 | 17.969 | 17.969 | 91730.1328 | 150 |
|
| 68 |
+
| 0.2513 | 3305.2227 | 13155.6016 | 2480.0 | 0.0549 | 18.218 | 18.218 | 73644.4141 | 50 |
|
| 69 |
+
|
| 70 |
+
### Framework versions
|
| 71 |
+
- Distily 0.1.0
|
| 72 |
+
- Transformers 4.43.3
|
| 73 |
+
- Pytorch 2.3.0
|
| 74 |
+
- Datasets 2.20.0
|