---
base_model: gpt2
library_name: Distily
license: mit
tags:
- generated_from_trainer
model-index:
- name: distily_bench_obj_cross_v2.13_gpt2
  results: []
---

# distily_bench_obj_cross_v2.13_gpt2

This student model is distilled from the teacher model [gpt2](https://huggingface.co/gpt2) using the dataset (unspecified).

The [Distily](https://github.com/lapp0/distily) library was used for this distillation.

It achieves the following results on the evaluation set:
- eval_enwikippl: 1376.0
- eval_frwikippl: 5856.0
- eval_zhwikippl: 111104.0
- eval_tinystoriesppl: 964.0
- eval_loss: 3.1072
- eval_runtime: 12.9331
- eval_samples_per_second: 46.392
- eval_steps_per_second: 11.598

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment.

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed
-->

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- 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=1.0, loss_fn=cos, layer_mapper=last, projector=None), attn_loss_component=LossComponent(label=attn, weight=0, loss_fn=None, layer_mapper=None, projector=None))
- train_embeddings: True
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.5
- num_epochs: 1.0

### Resource Usage
Peak GPU Memory: 8.0905 GB

### Eval-Phase Metrics
| step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | tinystoriesppl | zhwikippl |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| **teacher eval** |  | 43.75 | 61.75 |  |  |  |  | 11.8125 | 19.125 |
| 0 | 0 | 1821066133504.0 | 158329674399744.0 | 20.2008 | 12.9107 | 46.473 | 11.618 | 12079595520.0 | 98956046499840.0 |
| 750 | 0.1010 | 1376.0 | 5856.0 | 3.1072 | 12.9331 | 46.392 | 11.598 | 964.0 | 111104.0 |
| 1500 | 0.2020 | 580.0 | 3600.0 | 2.2189 | 12.9266 | 46.416 | 11.604 | 438.0 | 1020.0 |
| 2250 | 0.3030 | 376.0 | 1904.0 | 1.9249 | 12.9295 | 46.405 | 11.601 | 312.0 | 374.0 |
| 3000 | 0.4040 | 268.0 | 1080.0 | 1.6655 | 12.9091 | 46.479 | 11.62 | 238.0 | 208.0 |
| 3750 | 0.5051 | 211.0 | 732.0 | 1.4810 | 12.9336 | 46.391 | 11.598 | 172.0 | 217.0 |
| 4500 | 0.6061 | 167.0 | 580.0 | 1.2985 | 12.9202 | 46.439 | 11.61 | 143.0 | 146.0 |
| 5250 | 0.7071 | 135.0 | 486.0 | 1.1339 | 12.9225 | 46.431 | 11.608 | 112.5 | 133.0 |
| 6000 | 0.8081 | 124.5 | 452.0 | 1.0647 | 12.9107 | 46.473 | 11.618 | 101.5 | 125.5 |
| 6750 | 0.9091 | 118.5 | 436.0 | 1.0324 | 12.9153 | 46.456 | 11.614 | 97.5 | 120.0 |
| 7425 | 1.0 | 117.5 | 432.0 | 1.0264 | 13.1576 | 45.601 | 11.4 | 95.5 | 118.5 |

### Framework versions
- Distily 0.2.0
- Transformers 4.44.0
- Pytorch 2.3.0
- Datasets 2.21.0