metadata
license: apache-2.0
base_model: EleutherAI/pythia-410m
tags:
- generated_from_trainer
model-index:
- name: pythia410m-taylorswift-2000steps
results: []
pythia410m-taylorswift-2000steps
This model is a fine-tuned version of EleutherAI/pythia-410m on lamini/taylor_swift dataset. It achieves the following results on the evaluation set:
- Loss: 1.6837
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:
- learning_rate: 1e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: tpu
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1
- training_steps: 2000
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.8241 | 0.61 | 120 | 1.7255 |
| 1.0032 | 1.23 | 240 | 1.5268 |
| 1.141 | 1.84 | 360 | 1.4654 |
| 0.6062 | 2.45 | 480 | 1.4594 |
| 0.4412 | 3.07 | 600 | 1.4708 |
| 0.4085 | 3.68 | 720 | 1.5156 |
| 0.279 | 4.29 | 840 | 1.5411 |
| 0.2549 | 4.9 | 960 | 1.5015 |
| 0.2972 | 5.52 | 1080 | 1.5476 |
| 0.2389 | 6.13 | 1200 | 1.5731 |
| 0.2217 | 6.74 | 1320 | 1.5751 |
| 0.1253 | 7.36 | 1440 | 1.6184 |
| 0.2071 | 7.97 | 1560 | 1.6103 |
| 0.2004 | 8.58 | 1680 | 1.6501 |
| 0.1017 | 9.2 | 1800 | 1.6778 |
| 0.1549 | 9.81 | 1920 | 1.6837 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0