metadata
library_name: transformers
license: apache-2.0
base_model: EleutherAI/pythia-160m
tags:
- generated_from_trainer
model-index:
- name: pythia_160m_sft
results: []
pythia_160m_sft
This model is a fine-tuned version of EleutherAI/pythia-160m on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.1033
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: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.3332 | 0.0889 | 100 | 2.2776 |
| 2.3071 | 0.1778 | 200 | 2.2564 |
| 2.2941 | 0.2667 | 300 | 2.2301 |
| 2.2548 | 0.3556 | 400 | 2.2094 |
| 2.2246 | 0.4444 | 500 | 2.1973 |
| 2.2189 | 0.5333 | 600 | 2.1705 |
| 2.181 | 0.6222 | 700 | 2.1498 |
| 2.1515 | 0.7111 | 800 | 2.1399 |
| 2.1333 | 0.8 | 900 | 2.1212 |
| 2.1139 | 0.8889 | 1000 | 2.1050 |
| 2.0778 | 0.9778 | 1100 | 2.0970 |
| 1.8312 | 1.0667 | 1200 | 2.1071 |
| 1.7405 | 1.1556 | 1300 | 2.1022 |
| 1.7284 | 1.2444 | 1400 | 2.1049 |
| 1.7554 | 1.3333 | 1500 | 2.1023 |
| 1.732 | 1.4222 | 1600 | 2.0934 |
| 1.7474 | 1.5111 | 1700 | 2.0917 |
| 1.7495 | 1.6 | 1800 | 2.0820 |
| 1.7449 | 1.6889 | 1900 | 2.0770 |
| 1.7474 | 1.7778 | 2000 | 2.0708 |
| 1.7447 | 1.8667 | 2100 | 2.0643 |
| 1.7138 | 1.9556 | 2200 | 2.0550 |
| 1.5662 | 2.0444 | 2300 | 2.1122 |
| 1.4506 | 2.1333 | 2400 | 2.1203 |
| 1.4282 | 2.2222 | 2500 | 2.1225 |
| 1.4302 | 2.3111 | 2600 | 2.1173 |
| 1.4471 | 2.4 | 2700 | 2.1156 |
| 1.4217 | 2.4889 | 2800 | 2.1168 |
| 1.428 | 2.5778 | 2900 | 2.1126 |
| 1.4206 | 2.6667 | 3000 | 2.1059 |
| 1.4315 | 2.7556 | 3100 | 2.1068 |
| 1.4345 | 2.8444 | 3200 | 2.1037 |
| 1.4034 | 2.9333 | 3300 | 2.1033 |
Framework versions
- Transformers 4.46.2
- Pytorch 2.5.1+cu121
- Datasets 3.1.0
- Tokenizers 0.20.3