File size: 1,797 Bytes
1087326 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 |
---
license: mit
library_name: peft
tags:
- generated_from_trainer
base_model: EleutherAI/gpt-neo-125m
metrics:
- accuracy
model-index:
- name: mod_dep_fl-PEFT_mix-instruct
results: []
---
<!-- 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. -->
# mod_dep_fl-PEFT_mix-instruct
This model is a fine-tuned version of [EleutherAI/gpt-neo-125m](https://huggingface.co/EleutherAI/gpt-neo-125m) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.5502
- Accuracy: -1701.4273
## 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: 0.005
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:------:|:----:|:---------------:|:----------:|
| 4.5917 | 0.9524 | 10 | 2.8083 | -1780.3744 |
| 3.7289 | 2.0 | 21 | 2.6099 | -1691.0399 |
| 3.548 | 2.9524 | 31 | 2.5788 | -1751.4823 |
| 3.5153 | 4.0 | 42 | 2.5515 | -1766.1752 |
| 3.4735 | 4.7619 | 50 | 2.5502 | -1701.4273 |
### Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1 |