File size: 2,024 Bytes
4f2c4d0 |
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 |
---
library_name: peft
tags:
- generated_from_trainer
datasets:
- patent-classification
metrics:
- accuracy
base_model: NousResearch/Llama-2-7b-hf
model-index:
- name: llama-2-7b-flash-attention2-lora-patent-classification
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. -->
# llama-2-7b-flash-attention2-lora-patent-classification
This model is a fine-tuned version of [NousResearch/Llama-2-7b-hf](https://huggingface.co/NousResearch/Llama-2-7b-hf) on the patent-classification dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5598
- Accuracy: 0.436
- Precision Macro: 0.4276
- Recall Macro: 0.3658
- F1-score Macro: 0.3707
## 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: 0.0002
- train_batch_size: 4
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision Macro | Recall Macro | F1-score Macro |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:---------------:|:------------:|:--------------:|
| 1.4059 | 1.0 | 6250 | 1.9046 | 0.3748 | 0.3815 | 0.3173 | 0.3012 |
| 1.1153 | 2.0 | 12500 | 1.6457 | 0.419 | 0.4162 | 0.3461 | 0.3466 |
| 1.0234 | 3.0 | 18750 | 1.5598 | 0.436 | 0.4276 | 0.3658 | 0.3707 |
### Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.2.dev0
- Tokenizers 0.15.0 |