File size: 1,891 Bytes
03081da
83e3e02
e1bcec7
 
 
 
1f4e8d6
03081da
 
e1bcec7
 
03081da
e1bcec7
03081da
e1bcec7
 
 
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
03081da
e1bcec7
 
 
 
 
 
 
 
 
 
 
03081da
e1bcec7
03081da
e1bcec7
 
 
 
 
 
 
 
 
 
 
 
03081da
 
e1bcec7
03081da
e1bcec7
 
 
 
 
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: other
base_model: meta-llama/Meta-Llama-3-8B-Instruct
model-index:
- name: LLAMA3-8BI-APPS
  results: []
library_name: peft
---

<!-- 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. -->

# LLAMA3-8BI-APPS

This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1490

## 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: 5e-06
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 1000

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 0.9027        | 0.1   | 100  | 0.9320          |
| 0.8632        | 0.2   | 200  | 0.9143          |
| 0.8572        | 0.3   | 300  | 1.0150          |
| 0.937         | 0.4   | 400  | 1.0545          |
| 1.0336        | 0.5   | 500  | 1.1029          |
| 1.0056        | 0.6   | 600  | 1.1267          |
| 1.0125        | 0.7   | 700  | 1.1307          |
| 1.028         | 0.8   | 800  | 1.1398          |
| 1.0692        | 0.9   | 900  | 1.1482          |
| 1.0361        | 1.0   | 1000 | 1.1490          |


### Framework versions

- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2