File size: 3,470 Bytes
140e8c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ea30ce0
140e8c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ea30ce0
140e8c5
 
 
 
 
 
 
 
 
 
 
 
 
487290b
 
 
 
 
 
 
 
e3cdada
 
 
 
 
 
 
 
f5a25e5
 
 
 
 
 
 
 
ea30ce0
 
 
 
 
 
 
 
140e8c5
 
 
 
 
 
 
 
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
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
---
base_model: meta-llama/Llama-2-7b-hf
tags:
- generated_from_trainer
model-index:
- name: sparse_llama_7b_hf_refined_web_50p_2024-03-24
  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. -->

# sparse_llama_7b_hf_refined_web_50p_2024-03-24

This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.1031

## 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: 0
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- total_eval_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 1000

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.1824        | 0.01  | 25   | 2.4333          |
| 2.1815        | 0.02  | 50   | 2.4313          |
| 2.2914        | 0.02  | 75   | 2.4244          |
| 2.2586        | 0.03  | 100  | 2.4192          |
| 2.3395        | 0.04  | 125  | 2.4108          |
| 2.1753        | 0.05  | 150  | 2.4039          |
| 2.1433        | 0.06  | 175  | 2.3947          |
| 2.3055        | 0.06  | 200  | 2.3859          |
| 2.2679        | 0.07  | 225  | 2.3842          |
| 2.2177        | 0.08  | 250  | 2.3817          |
| 2.1572        | 0.09  | 275  | 2.3830          |
| 2.1926        | 0.1   | 300  | 2.3829          |
| 2.2406        | 0.1   | 325  | 2.3817          |
| 2.21          | 0.11  | 350  | 2.3771          |
| 2.1296        | 0.12  | 375  | 2.3797          |
| 2.232         | 0.13  | 400  | 2.3764          |
| 2.2167        | 0.14  | 425  | 2.3746          |
| 2.18          | 0.14  | 450  | 2.3739          |
| 2.2508        | 0.15  | 475  | 2.3734          |
| 2.2584        | 0.16  | 500  | 2.3707          |
| 2.1665        | 0.17  | 525  | 2.3725          |
| 2.1627        | 0.18  | 550  | 2.3730          |
| 2.2769        | 0.18  | 575  | 2.3687          |
| 2.1621        | 0.19  | 600  | 2.3702          |
| 2.191         | 0.2   | 625  | 2.3696          |
| 2.274         | 0.21  | 650  | 2.3692          |
| 2.172         | 0.22  | 675  | 2.3720          |
| 2.1948        | 0.22  | 700  | 2.3704          |
| 2.2184        | 0.23  | 725  | 2.3699          |
| 2.1154        | 0.24  | 750  | 2.3693          |
| 2.1967        | 0.25  | 775  | 2.3699          |
| 2.2482        | 0.26  | 800  | 2.3668          |
| 2.1999        | 0.26  | 825  | 2.3679          |
| 2.155         | 0.27  | 850  | 2.3681          |
| 2.162         | 0.28  | 875  | 2.3651          |
| 2.1416        | 0.29  | 900  | 2.3676          |
| 2.3175        | 0.3   | 925  | 2.3686          |
| 2.2771        | 0.3   | 950  | 2.3667          |
| 2.2253        | 0.31  | 975  | 2.3639          |
| 2.1176        | 0.32  | 1000 | 2.3649          |


### Framework versions

- Transformers 4.36.2
- Pytorch 2.1.1+cu121
- Datasets 2.15.0
- Tokenizers 0.15.2