File size: 3,543 Bytes
c137651
6c0d620
 
c137651
 
 
 
 
58d6d5b
c137651
 
 
 
 
 
58d6d5b
c137651
6c0d620
c137651
58d6d5b
c137651
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
58d6d5b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c137651
 
 
 
58d6d5b
c137651
58d6d5b
c137651
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
102
---
license: apache-2.0
base_model: mistralai/Mistral-7B-Instruct-v0.2
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: UTI_M2_1000steps_1e6rate_SFT
  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. -->

# UTI_M2_1000steps_1e6rate_SFT

This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.7960

## 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-06
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- 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 |
|:-------------:|:-------:|:----:|:---------------:|
| 2.2167        | 0.3333  | 25   | 1.1865          |
| 0.9806        | 0.6667  | 50   | 0.9618          |
| 0.936         | 1.0     | 75   | 0.9371          |
| 0.8294        | 1.3333  | 100  | 0.9512          |
| 0.8273        | 1.6667  | 125  | 0.9369          |
| 0.7851        | 2.0     | 150  | 0.9036          |
| 0.5263        | 2.3333  | 175  | 0.9990          |
| 0.5512        | 2.6667  | 200  | 0.9589          |
| 0.5272        | 3.0     | 225  | 0.9576          |
| 0.2888        | 3.3333  | 250  | 1.1371          |
| 0.2968        | 3.6667  | 275  | 1.1164          |
| 0.3381        | 4.0     | 300  | 1.1144          |
| 0.1802        | 4.3333  | 325  | 1.1697          |
| 0.2025        | 4.6667  | 350  | 1.1946          |
| 0.2273        | 5.0     | 375  | 1.2614          |
| 0.1417        | 5.3333  | 400  | 1.3260          |
| 0.1524        | 5.6667  | 425  | 1.3343          |
| 0.136         | 6.0     | 450  | 1.3735          |
| 0.117         | 6.3333  | 475  | 1.3843          |
| 0.1284        | 6.6667  | 500  | 1.3742          |
| 0.1172        | 7.0     | 525  | 1.4114          |
| 0.0905        | 7.3333  | 550  | 1.5000          |
| 0.1027        | 7.6667  | 575  | 1.5142          |
| 0.097         | 8.0     | 600  | 1.4912          |
| 0.0837        | 8.3333  | 625  | 1.5974          |
| 0.0832        | 8.6667  | 650  | 1.6185          |
| 0.0781        | 9.0     | 675  | 1.6203          |
| 0.0698        | 9.3333  | 700  | 1.6833          |
| 0.0722        | 9.6667  | 725  | 1.6960          |
| 0.0681        | 10.0    | 750  | 1.7139          |
| 0.0635        | 10.3333 | 775  | 1.7732          |
| 0.0654        | 10.6667 | 800  | 1.7704          |
| 0.0663        | 11.0    | 825  | 1.7647          |
| 0.0604        | 11.3333 | 850  | 1.7840          |
| 0.0628        | 11.6667 | 875  | 1.7916          |
| 0.0627        | 12.0    | 900  | 1.7947          |
| 0.061         | 12.3333 | 925  | 1.7962          |
| 0.062         | 12.6667 | 950  | 1.7967          |
| 0.0607        | 13.0    | 975  | 1.7960          |
| 0.0605        | 13.3333 | 1000 | 1.7960          |


### Framework versions

- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1