File size: 4,502 Bytes
a6814b6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: mistralai/Mistral-7B-Instruct-v0.3
datasets:
- generator
metrics:
- bleu
- rouge
model-index:
- name: Mistral-7B-Instruct-v0.3-advisegpt-v0.1
  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. -->

# Mistral-7B-Instruct-v0.3-advisegpt-v0.1

This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.3](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3) on the generator dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0762
- Bleu: {'bleu': 0.9585152983456746, 'precisions': [0.9779106264925121, 0.9626412004947897, 0.951895206199588, 0.9430042745426802], 'brevity_penalty': 0.999729358235579, 'length_ratio': 0.9997293948524543, 'translation_length': 1289353, 'reference_length': 1289702}
- Rouge: {'rouge1': 0.9761616288162651, 'rouge2': 0.9590944581779459, 'rougeL': 0.9748018206627191, 'rougeLsum': 0.9758991028742771}

## 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: 2e-05
- train_batch_size: 3
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 10
- total_train_batch_size: 30
- 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 | Bleu                                                                                                                                                                                                                                                                | Validation Loss | Rouge                                                                                                                       |
|:-------------:|:------:|:----:|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:---------------:|:---------------------------------------------------------------------------------------------------------------------------:|
| 0.0675        | 0.9998 | 809  | {'bleu': 0.9495787144110293, 'brevity_penalty': 0.9993236461934566, 'length_ratio': 0.9993238748175935, 'precisions': [0.9735806894625358, 0.9547064588389024, 0.9417775802515341, 0.9313417436570839], 'reference_length': 1289702, 'translation_length': 1288830} | 0.0936          | {'rouge1': 0.9713550622229471, 'rouge2': 0.9502301622796694, 'rougeL': 0.9691228372678113, 'rougeLsum': 0.9708685856330016} |
| 0.0548        | 1.9998 | 1618 | 0.0771                                                                                                                                                                                                                                                              | {'bleu': 0.9571495232321637, 'precisions': [0.9773150683231548, 0.9614637013631232, 0.95029828201407, 0.9410538932261553], 'brevity_penalty': 0.9996991100504438, 'length_ratio': 0.9996991553087458, 'translation_length': 1289314, 'reference_length': 1289702}| {'rouge1': 0.9755343391324649, 'rouge2': 0.9577790978374392, 'rougeL': 0.9740177474237091, 'rougeLsum': 0.9752585254668996} |
| 0.0439        | 2.9995 | 2427 | 0.0762                                                                                                                                                                                                                                                              | {'bleu': 0.9585152983456746, 'precisions': [0.9779106264925121, 0.9626412004947897, 0.951895206199588, 0.9430042745426802], 'brevity_penalty': 0.999729358235579, 'length_ratio': 0.9997293948524543, 'translation_length': 1289353, 'reference_length': 1289702}| {'rouge1': 0.9761616288162651, 'rouge2': 0.9590944581779459, 'rougeL': 0.9748018206627191, 'rougeLsum': 0.9758991028742771} |


### Framework versions

- PEFT 0.10.0
- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1