File size: 5,570 Bytes
5d6fdcd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
103
104
105
106
107
108
109
110
111
112
113
---
library_name: peft
base_model: peiyi9979/math-shepherd-mistral-7b-prm
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: v3c_mistral_lora
  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. -->

# v3c_mistral_lora

This model is a fine-tuned version of [peiyi9979/math-shepherd-mistral-7b-prm](https://huggingface.co/peiyi9979/math-shepherd-mistral-7b-prm) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2939
- Accuracy: 0.8636
- Precision: 0.8421
- Recall: 0.6324
- F1: 0.7223

## 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: 8
- eval_batch_size: 8
- seed: 765837
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Accuracy | Precision | Recall | F1     |
|:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| No log        | 0      | 0    | 0.6026          | 0.7339   | 0.6       | 0.1542 | 0.2453 |
| 0.6391        | 0.0248 | 20   | 0.5954          | 0.7361   | 0.6119    | 0.1621 | 0.2562 |
| 0.5891        | 0.0495 | 40   | 0.5570          | 0.7550   | 0.6176    | 0.3320 | 0.4319 |
| 0.4606        | 0.0743 | 60   | 0.4962          | 0.7794   | 0.6667    | 0.4269 | 0.5205 |
| 0.4229        | 0.0990 | 80   | 0.4433          | 0.7905   | 0.6649    | 0.5099 | 0.5772 |
| 0.3836        | 0.1238 | 100  | 0.4297          | 0.8160   | 0.7605    | 0.5020 | 0.6048 |
| 0.3363        | 0.1485 | 120  | 0.3676          | 0.8381   | 0.7892    | 0.5771 | 0.6667 |
| 0.2483        | 0.1733 | 140  | 0.3537          | 0.8404   | 0.8225    | 0.5494 | 0.6588 |
| 0.2803        | 0.1980 | 160  | 0.3468          | 0.8415   | 0.8481    | 0.5296 | 0.6521 |
| 0.2782        | 0.2228 | 180  | 0.3493          | 0.8237   | 0.8310    | 0.4664 | 0.5975 |
| 0.2174        | 0.2475 | 200  | 0.3329          | 0.8492   | 0.8232    | 0.5889 | 0.6866 |
| 0.2965        | 0.2723 | 220  | 0.3314          | 0.8448   | 0.8343    | 0.5573 | 0.6682 |
| 0.2379        | 0.2970 | 240  | 0.3736          | 0.8149   | 0.8468    | 0.4150 | 0.5570 |
| 0.1587        | 0.3218 | 260  | 0.3315          | 0.8404   | 0.8609    | 0.5138 | 0.6436 |
| 0.1769        | 0.3465 | 280  | 0.3329          | 0.8370   | 0.8313    | 0.5257 | 0.6441 |
| 0.1984        | 0.3713 | 300  | 0.3211          | 0.8537   | 0.8712    | 0.5613 | 0.6827 |
| 0.2109        | 0.3960 | 320  | 0.3064          | 0.8570   | 0.8333    | 0.6126 | 0.7062 |
| 0.1961        | 0.4208 | 340  | 0.3035          | 0.8625   | 0.8413    | 0.6285 | 0.7195 |
| 0.2369        | 0.4455 | 360  | 0.2959          | 0.8747   | 0.8365    | 0.6877 | 0.7549 |
| 0.2355        | 0.4703 | 380  | 0.3176          | 0.8537   | 0.8380    | 0.5929 | 0.6944 |
| 0.1538        | 0.4950 | 400  | 0.3098          | 0.8503   | 0.8554    | 0.5613 | 0.6778 |
| 0.2261        | 0.5198 | 420  | 0.2964          | 0.8659   | 0.8235    | 0.6640 | 0.7352 |
| 0.1894        | 0.5446 | 440  | 0.3085          | 0.8625   | 0.8772    | 0.5929 | 0.7075 |
| 0.2089        | 0.5693 | 460  | 0.3103          | 0.8592   | 0.8621    | 0.5929 | 0.7026 |
| 0.225         | 0.5941 | 480  | 0.2933          | 0.8670   | 0.8519    | 0.6364 | 0.7285 |
| 0.2837        | 0.6188 | 500  | 0.2955          | 0.8636   | 0.8283    | 0.6482 | 0.7273 |
| 0.2046        | 0.6436 | 520  | 0.2943          | 0.8647   | 0.8429    | 0.6364 | 0.7252 |
| 0.1548        | 0.6683 | 540  | 0.3003          | 0.8636   | 0.8421    | 0.6324 | 0.7223 |
| 0.1626        | 0.6931 | 560  | 0.2982          | 0.8625   | 0.8603    | 0.6087 | 0.7130 |
| 0.2065        | 0.7178 | 580  | 0.2877          | 0.8636   | 0.8186    | 0.6601 | 0.7309 |
| 0.1423        | 0.7426 | 600  | 0.3031          | 0.8603   | 0.8757    | 0.5850 | 0.7014 |
| 0.1743        | 0.7673 | 620  | 0.2920          | 0.8659   | 0.8511    | 0.6324 | 0.7256 |
| 0.1281        | 0.7921 | 640  | 0.2912          | 0.8659   | 0.8474    | 0.6364 | 0.7269 |
| 0.1879        | 0.8168 | 660  | 0.2938          | 0.8625   | 0.8449    | 0.6245 | 0.7182 |
| 0.1741        | 0.8416 | 680  | 0.2965          | 0.8625   | 0.8486    | 0.6206 | 0.7169 |
| 0.1429        | 0.8663 | 700  | 0.2911          | 0.8647   | 0.8359    | 0.6443 | 0.7277 |
| 0.2218        | 0.8911 | 720  | 0.2950          | 0.8625   | 0.8449    | 0.6245 | 0.7182 |
| 0.1608        | 0.9158 | 740  | 0.2995          | 0.8603   | 0.8508    | 0.6087 | 0.7097 |
| 0.2056        | 0.9406 | 760  | 0.2967          | 0.8592   | 0.8424    | 0.6126 | 0.7094 |
| 0.2127        | 0.9653 | 780  | 0.2944          | 0.8625   | 0.8413    | 0.6285 | 0.7195 |
| 0.2252        | 0.9901 | 800  | 0.2939          | 0.8636   | 0.8421    | 0.6324 | 0.7223 |


### Framework versions

- PEFT 0.13.2
- Transformers 4.46.0
- Pytorch 2.5.1+cu124
- Datasets 3.1.0
- Tokenizers 0.20.3