File size: 1,996 Bytes
3147092
9dbe82c
3147092
 
9dbe82c
 
3147092
 
 
 
 
 
 
 
 
 
9dbe82c
 
 
 
3147092
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9dbe82c
0635e22
3147092
 
 
0635e22
3147092
9dbe82c
 
3147092
 
 
 
9dbe82c
 
 
 
 
 
 
 
 
 
 
 
3147092
 
 
 
0635e22
3147092
 
9dbe82c
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
---
base_model: cosmin/falcon-7b-sharded-bf16
tags:
- generated_from_trainer
metrics:
- f1
model-index:
- name: falcon-7b-sharded
  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. -->

# falcon-7b-sharded

This model is a fine-tuned version of [cosmin/falcon-7b-sharded-bf16](https://huggingface.co/cosmin/falcon-7b-sharded-bf16) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.3060
- F1: 0.0027

## 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: 0.0001
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 500

### Training results

| Training Loss | Epoch | Step | Validation Loss | F1     |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 3.0776        | 1.0   | 55   | 2.8379          | 0.0027 |
| 2.3055        | 1.99  | 110  | 2.5165          | 0.0220 |
| 2.3104        | 2.99  | 165  | 2.4452          | 0.0027 |
| 2.1221        | 4.0   | 221  | 2.3845          | 0.0027 |
| 2.2114        | 5.0   | 276  | 2.3660          | 0.0151 |
| 2.0432        | 5.99  | 331  | 2.3325          | 0.0124 |
| 2.0811        | 6.99  | 386  | 2.3185          | 0.0027 |
| 2.0372        | 8.0   | 442  | 2.3066          | 0.0027 |
| 2.019         | 9.0   | 497  | 2.3058          | 0.0027 |
| 2.0906        | 9.05  | 500  | 2.3060          | 0.0027 |


### Framework versions

- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.1