File size: 1,577 Bytes
9c1f593
23d7d18
 
 
 
 
 
 
9c1f593
 
23d7d18
 
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
9c1f593
23d7d18
 
 
 
 
 
 
 
 
 
 
9c1f593
23d7d18
9c1f593
83478ac
9c1f593
83478ac
9c1f593
23d7d18
9c1f593
23d7d18
 
 
 
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
---
license: bigscience-bloom-rail-1.0
base_model: alonzogarbanzo/Bloom-1b7-winograd-wsc-IT-baseline
tags:
- generated_from_trainer
model-index:
- name: Bloom-1b7-ropes-Cont-IT-Step2
  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. -->

# Bloom-1b7-ropes-Cont-IT-Step2

This model is a fine-tuned version of [alonzogarbanzo/Bloom-1b7-winograd-wsc-IT-baseline](https://huggingface.co/alonzogarbanzo/Bloom-1b7-winograd-wsc-IT-baseline) on an unknown dataset.

## 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: 5e-05
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

Final Results: {'loss': 0.0261, 'grad_norm': 1.9494764804840088, 'learning_rate': 3.0000000000000004e-07, 'epoch': 10.0}

Average Results: {'train_runtime': 858.2936, 'train_samples_per_second': 2.33, 'train_steps_per_second': 0.583, 'train_loss': 0.4610937827527523, 'epoch': 10.0}

### Framework versions

- Transformers 4.38.1
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2