Transformers
PyTorch
Inference Endpoints
File size: 1,427 Bytes
fc72051
 
debca9a
 
 
 
 
 
fc72051
debca9a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
base_model: state-spaces/mamba-2.8b-slimpj
datasets:
- HuggingFaceH4/ultrachat_200k
model-index:
- name: mamba-2.8b-ultrachat
  results: []
---

# mamba-2.8b-ultrachat

This model is a fine-tuned version of [state-spaces/mamba-2.8b-slimpj](https://huggingface.co/state-spaces/mamba-2.8b-slimpj) on the [HuggingFaceH4/ultrachat_200k](https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k) dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1858

## 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: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 16
- total_train_batch_size: 512
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 1

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.0106        | 0.0   | 1    | 1.9092          |
| 1.1783        | 0.62  | 250  | 1.1858          |


### Framework versions

- Transformers 4.35.0
- Pytorch 2.1.1+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1