File size: 1,473 Bytes
07fa842
 
a783b08
 
 
 
 
 
 
 
07fa842
 
 
 
 
 
 
 
 
 
a783b08
07fa842
a783b08
07fa842
 
a783b08
 
07fa842
a783b08
07fa842
 
a783b08
07fa842
 
 
 
 
a783b08
 
07fa842
 
 
 
 
a783b08
07fa842
 
 
a783b08
07fa842
 
 
 
a783b08
07fa842
 
a783b08
07fa842
a783b08
07fa842
 
 
 
 
 
a783b08
07fa842
 
 
a783b08
07fa842
 
 
a783b08
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
---
library_name: transformers
tags:
- moe
- moah
license: apache-2.0
datasets:
- Locutusque/UltraTextbooks
language:
- en
---

# Model Card for Model ID

## Model Details

### Model Description

<!-- Provide a longer summary of what this model is. -->

This Model is a first test to combine [Jamba](https://huggingface.co/ai21labs/Jamba-v0.1) architecture with 1.58 bits linear layers and mixture of attention head.

The goal is to developpe and test if this kind of architectures have not too much quality loss for a fast inference.


- **Model type:** Mixture of attention head and mixture of expert 1.58bit linear layers 
- **License:** Apache licence 2.0

### Model Sources [optional]


- **Repository:** https://github.com/ostix360/optimized-LLM


## How to Get Started with the Model


If you want to test  this model please look at this repo at this [commit](https://github.com/ostix360/optimized-LLM/tree/8878e0f0bd764f85ce2ea56790a95f9837fb2fe4)


## Training Details

### Training Data

We use the first 100k data of Locutusque/UltraTextbooks to train this model

### Training Procedure

We use adam-8 bits with default betas and epsilon values

#### Preprocessing [optional]


The data fit the model max length i.e. 512 tokens


#### Training Hyperparameters

Please look at this file to see the hyperparameters


## Technical Specifications [optional]

### Compute Infrastructure



#### Hardware

- one 4070 ti GPU 

#### Software

- pytorch, transformers etc