File size: 2,550 Bytes
334e16a
 
 
 
 
 
 
02ba2fe
 
334e16a
 
536f323
334e16a
 
 
02ba2fe
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
334e16a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02ba2fe
334e16a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02ba2fe
334e16a
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
---
language:
- en
tags:
- llm-rs
- ggml
pipeline_tag: text-generation
datasets:
- the_pile
---

# GGML converted versions of [EleutherAI](https://huggingface.co/EleutherAI)'s GPT-J model

## Description

GPT-J 6B is a transformer model trained using Ben Wang's [Mesh Transformer JAX](https://github.com/kingoflolz/mesh-transformer-jax/). "GPT-J" refers to the class of model, while "6B" represents the number of trainable parameters.

<figure>

| Hyperparameter       | Value      |
|----------------------|------------|
| \\(n_{parameters}\\) | 6053381344 |
| \\(n_{layers}\\)     | 28&ast;    |
| \\(d_{model}\\)      | 4096       |
| \\(d_{ff}\\)         | 16384      |
| \\(n_{heads}\\)      | 16         |
| \\(d_{head}\\)       | 256        |
| \\(n_{ctx}\\)        | 2048       |
| \\(n_{vocab}\\)      | 50257/50400&dagger; (same tokenizer as GPT-2/3)  |
| Positional Encoding  | [Rotary Position Embedding (RoPE)](https://arxiv.org/abs/2104.09864) |
| RoPE Dimensions      | [64](https://github.com/kingoflolz/mesh-transformer-jax/blob/f2aa66e0925de6593dcbb70e72399b97b4130482/mesh_transformer/layers.py#L223) |
<figcaption><p><strong>&ast;</strong> Each layer consists of one feedforward block and one self attention block.</p>
<p><strong>&dagger;</strong> Although the embedding matrix has a size of 50400, only 50257 entries are used by the GPT-2 tokenizer.</p></figcaption></figure>

The model consists of 28 layers with a model dimension of 4096, and a feedforward dimension of 16384. The model
dimension is split into 16 heads, each with a dimension of 256. Rotary Position Embedding (RoPE) is applied to 64
dimensions of each head. The model is trained with a tokenization vocabulary of 50257, using the same set of BPEs as
GPT-2/GPT-3.


## Converted Models
$MODELS$

## Usage

### Python via [llm-rs](https://github.com/LLukas22/llm-rs-python):

#### Installation
Via pip: `pip install llm-rs`

#### Run inference
```python
from llm_rs import AutoModel

#Load the model, define any model you like from the list above as the `model_file`
model = AutoModel.from_pretrained("rustformers/gpt-j-ggml",model_file="gpt-j-6b-q4_0-ggjt.bin")

#Generate
print(model.generate("The meaning of life is"))
```

### Rust via [Rustformers/llm](https://github.com/rustformers/llm): 

#### Installation
```
git clone --recurse-submodules https://github.com/rustformers/llm.git
cd llm
cargo build --release
```

#### Run inference
```
cargo run --release -- gptj infer -m path/to/model.bin  -p "Tell me how cool the Rust programming language is:"
```