File size: 5,135 Bytes
bd6e8b2
86b46ed
 
 
 
 
 
 
 
 
bd6e8b2
86b46ed
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
---
language:
- en
tags:
- text generation
- pytorch
- causal-lm
license: cc-by-4.0
pipeline_tag: text-generation
library_name: transformers
---
# Writer-5B

<style>
img {
 display: inline;
}
</style>

|[![Model architecture](https://img.shields.io/badge/Model%20Arch-Transformer%20Decoder-green)](#model-architecture)|[![Model size](https://img.shields.io/badge/Params-5B-green)](#model-architecture)|[![Language](https://img.shields.io/badge/Language-en--US-lightgrey#model-badge)](#datasets)


## Model Description

Model description
Writer LLM base was primarily pretrained with English text, there is still a trace amount of non-English data present within the training corpus that was accessed through CommonCrawl. A causal language modeling (CLM) objective was utilized during the process of the model's pretraining. Similar to GPT-3, Writer LLM base is a member of the same family of models that only contain a decoder. As a result, it was pretrained utilizing the objective of self-supervised causal language modeling.
Writer LLM base uses the prompts and general experimental setup from GPT-3 in order to conduct its evaluation in accordance with GPT-3. Read the official paper if you want more information about this.

## Getting started

### Step 1: Install NeMo and dependencies

You will need to install NVIDIA Apex and NeMo. 

```
git clone https://github.com/ericharper/apex.git
cd apex
git checkout nm_v1.11.0
pip install -v --disable-pip-version-check --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" --global-option="--fast_layer_norm" --global-option="--distributed_adam" --global-option="--deprecated_fused_adam" ./
```

```
pip install nemo_toolkit['nlp']==1.11.0
``` 

### Step 2: Launch eval server 

**Note.** The example below launches a model variant with Tensor Parallelism (TP) of 4 and Pipeline Parallelism (PP) of 1 on two GPUs.


```
git clone https://github.com/NVIDIA/NeMo.git 
cd NeMo/examples/nlp/language_modeling
git checkout v1.11.0
python megatron_gpt_eval.py gpt_model_file=palmyara_gpt_5b.nemo server=True tensor_model_parallel_size=4 trainer.devices=4
```

### Step 3: Send prompts to your model!
```python
import json
import requests

port_num = 5555
headers = {"Content-Type": "application/json"}

def request_data(data):
    resp = requests.put('http://localhost:{}/generate'.format(port_num),
                        data=json.dumps(data),
                        headers=headers)
    sentences = resp.json()['sentences']
    return sentences


data = {
    "sentences": ["Tell me an interesting fact about space travel."]*1,
    "tokens_to_generate": 50,
    "temperature": 1.0,
    "add_BOS": True,
    "top_k": 0,
    "top_p": 0.9,
    "greedy": False,
    "all_probs": False,
    "repetition_penalty": 1.2,
    "min_tokens_to_generate": 2,
}

sentences = request_data(data)
print(sentences)
```


## Training Data

|    part        | MassiveText (sampling) | tokens (B) | sampling ratio |
|:---------------|-----------------------:|:----------:| --------------:|
| mc4 filtered   |   MassiveWeb (48%)     |   1331     |   58%          |
| TrustedWeb     |   -                    |    -       |   -            |
| realnews       |   News (10%)           |    21      |   10%          |
| c4             |   c4  (10%)            |    -       |   -            |
| wikipedia-40B  |   wikipedia  (2%)      |    2       |   5%           |
| github         |   github  (3%)         |    -       |   -            |
| books          |   books (27%)          |    24      |   27%          |
| youtube        |   -                    |    -       |   -            |



## Evaluation results

*Zero-shot performance.* Evaluated using [LM Evaluation Test Suite from AI21](https://github.com/AI21Labs/lm-evaluation)

| ARC-Challenge	| ARC-Easy | RACE-middle | RACE-high | Winogrande | RTE | BoolQA | HellaSwag | PiQA |
| ------------- | -------- | ----------- | --------- | ---------- | --- | ------ | --------- | ---- |
| 0.3976        | 0.5566  | 0.5007       | 0.4171    | 0.6133     | 0.5812 | 0.6356 | 0.6298 | 0.7492 | 

## Limitations

The model was trained on the data originally crawled from the Internet. This data contains toxic language and societal biases. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts.   

## References

[1] [Improving Language Understanding by Generative Pre-Training](https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf)

[2] [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/pdf/1909.08053.pdf)

[3] [NVIDIA NeMo Toolkit](https://github.com/NVIDIA/NeMo)

[4] [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://arxiv.org/abs/2101.00027)

## Licence

License to use this model is covered by the [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/). By downloading the public and release version of the model, you accept the terms and conditions of the [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/) license.