File size: 9,866 Bytes
7756256
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cef9067
7756256
 
 
 
 
 
 
c41b073
7756256
 
 
cef9067
 
 
c36667f
 
 
 
 
 
 
 
cef9067
7756256
 
 
c36667f
 
 
7756256
cef9067
c36667f
7756256
 
cef9067
c36667f
 
7756256
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
084d672
 
7756256
 
 
 
 
 
 
 
 
946e808
7756256
 
 
 
 
 
 
cef9067
7756256
 
 
 
 
 
 
 
cef9067
7756256
 
 
cef9067
7756256
 
 
 
 
 
 
 
cef9067
 
 
 
7756256
cef9067
 
7756256
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7155548
 
 
 
 
 
ebbcb23
7155548
 
 
 
 
 
 
 
7c2f422
 
7756256
 
cef9067
7756256
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ebebfbf
7756256
 
 
 
 
 
 
 
 
 
 
 
 
 
c41b073
 
7756256
 
cef9067
 
7756256
 
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
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
---
license: cc-by-nc-sa-4.0
datasets:
- camel-ai/code
- ehartford/wizard_vicuna_70k_unfiltered
- anon8231489123/ShareGPT_Vicuna_unfiltered
- teknium1/GPTeacher/roleplay-instruct-v2-final
- teknium1/GPTeacher/codegen-isntruct
- timdettmers/openassistant-guanaco
- camel-ai/math
- project-baize/baize-chatbot/medical_chat_data
- project-baize/baize-chatbot/quora_chat_data
- project-baize/baize-chatbot/stackoverflow_chat_data
- camel-ai/biology
- camel-ai/chemistry
- camel-ai/ai_society
- jondurbin/airoboros-gpt4-1.2
- LongConversations
- camel-ai/physics
tags:
- Composer
- MosaicML
- llm-foundry
inference: false
---

# MPT-30B-Chat

MPT-30B-Chat is a chatbot-like model for dialogue generation.
It was built by finetuning [MPT-30B](https://huggingface.co/mosaicml/mpt-30b) on the [ShareGPT-Vicuna](https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered), [Camel-AI](https://huggingface.co/camel-ai),
 [GPTeacher](https://github.com/teknium1/GPTeacher), [Guanaco](https://huggingface.co/datasets/timdettmers/openassistant-guanaco), [Baize](https://github.com/project-baize/baize-chatbot) and some generated datasets.
  * License: _CC-By-NC-SA-4.0_ (non-commercial use only)


This model was trained by [MosaicML](https://www.mosaicml.com) and follows a modified decoder-only transformer architecture.

## Model Date

June 22, 2023

## Model License

_CC-By-NC-SA-4.0_ (non-commercial use only)

## Documentation

* [Blog post: Raising the bar for open-source foundation models](https://www.mosaicml.com/blog/mpt-30b)
* [Codebase (mosaicml/llm-foundry repo)](https://github.com/mosaicml/llm-foundry/)
* Questions: Feel free to contact us via the [MosaicML Community Slack](https://mosaicml.me/slack)!

### Example Dialogue

**ksreenivasan**:
> I want you to come up with a tweet based on this summary of the article:
> 
> We are excited to extend the MosaicML Foundation Series with MPT-30B, a new commercially licensed open-source model that is significantly more powerful, and outperforms the
> original GPT-3. In addition, we are releasing two fine-tuned models, MPT-30B-Instruct and MPT-30B-Chat, that are built on top of MPT-30B and excel at short-form instruction
> following and multi-turn conversations, respectively.
> All MPT-30B models come with special features that differentiate them from other LLMs, including an 8k token context window (which can be further extended via finetuning;
> see MPT-7B-StoryWriter), support for context-length extrapolation via ALiBi, and efficient inference + training performance via FlashAttention. It also has strong coding
> abilities thanks to its pretraining data mixture.



**MPT-30B-Chat**:
> "Revolutionize your AI capabilities with MosaicML's latest addition, MPT-30B! Outperforming GPT-3, this open-source model is significantly more powerful and comes with
> finely-tuned models for exceptional instruction following and multi-turn conversations. Unleash the potential of an 8k token context window, ALiBi support, and FlashAttention.
> Experience the future of AI with MosaicML! #AI #ML #innovation"

**ksreenivasan**:
> Is there anything you want to say to MosaicML about this exciting news?

**MPT-30B-Chat**:
> As your AI marketing assistant, I would say that this is truly exciting news for the AI community! MosaicML's latest addition, MPT-30B, is poised to revolutionize the industry with its powerful open-source capabilities and finely-tuned models for instruction following and multi-turn conversations.
> The special features, including the 8k token context window, ALiBi support, and FlashAttention, are sure to set MosaicML's MPT-30B apart from other LLMs.
> Congratulations to the MosaicML team on this incredible achievement!

## How to Use

This model is best used with the MosaicML [llm-foundry repository](https://github.com/mosaicml/llm-foundry) for training and finetuning.

```python
import transformers
model = transformers.AutoModelForCausalLM.from_pretrained(
  'mosaicml/mpt-30b-chat',
  trust_remote_code=True
)
```
Note: This model requires that `trust_remote_code=True` be passed to the `from_pretrained` method.
This is because we use a custom `MPT` model architecture that is not yet part of the Hugging Face `transformers` package.
`MPT` includes options for many training efficiency features such as [FlashAttention](https://arxiv.org/pdf/2205.14135.pdf), [ALiBi](https://arxiv.org/abs/2108.12409), [QK LayerNorm](https://arxiv.org/abs/2010.04245), and more.

To use the optimized [triton implementation](https://github.com/openai/triton) of FlashAttention, you can load the model on GPU (`cuda:0`) with `attn_impl='triton'` and with `bfloat16` precision:
```python
import torch
import transformers

name = 'mosaicml/mpt-30b-chat'

config = transformers.AutoConfig.from_pretrained(name, trust_remote_code=True)
config.attn_config['attn_impl'] = 'triton'  # change this to use triton-based FlashAttention
config.init_device = 'cuda:0' # For fast initialization directly on GPU!

model = transformers.AutoModelForCausalLM.from_pretrained(
  name,
  config=config,
  torch_dtype=torch.bfloat16, # Load model weights in bfloat16
  trust_remote_code=True
)
```

The model was trained initially with a sequence length of 2048 with an additional pretraining stage for sequence length adapation up to 8192. However, ALiBi enables users to increase the maximum sequence length even further during finetuning and/or inference. For example:

```python
import transformers

name = 'mosaicml/mpt-30b-chat'

config = transformers.AutoConfig.from_pretrained(name, trust_remote_code=True)
config.max_seq_len = 16384 # (input + output) tokens can now be up to 16384

model = transformers.AutoModelForCausalLM.from_pretrained(
  name,
  config=config,
  trust_remote_code=True
)
```

This model was trained with the MPT-30B tokenizer which is based on the [EleutherAI/gpt-neox-20b](https://huggingface.co/EleutherAI/gpt-neox-20b) tokenizer and includes additional padding and eos tokens.

```python
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained('mosaicml/mpt-30b')
```

The model can then be used, for example, within a text-generation pipeline.  
Note: when running Torch modules in lower precision, it is best practice to use the [torch.autocast context manager](https://pytorch.org/docs/stable/amp.html).

```python
from transformers import pipeline

with torch.autocast('cuda', dtype=torch.bfloat16):
    inputs = tokenizer('Here is a recipe for vegan banana bread:\n', return_tensors="pt").to('cuda')
    outputs = model.generate(**inputs, max_new_tokens=100)
    print(tokenizer.batch_decode(outputs, skip_special_tokens=True))

# or using the HF pipeline
pipe = pipeline('text-generation', model=model, tokenizer=tokenizer, device='cuda:0')
with torch.autocast('cuda', dtype=torch.bfloat16):
    print(
        pipe('Here is a recipe for vegan banana bread:\n',
            max_new_tokens=100,
            do_sample=True,
            use_cache=True))
```

## Model Description

The architecture is a modification of a standard decoder-only transformer.

The model has been modified from a standard transformer in the following ways:
* It uses [FlashAttention](https://arxiv.org/pdf/2205.14135.pdf)
* It uses [ALiBi (Attention with Linear Biases)](https://arxiv.org/abs/2108.12409) and does not use positional embeddings
* It does not use biases


| Hyperparameter | Value |
|----------------|-------|
|n_parameters | 29.95B |
|n_layers | 48 |
| n_heads | 64 |
| d_model | 7168 |
| vocab size | 50432 |
| sequence length | 8192 |

## Data Mix

The model was trained on the following data mix:

| Data Source | Number of Tokens in Source | Proportion | 
|-------------|----------------------------|------------|
| Airoboros/GPT4-1.2 | 26.4M | 1.71% |
| Baize | 55.0M | 3.57% |
| Camel	| 301M | 19.54% | 
| GPTeacher	| 7.56M | 0.49% | 
| Guanaco | 15.6M | 1.02% | 
| LongCoversations | 18.4M | 1.19% | 
| ShareGPT | 821M | 53.24% |
| WizardLM | 297M | 19.23% |

"LongConversations" is a GPT3.5/4-generated dataset, details of which will be released at a later date.

### Training Configuration

This model was trained on 64 H100s for about 7.6 hours using the [MosaicML Platform](https://www.mosaicml.com/platform).
The model was trained with sharded data parallelism using [FSDP](https://pytorch.org/docs/stable/fsdp.html) and used the AdamW optimizer.

## Limitations and Biases

_The following language is modified from [EleutherAI's GPT-NeoX-20B](https://huggingface.co/EleutherAI/gpt-neox-20b)_

MPT-30B-Chat can produce factually incorrect output, and should not be relied on to produce factually accurate information.
MPT-30B-Chat was trained on various public datasets.
While great efforts have been taken to clean the pretraining data, it is possible that this model could generate lewd, biased or otherwise offensive outputs.

## Acknowledgements

This model was finetuned by Sam Havens and the MosaicML NLP team

## Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.


## MosaicML Platform

If you're interested in [training](https://www.mosaicml.com/training) and [deploying](https://www.mosaicml.com/inference) your own MPT or LLMs on the MosaicML Platform, [sign up here](https://forms.mosaicml.com/demo?utm_source=huggingface&utm_medium=referral&utm_campaign=mpt-7b).


## Citation

Please cite this model using the following format:

```
@online{MosaicML2023Introducing,
    author    = {MosaicML NLP Team},
    title     = {Introducing MPT-30B: Raising the bar
for open-source foundation models},
    year      = {2023},
    url       = {www.mosaicml.com/blog/mpt-30b},
    note      = {Accessed: 2023-06-22},
    urldate   = {2023-06-22}
}
```