File size: 2,060 Bytes
a50cb1a
5f9eb1b
a50cb1a
b2d761b
 
 
 
 
0a8a6dd
b2d761b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7e628ed
 
9de792b
 
 
 
 
 
b92fabd
9de792b
 
b2d761b
 
 
 
 
 
 
 
 
 
 
 
 
 
5f9eb1b
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
---
{}
---

# XGen-7B-8K-Inst

Official research release for the family of **XGen** models (`7B`) by Salesforce AI Research:

*Title*: [Long Sequence Modeling with XGen: A 7B LLM Trained on 8K Input Sequence Length](https://blog.salesforceairesearch.com/xgen/)

## Models

### Base models
* [XGen-7B-4K-Base](https://huggingface.co/Salesforce/xgen-7b-4k-base): XGen-7B model pre-trained under 4K sequence length.
  * License: Apache-2.0
* [XGen-7B-8K-Base](https://huggingface.co/Salesforce/xgen-7b-8k-base): XGen-7B model pre-trained under 8K sequence length.
  * License: Apache-2.0

### Instruction-finetuned models

Supervised finetuned model on public domain instructional data. Released for ***research purpose*** only.

* [XGen-7B-8K-Inst](https://huggingface.co/Salesforce/xgen-7b-8k-inst)

## How to run

The training data for the models are tokenized with OpenAI Tiktoken library.
To use this model, install the package via `pip`:

```sh
pip install tiktoken
```

The models can be used as auto-regressive samplers as follows:

```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Salesforce/xgen-7b-8k-inst", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("Salesforce/xgen-7b-8k-inst", torch_dtype=torch.bfloat16)

header = (
    "A chat between a curious human and an artificial intelligence assistant. "
    "The assistant gives helpful, detailed, and polite answers to the human's questions.\n\n"
)
article = ""  # insert a document here
prompt = f"### Human: Please summarize the following article.\n\n{article}.\n###"

inputs = tokenizer(header + prompt, return_tensors="pt")
sample = model.generate(**inputs, max_length=128)
print(tokenizer.decode(sample[0]))
```

## Citation

```bibtex
@misc{XGen,
  title={Long Sequence Modeling with XGen: A 7B LLM Trained on 8K Input Sequence Length},
  author={Salesforce AI Research},
  howpublished={Salesforce AI Research Blog},
  year={2023},
  url={https://blog.salesforceairesearch.com/xgen-7b/}
}
```