instruction-pretrain
commited on
Commit
•
95649a1
1
Parent(s):
1849b49
Upload 7 files
Browse files- README.md +9 -11
- config.json +24 -0
- generation_config.json +6 -0
- special_tokens_map.json +5 -0
- tokenizer.json +0 -0
- tokenizer.model +3 -0
- tokenizer_config.json +42 -0
README.md
CHANGED
@@ -1,10 +1,7 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
---
|
4 |
# Instruction Pre-Training: Language Models are Supervised Multitask Learners
|
5 |
This repo contains the **context-based instruction synthesizer** used in our paper **Instruction Pre-Training: Language Models are Supervised Multitask Learners**.
|
6 |
|
7 |
-
|
8 |
|
9 |
<p align='center'>
|
10 |
<img src="./hf_intro.png" width="400">
|
@@ -71,14 +68,15 @@ for index, pair in enumerate(instruction_response_pairs):
|
|
71 |
print(f'## Instruction {index + 1}:\n{pair["Q"]}\n## Response {index + 1}:\n{pair["A"]}\n')
|
72 |
```
|
73 |
|
74 |
-
|
75 |
## Citation
|
76 |
If you find our work helpful, please cite us:
|
77 |
```bibtex
|
78 |
-
@
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
|
|
|
|
83 |
}
|
84 |
-
```
|
|
|
|
|
|
|
|
|
1 |
# Instruction Pre-Training: Language Models are Supervised Multitask Learners
|
2 |
This repo contains the **context-based instruction synthesizer** used in our paper **Instruction Pre-Training: Language Models are Supervised Multitask Learners**.
|
3 |
|
4 |
+
We explore supervised multitask pre-training by proposing ***Instruction Pre-Training***, a framework that scalably augments massive raw corpora with instruction-response pairs to pre-train language models. The instruction-response pairs are generated by an efficient instruction synthesizer built on open-source models. In our experiments, we synthesize 200M instruction-response pairs covering 40+ task categories to verify the effectiveness of *Instruction Pre-Training*. ***Instruction Pre-Training* outperforms *Vanilla Pre-training* in both general pre-training from scratch and domain-adaptive continued pre-training.** In pre-training from scratch, *Instruction Pre-Training* not only improves pre-trained base models but also benefits more from further instruction tuning. In continual pre-training, *Instruction Pre-Training* enables Llama3-8B to be comparable to or even outperform Llama3-70B.
|
5 |
|
6 |
<p align='center'>
|
7 |
<img src="./hf_intro.png" width="400">
|
|
|
68 |
print(f'## Instruction {index + 1}:\n{pair["Q"]}\n## Response {index + 1}:\n{pair["A"]}\n')
|
69 |
```
|
70 |
|
|
|
71 |
## Citation
|
72 |
If you find our work helpful, please cite us:
|
73 |
```bibtex
|
74 |
+
@inproceedings{
|
75 |
+
cheng2024adapting,
|
76 |
+
title={Adapting Large Language Models via Reading Comprehension},
|
77 |
+
author={Daixuan Cheng and Shaohan Huang and Furu Wei},
|
78 |
+
booktitle={The Twelfth International Conference on Learning Representations},
|
79 |
+
year={2024},
|
80 |
+
url={https://openreview.net/forum?id=y886UXPEZ0}
|
81 |
}
|
82 |
+
```
|
config.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"MistralForCausalLM"
|
4 |
+
],
|
5 |
+
"bos_token_id": 1,
|
6 |
+
"eos_token_id": 2,
|
7 |
+
"hidden_act": "silu",
|
8 |
+
"hidden_size": 4096,
|
9 |
+
"initializer_range": 0.02,
|
10 |
+
"intermediate_size": 14336,
|
11 |
+
"max_position_embeddings": 32768,
|
12 |
+
"model_type": "mistral",
|
13 |
+
"num_attention_heads": 32,
|
14 |
+
"num_hidden_layers": 32,
|
15 |
+
"num_key_value_heads": 8,
|
16 |
+
"rms_norm_eps": 1e-05,
|
17 |
+
"rope_theta": 10000.0,
|
18 |
+
"sliding_window": 4096,
|
19 |
+
"tie_word_embeddings": false,
|
20 |
+
"torch_dtype": "float16",
|
21 |
+
"transformers_version": "4.34.0.dev0",
|
22 |
+
"use_cache": true,
|
23 |
+
"vocab_size": 32000
|
24 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"transformers_version": "4.34.0.dev0"
|
6 |
+
}
|
special_tokens_map.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<s>",
|
3 |
+
"eos_token": "</s>",
|
4 |
+
"unk_token": "<unk>"
|
5 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
|
3 |
+
size 493443
|
tokenizer_config.json
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": false,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"added_tokens_decoder": {
|
5 |
+
"0": {
|
6 |
+
"content": "<unk>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": false,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false,
|
11 |
+
"special": true
|
12 |
+
},
|
13 |
+
"1": {
|
14 |
+
"content": "<s>",
|
15 |
+
"lstrip": false,
|
16 |
+
"normalized": false,
|
17 |
+
"rstrip": false,
|
18 |
+
"single_word": false,
|
19 |
+
"special": true
|
20 |
+
},
|
21 |
+
"2": {
|
22 |
+
"content": "</s>",
|
23 |
+
"lstrip": false,
|
24 |
+
"normalized": false,
|
25 |
+
"rstrip": false,
|
26 |
+
"single_word": false,
|
27 |
+
"special": true
|
28 |
+
}
|
29 |
+
},
|
30 |
+
"additional_special_tokens": [],
|
31 |
+
"bos_token": "<s>",
|
32 |
+
"clean_up_tokenization_spaces": false,
|
33 |
+
"eos_token": "</s>",
|
34 |
+
"legacy": true,
|
35 |
+
"model_max_length": 1000000000000000019884624838656,
|
36 |
+
"pad_token": null,
|
37 |
+
"sp_model_kwargs": {},
|
38 |
+
"spaces_between_special_tokens": false,
|
39 |
+
"tokenizer_class": "LlamaTokenizer",
|
40 |
+
"unk_token": "<unk>",
|
41 |
+
"use_default_system_prompt": false
|
42 |
+
}
|