zihanliu commited on
Commit
5ec6c8d
·
verified ·
1 Parent(s): 7a443cd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +95 -3
README.md CHANGED
@@ -1,3 +1,95 @@
1
- ---
2
- license: cc-by-nc-nd-4.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ language:
4
+ - en
5
+ pipeline_tag: text-generation
6
+ tags:
7
+ - nvidia
8
+ - AceInstruct
9
+ - code
10
+ - math
11
+ - general_domain
12
+ - instruct_model
13
+ - pytorch
14
+ ---
15
+
16
+ ## Introduction
17
+ We introduce AceInstruct, a family of advanced SFT models for coding, mathematics, and general-purpose tasks. The AceInstruct family, which includes AceInstruct-1.5B, 7B, and 72B, is <b>Improved using Qwen</b>.
18
+ These models are fine-tuned on Qwen2.5-Base using [general SFT datasets](https://huggingface.co/datasets/nvidia/AceMath-Instruct-Training-Data). These same datasets are also used in the training of [AceMath-Instruct](https://huggingface.co/nvidia/AceMath-72B-Instruct). Different from AceMath-Instruct which is specialized for math questions, AceInstruct is versatile and can be applied to a wide range of domains. Benchmark evaluations across coding, mathematics, and general knowledge tasks demonstrate that AceInstruct delivers performance comparable to Qwen2.5-Instruct.
19
+
20
+ For more information about AceInstruct, check our [website](https://research.nvidia.com/labs/adlr/acemath/) and [paper](https://arxiv.org/abs/2412.15084).
21
+
22
+
23
+ ## Benchmark Results
24
+ | | Qwen2.5-1.5B-Instruct | AceInstruct-1.5B | Qwen2.5-7B-Instruct | AceInstruct-7B | Qwen2.5-72B-Instruct | AceInstruct-72B |
25
+ |-----------|-------|-------|-------|-------|-------|-------|
26
+ | HumanEval | 61.60 | 73.17 | 84.80 | 85.37 | 86.60 | 89.63 |
27
+ | MBPP | 63.20 | 65.76 | 79.20 | 74.32 | 88.20 | 83.66 |
28
+ | GSM8K | 73.20 | 80.44 | 91.60 | 93.10 | 95.80 | 96.36 |
29
+ | MATH | 55.20 | 60.34 | 75.50 | 76.40 | 83.10 | 84.50 |
30
+ | MMLU | 58.37 | 58.17 | 74.51 | 74.68 | 84.67 | 83.88 |
31
+ | MMLU Pro | 32.40 | 33.78 | 56.30 | 54.50 | 71.10 | 66.10 |
32
+ | Average | 57.33 | 61.94 | 76.99 | 76.40 | 84.91 | 84.02 |
33
+
34
+ ## All Resources
35
+ ### AceMath Instruction Models
36
+ - [AceMath-1.5B-Instruct](https://huggingface.co/nvidia/AceMath-1.5B-Instruct), [AceMath-7B-Instruct](https://huggingface.co/nvidia/AceMath-7B-Instruct), [AceMath-72B-Instruct](https://huggingface.co/nvidia/AceMath-72B-Instruct)
37
+
38
+ ### AceMath Reward Models
39
+ - [AceMath-7B-RM](https://huggingface.co/nvidia/AceMath-7B-RM), [AceMath-72B-RM](https://huggingface.co/nvidia/AceMath-72B-RM)
40
+
41
+ ### Evaluation & Training Data
42
+ - [AceMath-RewardBench](https://huggingface.co/datasets/nvidia/AceMath-RewardBench), [AceMath-Instruct Training Data](https://huggingface.co/datasets/nvidia/AceMath-Instruct-Training-Data), [AceMath-RM Training Data](https://huggingface.co/datasets/nvidia/AceMath-RM-Training-Data)
43
+
44
+ ### General Instruction Models
45
+ - [AceInstruct-1.5B](https://huggingface.co/nvidia/AceInstruct-1.5B), [AceInstruct-7B](https://huggingface.co/nvidia/AceInstruct-7B), [AceInstruct-72B](https://huggingface.co/nvidia/AceInstruct-72B)
46
+
47
+
48
+ ## How to use
49
+ ```python
50
+ from transformers import AutoModelForCausalLM, AutoTokenizer
51
+
52
+ model_name = "AceInstruct-1.5B"
53
+ tokenizer = AutoTokenizer.from_pretrained(model_name)
54
+ model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")
55
+
56
+ prompt = "Tell me something about artificial intelligence."
57
+ messages = [{"role": "user", "content": prompt}]
58
+
59
+ text = tokenizer.apply_chat_template(
60
+ messages,
61
+ tokenize=False,
62
+ add_generation_prompt=True
63
+ )
64
+ model_inputs = tokenizer([text], return_tensors="pt").to("cuda")
65
+
66
+ generated_ids = model.generate(
67
+ **model_inputs,
68
+ max_new_tokens=1024
69
+ )
70
+ generated_ids = [
71
+ output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
72
+ ]
73
+
74
+ response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
75
+ ```
76
+
77
+
78
+ ## Correspondence to
79
+ Zihan Liu ([email protected]), Yang Chen ([email protected]), Wei Ping ([email protected])
80
+
81
+
82
+ ## Citation
83
+ If you find our work helpful, we’d appreciate it if you could cite us.
84
+ <pre>
85
+ @article{acemath2024,
86
+ title={AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling},
87
+ author={Liu, Zihan and Chen, Yang and Shoeybi, Mohammad and Catanzaro, Bryan and Ping, Wei},
88
+ journal={arXiv preprint},
89
+ year={2024}
90
+ }
91
+ </pre>
92
+
93
+
94
+ ## License
95
+ All models in the AceInstruct family are for non-commercial use only, subject to [Terms of Use](https://openai.com/policies/row-terms-of-use/) of the data generated by OpenAI. We put the AceInstruct models under the license of [Creative Commons Attribution: Non-Commercial 4.0 International](https://spdx.org/licenses/CC-BY-NC-4.0).