Text Generation
Transformers
Safetensors
llama
conversational
text-generation-inference
Inference Endpoints
yuxiang630 commited on
Commit
a699712
1 Parent(s): 7e37b87

feat: README

Browse files
Files changed (3) hide show
  1. README.md +103 -1
  2. assets/overview.svg +0 -0
  3. assets/result.png +0 -0
README.md CHANGED
@@ -6,4 +6,106 @@ datasets:
6
  - ise-uiuc/Magicoder-Evol-Instruct-110K
7
  library_name: transformers
8
  pipeline_tag: text-generation
9
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6
  - ise-uiuc/Magicoder-Evol-Instruct-110K
7
  library_name: transformers
8
  pipeline_tag: text-generation
9
+ ---
10
+ # 🎩 Magicoder: Source Code Is All You Need
11
+
12
+ * 🎩**Magicoder** is a model family empowered by 🪄**OSS-Instruct**, a novel approach to enlightening LLMs with open-source code snippets for generating *low-bias* and *high-quality* instruction data for code.
13
+ * 🪄**OSS-Instruct** mitigates the *inherent bias* of the LLM-synthesized instruction data by empowering them with *a wealth of open-source references* to produce more diverse, realistic, and controllable data.
14
+
15
+ ![Overview of OSS-Instruct](assets/overview.svg)
16
+ ![Overview of Result](assets/result.png)
17
+
18
+ ## Model Details
19
+
20
+ ### Model Description
21
+
22
+ * **Developed by:**
23
+ [Yuxiang Wei](https://yuxiang.cs.illinois.edu),
24
+ [Zhe Wang](https://github.com/zhewang2001),
25
+ [Jiawei Liu](https://jiawei-site.github.io),
26
+ [Yifeng Ding](https://yifeng-ding.com),
27
+ [Lingming Zhang](https://lingming.cs.illinois.edu)
28
+ * **License:** [DeepSeek](https://github.com/deepseek-ai/DeepSeek-Coder/blob/main/LICENSE-MODEL)
29
+ * **Finetuned from model:** [deepseek-coder-6.7b-base](https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-base)
30
+
31
+ ### Model Sources
32
+
33
+ * **Repository:** <https://github.com/ise-uiuc/magicoder>
34
+ * **Paper:** <https://arxiv.org/pdf/2312.02120.pdf>
35
+ * **Demo (powered by [Gradio](https://www.gradio.app)):**
36
+ <https://github.com/ise-uiuc/magicoder/tree/main/demo>
37
+
38
+ ## Uses
39
+
40
+ ### Direct Use
41
+
42
+ Magicoders are designed and best suited for **coding tasks**.
43
+
44
+ ### Out-of-Scope Use
45
+
46
+ Magicoders may not work well in non-coding tasks.
47
+
48
+ ## Bias, Risks, and Limitations
49
+
50
+ Magicoders may sometimes make errors, producing misleading contents, or struggle to manage tasks that are not related to coding.
51
+
52
+ ### Recommendations
53
+
54
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
55
+
56
+ ## How to Get Started with the Model
57
+
58
+ Use the code below to get started with the model. Make sure you installed the [transformers](https://huggingface.co/docs/transformers/index) library.
59
+
60
+ ```python
61
+ from transformers import pipeline
62
+ import torch
63
+
64
+ MAGICODER_PROMPT = """You are an exceptionally intelligent coding assistant that consistently delivers accurate and reliable responses to user instructions.
65
+
66
+ @@ Instruction
67
+ {instruction}
68
+
69
+ @@ Response
70
+ """
71
+
72
+ instruction = <Your code instruction here>
73
+
74
+ prompt = MAGICODER_PROMPT.format(instruction=instruction)
75
+ generator = pipeline(
76
+ model="ise-uiuc/Magicoder-S-DS-6.7B",
77
+ task="text-generation",
78
+ torch_dtype=torch.bfloat16,
79
+ device_map="auto",
80
+ )
81
+ result = generator(prompt, max_length=1024, num_return_sequences=1, temperature=0.0)
82
+ print(result[0]["generated_text"])
83
+ ```
84
+
85
+ ## Technical Details
86
+
87
+ Refer to our GitHub repo: [ise-uiuc/magicoder](https://github.com/ise-uiuc/magicoder/).
88
+
89
+ ## 📝 Citation
90
+
91
+ ```bibtex
92
+ @misc{magicoder,
93
+ title={Magicoder: Source Code Is All You Need},
94
+ author={Yuxiang Wei and Zhe Wang and Jiawei Liu and Yifeng Ding and Lingming Zhang},
95
+ year={2023},
96
+ eprint={2312.02120},
97
+ archivePrefix={arXiv},
98
+ primaryClass={cs.CL}
99
+ }
100
+ ```
101
+
102
+ ## 🙏 Acknowledgements
103
+
104
+ * [WizardCoder](https://github.com/nlpxucan/WizardLM/tree/main/WizardCoder): Evol-Instruct
105
+ * [DeepSeek-Coder](https://github.com/deepseek-ai/DeepSeek-Coder): Base model for Magicoder-DS
106
+ * [CodeLlama](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/): Base model for Magicoder-CL
107
+ * [StarCoder](https://arxiv.org/abs/2305.06161): Data decontamination
108
+
109
+ ## Important Note
110
+
111
+ Magicoder models are trained on the synthetic data generated by OpenAI models. Please pay attention to OpenAI's [terms of use](https://openai.com/policies/terms-of-use) when using the models and the datasets. Magicoders will not compete with OpenAI's commercial products.
assets/overview.svg ADDED
assets/result.png ADDED