Hugging Face's logo

RichardErkhov
/
SenseLLM_-_ReflectionCoder-CL-34B-gguf

GGUF
Inference Endpoints
conversational
Model card Files Community
RichardErkhov commited on
Commit
94ade6b
•
1 Parent(s): fe8ea4b

uploaded readme

Browse files
Files changed (1) hide show
  1. README.md +134 -0
README.md ADDED
@@ -0,0 +1,134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Quantization made by Richard Erkhov.
2
+
3
+ [Github](https://github.com/RichardErkhov)
4
+
5
+ [Discord](https://discord.gg/pvy7H8DZMG)
6
+
7
+ [Request more models](https://github.com/RichardErkhov/quant_request)
8
+
9
+
10
+ ReflectionCoder-CL-34B - GGUF
11
+ - Model creator: https://huggingface.co/SenseLLM/
12
+ - Original model: https://huggingface.co/SenseLLM/ReflectionCoder-CL-34B/
13
+
14
+
15
+ | Name | Quant method | Size |
16
+ | ---- | ---- | ---- |
17
+ | [ReflectionCoder-CL-34B.Q2_K.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q2_K.gguf) | Q2_K | 11.65GB |
18
+ | [ReflectionCoder-CL-34B.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q3_K_S.gguf) | Q3_K_S | 13.6GB |
19
+ | [ReflectionCoder-CL-34B.Q3_K.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q3_K.gguf) | Q3_K | 15.19GB |
20
+ | [ReflectionCoder-CL-34B.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q3_K_M.gguf) | Q3_K_M | 15.19GB |
21
+ | [ReflectionCoder-CL-34B.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q3_K_L.gguf) | Q3_K_L | 16.55GB |
22
+ | [ReflectionCoder-CL-34B.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.IQ4_XS.gguf) | IQ4_XS | 16.99GB |
23
+ | [ReflectionCoder-CL-34B.Q4_0.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q4_0.gguf) | Q4_0 | 17.74GB |
24
+ | [ReflectionCoder-CL-34B.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.IQ4_NL.gguf) | IQ4_NL | 17.92GB |
25
+ | [ReflectionCoder-CL-34B.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q4_K_S.gguf) | Q4_K_S | 17.87GB |
26
+ | [ReflectionCoder-CL-34B.Q4_K.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q4_K.gguf) | Q4_K | 18.83GB |
27
+ | [ReflectionCoder-CL-34B.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q4_K_M.gguf) | Q4_K_M | 18.83GB |
28
+ | [ReflectionCoder-CL-34B.Q4_1.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q4_1.gguf) | Q4_1 | 19.69GB |
29
+ | [ReflectionCoder-CL-34B.Q5_0.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q5_0.gguf) | Q5_0 | 21.64GB |
30
+ | [ReflectionCoder-CL-34B.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q5_K_S.gguf) | Q5_K_S | 21.64GB |
31
+ | [ReflectionCoder-CL-34B.Q5_K.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q5_K.gguf) | Q5_K | 22.2GB |
32
+ | [ReflectionCoder-CL-34B.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q5_K_M.gguf) | Q5_K_M | 22.2GB |
33
+ | [ReflectionCoder-CL-34B.Q5_1.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q5_1.gguf) | Q5_1 | 23.59GB |
34
+ | [ReflectionCoder-CL-34B.Q6_K.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q6_K.gguf) | Q6_K | 25.78GB |
35
+ | [ReflectionCoder-CL-34B.Q8_0.gguf](https://huggingface.co/RichardErkhov/SenseLLM_-_ReflectionCoder-CL-34B-gguf/blob/main/ReflectionCoder-CL-34B.Q8_0.gguf) | Q8_0 | 33.39GB |
36
+
37
+
38
+
39
+
40
+ Original model description:
41
+ ---
42
+ license: apache-2.0
43
+ datasets:
44
+ - SenseLLM/ReflectionSeq-GPT
45
+ - SenseLLM/ReflectionSeq-DS
46
+ language:
47
+ - en
48
+ ---
49
+ ## ReflectionCoder: Learning from Reflection Sequence for Enhanced One-off Code Generation
50
+
51
+ <p align="center">
52
+ <a href="https://arxiv.org/abs/2405.17057">📄 Paper</a> •
53
+ <a href="https://github.com/SenseLLM/ReflectionCoder">🏠 Repo</a> •
54
+ <a href="https://huggingface.co/SenseLLM/ReflectionCoder-DS-33B">🤖 Models</a> •
55
+ <a href="https://huggingface.co/datasets/SenseLLM/ReflectionSeq-GPT">📚 Datasets </a>
56
+ </p>
57
+
58
+ ## Introduction
59
+ ReflectionCoder is a novel approach that effectively leverages reflection sequences constructed by integrating compiler feedback to improve one-off code generation performance. Please refer to our paper and repo for more details!
60
+
61
+ ![](method.png)
62
+
63
+ <hr>
64
+
65
+ ## Models
66
+
67
+ | Model | Checkpoint | Size | HumanEval (+) | MBPP (+) | License|
68
+ |:-------|:------------|:------|:---------------|:----------|:--------|
69
+ | ReflectionCoder-CL-7B | 🤗 [HF Link](https://huggingface.co/SenseLLM/ReflectionCoder-CL-7B) | 7B | 75.0 (68.9) | 72.2 (61.4) | [Llama2](https://ai.meta.com/llama/license/) |
70
+ | ReflectionCoder-CL-34B | 🤗 [HF Link](https://huggingface.co/SenseLLM/ReflectionCoder-CL-34B) | 34B | 70.7 (66.5) | 68.4 (56.6) | [Llama2](https://ai.meta.com/llama/license/) |
71
+ | ReflectionCoder-DS-6.7B | 🤗 [HF Link](https://huggingface.co/SenseLLM/ReflectionCoder-DS-6.7B) | 6.7B | 80.5 (74.4) | 81.5 (69.6) | [DeepSeek](https://github.com/deepseek-ai/DeepSeek-Coder/blob/main/LICENSE-MODEL) |
72
+ | ReflectionCoder-DS-33B | 🤗 [HF Link](https://huggingface.co/SenseLLM/ReflectionCoder-DS-33B) | 33B | 82.9 (76.8) | 84.1 (72.0) | [DeepSeek](https://github.com/deepseek-ai/DeepSeek-Coder/blob/main/LICENSE-MODEL) |
73
+
74
+ ## Datasets
75
+
76
+ | Dataset | Link | License |
77
+ |:-------------------|:----------------|:----------------------------------------------|
78
+ | ReflectionSeq-GPT | 🤗 [HF Link](https://huggingface.co/datasets/SenseLLM/ReflectionSeq-GPT) | [License](LICENSE) |
79
+ | ReflectionSeq-DS | 🤗 [HF Link](https://huggingface.co/datasets/SenseLLM/ReflectionSeq-DS) | [License](LICENSE) |
80
+
81
+
82
+ ## How to Use
83
+
84
+ #### Chat Format
85
+ Following chat templates of most models, we use two special tokens to wrap the message of user and assistant, *i.e.*, ``<|user|>``, ``<|assistant|>``, and ``<|endofmessage|>``. Furthermore, we use two special tokens to wrap the content of different blocks, *i.e.*, ``<|text|>`` and ``<|endofblock|>``. You can use the following template to prompt our ReflectionCoder.
86
+
87
+ ```python
88
+ import torch
89
+ from transformers import pipeline
90
+
91
+ chat = [
92
+ {"role": "user", "content": "<Your code instruction here>"}
93
+ ]
94
+
95
+ generator = pipeline(
96
+ model="SenseLLM/ReflectionCoder-CL-34B",
97
+ task="text-generation",
98
+ torch_dtype=torch.bfloat16,
99
+ device_map="auto",
100
+ )
101
+
102
+ result = generator(chat, max_length=128, num_return_sequences=1)
103
+
104
+ print(result)
105
+ ```
106
+
107
+ Please refer to our [GitHub Repo](https://github.com/SenseLLM/ReflectionCoder) for more technical details.
108
+
109
+ ## Citation
110
+
111
+ If you find this repo useful for your research, please kindly cite our paper:
112
+ ```
113
+ @misc{ren2024reflectioncoder,
114
+ title={ReflectionCoder: Learning from Reflection Sequence for Enhanced One-off Code Generation},
115
+ author={Houxing Ren and Mingjie Zhan and Zhongyuan Wu and Aojun Zhou and Junting Pan and Hongsheng Li},
116
+ year={2024},
117
+ eprint={2405.17057},
118
+ archivePrefix={arXiv},
119
+ primaryClass={cs.CL}
120
+ }
121
+ ```
122
+
123
+ ## Acknowledgments
124
+
125
+ We thank the following amazing projects that truly inspired us:
126
+
127
+ - [CodeLlama](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/)
128
+ - [DeepSeek-Coder](https://github.com/deepseek-ai/DeepSeek-Coder)
129
+ - [WizardCoder](https://github.com/nlpxucan/WizardLM/tree/main/WizardCoder)
130
+ - [Evol-CodeAlpaca-v1](https://huggingface.co/datasets/theblackcat102/evol-codealpaca-v1)
131
+ - [MagiCoder](https://github.com/ise-uiuc/magicoder/tree/main)
132
+ - [EvalPlus](https://github.com/evalplus/evalplus)
133
+ - [OpenCoderInterpreter](https://github.com/OpenCodeInterpreter/OpenCodeInterpreter/tree/main)
134
+