LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

DeciCoder 1B by Deci

 »  All LLMs  »  Deci  »  DeciCoder 1B   URL Share it on

  2354   
  Arxiv:2104.09864   Arxiv:2305.13245   Autotrain compatible   Custom code   Dataset:bigcode/starcoderdata   Deci ai   Decicoder   Model-index   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/Deci/DeciCoder-1b 

DeciCoder 1B Benchmarks

ARC: 21.16 vs 96.7 (so35)-78.1%
HellaSwag: 31.09 vs 95.3 (gpt4)-67.4%
MMLU: 24.34 vs 88.3 (so35)-72.4%
TruthfulQA: 47.05 vs 59 (gpt4)-20.3%
WinoGrande: 50.83 vs 87.5 (gpt4)-41.9%
GSM8K: 1.74 vs 96.4 (so35)-98.2%
HumanEvalA: 5.86
Python Eva: 19.32 vs 92 (so35)-79%
Java Eval: 15.3
Javascript: 17.85
CPP Eval: 6.87
PHP Eval: 2.01
D Eval: 6.08
R Eval: 0.1
Racket Eva: 0.47
Rust Eval: 1.72
Swift Eval: 0.63
LLME Score: 0.21203
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
DeciCoder 1B (Deci/DeciCoder-1b)

DeciCoder 1B Parameters and Internals

Model Type 
text-generation
Use Cases 
Areas:
text generation, code completion
Applications:
programming, software development
Primary Use Cases:
single/multiline code completion
Limitations:
primarily trained on English source code, may produce suboptimal code, may contain bugs or exploits
Considerations:
present a function signature and docstring to get accurate code completion
Additional Notes 
Model trained only on Python, Java, and Javascript subsets.
Supported Languages 
Python (high), Java (high), JavaScript (high)
Training Details 
Data Sources:
Starcoder Training Dataset
Data Volume:
446 billion tokens
Methodology:
Grouped Query Attention, Fill-in-the-Middle training objective
Context Length:
2048
Model Architecture:
Neural Architecture Search-based technology, AutoNAC
Input Output 
Input Format:
code snippets with comments or function signatures
Accepted Modalities:
text
Output Format:
completed code
Performance Tips:
Use up to 2048 tokens in the context window for optimal performance.
LLM NameDeciCoder 1B
Repository 🤗https://huggingface.co/Deci/DeciCoder-1b 
Model Size1b
Required VRAM2.2 GB
Updated2025-02-05
MaintainerDeci
Model Files  2.2 GB
Model ArchitectureDeciCoderForCausalLM
Licenseapache-2.0
Context Length2048
Model Max Length2048
Transformers Version4.31.0.dev0
Tokenizer ClassGPT2Tokenizer
Vocabulary Size49152
Torch Data Typebfloat16

Best Alternatives to DeciCoder 1B

Best Alternatives
Context / RAM
Downloads
Likes
...oder1B Personal Copilot Merged2K / 2.2 GB1101
Decicoder 1B Openvino Int82K / 1.1 GB371

Rank the DeciCoder 1B Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation  
Factuality and Completeness of Knowledge  
Censorship and Alignment  
Data Analysis and Insight Generation  
Text Generation  
Text Summarization and Feature Extraction  
Code Generation  
Multi-Language Support and Translation  

What open-source LLMs or SLMs are you in search of? 42625 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227
  Support LLM Explorer