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

Starcoder2 15B by bigcode

 »  All LLMs  »  bigcode  »  Starcoder2 15B   URL Share it on

  12652   
  Arxiv:2004.05150   Arxiv:2205.14135   Arxiv:2207.14255   Arxiv:2305.13245   Arxiv:2402.19173   Autotrain compatible   Code Dataset:bigcode/the-stack-v2-t...   Endpoints compatible   Model-index   Region:us   Safetensors   Sharded   Starcoder2   Tensorflow
Model Card on HF 🤗: https://huggingface.co/bigcode/starcoder2-15b 

Starcoder2 15B Benchmarks

MMLU Pro: 15.03
GPQA: 3.13
MUSR: 2.93
BBH: 20.37
IFEval: 27.8 vs 88 (so35)-68.4%
ARC: 47.35 vs 96.7 (so35)-51%
HellaSwag: 64.09 vs 95.3 (gpt4)-32.7%
MMLU: 51.35 vs 88.3 (so35)-41.8%
TruthfulQA: 37.87 vs 59 (gpt4)-35.8%
WinoGrande: 63.85 vs 87.5 (gpt4)-27%
GSM8K: 52.24 vs 96.4 (so35)-45.8%
HumanEvalA: 34.85
Python Eva: 44.15 vs 92 (so35)-52%
Java Eval: 33.86
Javascript: 44.24
CPP Eval: 41.44
PHP Eval: 39.48
Julia Eval: 33.19
D Eval: 23.64
Lua Eval: 43.75
R Eval: 19.81
Racket Eva: 22.41
Rust Eval: 38.03
Swift Eval: 34.18
MATH Lvl 5: 5.97
LLME Score: 0.31068
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Starcoder2 15B (bigcode/starcoder2-15b)

Starcoder2 15B Parameters and Internals

Model Type 
text-generation
Use Cases 
Areas:
Research, Educational
Applications:
Code generation, Assisting developers
Primary Use Cases:
Text generation based on programming language inputs
Limitations:
May not work with instruction-based commands, Generated code might contain bugs, Not guaranteed to work as intended
Additional Notes 
Requires proper attribution when utilizing generated code from the model due to licenses.
Supported Languages 
600+ programming languages ()
Training Details 
Data Sources:
The Stack v2, Arxiv, Wikipedia
Data Volume:
4+ trillion tokens
Methodology:
Grouped Query Attention, Sliding Window Attention, Fill-in-the-Middle objective
Context Length:
16384
Hardware Used:
NVIDIA DGX H100, 1024 x H100 GPUs
Model Architecture:
Transformer decoder
Input Output 
Input Format:
Text prompt
Output Format:
Generated text/code
LLM NameStarcoder2 15B
Repository 🤗https://huggingface.co/bigcode/starcoder2-15b 
Model Size15b
Required VRAM63.8 GB
Updated2025-04-19
Maintainerbigcode
Model Typestarcoder2
Model Files  4.6 GB: 1-of-14   4.6 GB: 2-of-14   4.6 GB: 3-of-14   4.6 GB: 4-of-14   4.6 GB: 5-of-14   4.6 GB: 6-of-14   4.6 GB: 7-of-14   4.6 GB: 8-of-14   4.6 GB: 9-of-14   4.6 GB: 10-of-14   4.6 GB: 11-of-14   4.6 GB: 12-of-14   4.6 GB: 13-of-14   4.0 GB: 14-of-14
Model ArchitectureStarcoder2ForCausalLM
Licensebigcode-openrail-m
Context Length16384
Model Max Length16384
Transformers Version4.37.0.dev0
Tokenizer ClassGPT2Tokenizer
Vocabulary Size49152
Torch Data Typefloat32

Quantized Models of the Starcoder2 15B

Model
Likes
Downloads
VRAM
Starcoder2 15B AWQ129 GB
StarCoder2 15B GGUF257746 GB

Best Alternatives to Starcoder2 15B

Best Alternatives
Context / RAM
Downloads
Likes
Starchat2 15B V0.116K / 31.9 GB6218110
Starcoder2 15B Instruct V0.116K / 31.9 GB762101
CodeFuse StarCoder2 15B16K / 31.9 GB122
Starcoder2 15B Finetuned Drake16K / 63.8 GB50
...aceH4 Starchat2 15B V0.1 4bits16K / 9.9 GB50
Starcoder2 15B Instruct V0.116K / 53.4 GB50
Dolphincoder Starcoder2 15B16K / 31.9 GB12769
Starchat2 15B Sft V0.116K / 31.9 GB25
Starcoder2 15B Instruct16K / 31.9 GB47
OpenCodeInterpreter SC2 15B16K / 31.9 GB34

Rank the Starcoder2 15B 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? 46490 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