Skywork MoE Base by Skywork

 ยป  All LLMs  ยป  Skywork  ยป  Skywork MoE Base   URL Share it on

  Arxiv:2406.00605   Arxiv:2406.06563   Autotrain compatible   Custom code   Moe   Pytorch   Region:us   Sharded   Skywork moe

Skywork MoE Base Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Skywork MoE Base (Skywork/Skywork-MoE-Base)

Skywork MoE Base Parameters and Internals

Model Type 
mixture-of-experts
Training Details 
Methodology:
Gating Logit Normalization, Adaptive Auxiliary Loss Coefficients
Model Architecture:
mixture-of-experts, 16 experts
Release Notes 
Version:
Skywork-MoE-Base
Date:
2024.6.3
Notes:
Release of the Skywork-MoE-Base model.
LLM NameSkywork MoE Base
Repository ๐Ÿค—https://huggingface.co/Skywork/Skywork-MoE-Base 
Required VRAM240.8 GB
Updated2025-01-23
MaintainerSkywork
Model Typeskywork_moe
Model Files  5.6 GB: 1-of-53   5.6 GB: 2-of-53   5.6 GB: 3-of-53   5.6 GB: 4-of-53   5.6 GB: 5-of-53   5.6 GB: 6-of-53   5.6 GB: 7-of-53   5.6 GB: 8-of-53   5.6 GB: 9-of-53   5.6 GB: 10-of-53   5.6 GB: 11-of-53   5.6 GB: 12-of-53   5.6 GB: 13-of-53   5.6 GB: 14-of-53   5.6 GB: 15-of-53   5.6 GB: 16-of-53   5.6 GB: 17-of-53   5.6 GB: 18-of-53   5.6 GB: 19-of-53   5.6 GB: 20-of-53   5.6 GB: 21-of-53   5.6 GB: 22-of-53   5.6 GB: 23-of-53   5.6 GB: 24-of-53   5.6 GB: 25-of-53   5.6 GB: 26-of-53   5.6 GB: 27-of-53   5.6 GB: 28-of-53   5.6 GB: 29-of-53   5.6 GB: 30-of-53   5.6 GB: 31-of-53   5.6 GB: 32-of-53   5.6 GB: 33-of-53   5.6 GB: 34-of-53   5.6 GB: 35-of-53   5.6 GB: 36-of-53   5.6 GB: 37-of-53   5.6 GB: 38-of-53   5.6 GB: 39-of-53   5.6 GB: 40-of-53   5.6 GB: 41-of-53   5.6 GB: 42-of-53   5.6 GB: 43-of-53
Model ArchitectureSkyworkForCausalLM
Licenseother
Context Length8192
Model Max Length8192
Transformers Version4.40.1
Vocabulary Size65532
Torch Data Typebfloat16

Rank the Skywork MoE Base Capabilities

๐Ÿ†˜ 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! ๐ŸŒŸ

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? 41774 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227