PowerMoE 3B by ibm

 ยป  All LLMs  ยป  ibm  ยป  PowerMoE 3B   URL Share it on

  Arxiv:2408.13359   Autotrain compatible   Granitemoe   Model-index   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF ๐Ÿค—: https://huggingface.co/ibm/PowerMoE-3b 

PowerMoE 3B Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
PowerMoE 3B (ibm/PowerMoE-3b)

PowerMoE 3B Parameters and Internals

Model Type 
text generation
Use Cases 
Applications:
natural language multi-choices, code generation, math reasoning
Additional Notes 
The model requires a custom branch of the Transformers library for usage: https://github.com/mayank31398/transformers/tree/granitemoe
Training Details 
Data Sources:
mix of open-source and proprietary datasets
Methodology:
Mixture-of-Experts (sMoE) approach with the Power learning rate scheduler
Model Architecture:
sparse Mixture-of-Experts model
LLM NamePowerMoE 3B
Repository ๐Ÿค—https://huggingface.co/ibm/PowerMoE-3b 
Model Size3b
Required VRAM13.5 GB
Updated2024-12-21
Maintaineribm
Model Typegranitemoe
Model Files  4.8 GB: 1-of-3   4.8 GB: 2-of-3   3.9 GB: 3-of-3
Model ArchitectureGraniteMoeForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.44.0.dev0
Tokenizer ClassGPT2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size49152
Torch Data Typefloat32
Activation Functionsilu

Best Alternatives to PowerMoE 3B

Best Alternatives
Context / RAM
Downloads
Likes
Granite 3.1 3B A800m Instruct128K / 6.6 GB2926
Granite 3.1 3B A800m Base128K / 6.6 GB1701
Granite 3.0 3B A800m Instruct4K / 6.8 GB403016
Granite 3.0 3B A800m Base4K / 13.5 GB30584
Note: green Score (e.g. "73.2") means that the model is better than ibm/PowerMoE-3b.

Rank the PowerMoE 3B 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? 40013 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241217