Viking 13B by LumiOpen

 ยป  All LLMs  ยป  LumiOpen  ยป  Viking 13B   URL Share it on

  Autotrain compatible   Da   Dataset:bigcode/starcoderdata Dataset:cerebras/slimpajama-62...   Dataset:mc4   En   Endpoints compatible   Fi   Is   Llama   Nn   No   Region:us   Safetensors   Sharded   Sv   Tensorflow
Model Card on HF ๐Ÿค—: https://huggingface.co/LumiOpen/Viking-13B 

Viking 13B Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Viking 13B (LumiOpen/Viking-13B)

Viking 13B Parameters and Internals

Model Type 
decoder-only transformer, text generation
Use Cases 
Areas:
research, commercial applications
Applications:
text generation, natural language understanding
Primary Use Cases:
translation, code generation
Limitations:
Limited proficiency outside supported languages
Additional Notes 
Base model, requiring further fine-tuning for specific use cases.
Supported Languages 
fi (fluent), en (fluent), da (fluent), sv (fluent), no (fluent), nn (fluent), is (fluent)
Training Details 
Data Sources:
cerebras/SlimPajama-627B, bigcode/starcoderdata, mc4
Data Volume:
2 trillion tokens
Context Length:
4096
Training Time:
since September 2023
Hardware Used:
512 AMD MI250X GPUs
Model Architecture:
GPT-like with rotary positional embeddings and flash attention
Responsible Ai Considerations 
Fairness:
May produce outputs that are inaccurate, prejudiced, or controversial due to its training data.
Mitigation Strategies:
Users should consider additional evaluation and customization.
Input Output 
Input Format:
text input with tokens
Accepted Modalities:
text
Output Format:
generated text
Release Notes 
Version:
1.0
Date:
2023-10-11
Notes:
Initial model release with partial training data.
LLM NameViking 13B
Repository ๐Ÿค—https://huggingface.co/LumiOpen/Viking-13B 
Model Size13b
Required VRAM28.1 GB
Updated2025-01-24
MaintainerLumiOpen
Model Typellama
Model Files  4.9 GB: 1-of-6   4.9 GB: 2-of-6   5.0 GB: 3-of-6   5.0 GB: 4-of-6   4.9 GB: 5-of-6   3.4 GB: 6-of-6
Supported Languagesfi en da sv no nn is
Model ArchitectureLlamaForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.40.0
Tokenizer ClassBloomTokenizer
Padding Token<pad>
Vocabulary Size131072
Torch Data Typebfloat16

Best Alternatives to Viking 13B

Best Alternatives
Context / RAM
Downloads
Likes
Yarn Llama 2 13B 128K128K / 26 GB4637113
Luminaura RP 13B128K / 26 GB190
Agent Llama2 13B 80K80K / 26.4 GB120
Chat Llama2 13B 80K80K / 52.8 GB90
LongAlign 13B 64K64K / 26 GB3913
LongAlign 13B 64K Base64K / 26 GB213
Yarn Llama 2 13B 64K64K / 26 GB268917
Openbuddy Llama2 13B V15p1 64K64K / 26.1 GB134
Openbuddy Llama2 13b64k V1564K / 26.1 GB171
Airoboros L2 13B 2.1 YaRN 64K64K / 26 GB167

Rank the Viking 13B 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? 41817 in total.

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