Llama3 German 8B 32K by DiscoResearch

 ยป  All LLMs  ยป  DiscoResearch  ยป  Llama3 German 8B 32K   URL Share it on

  Arxiv:2404.10830   Autotrain compatible   De   Endpoints compatible   Llama   Region:us   Safetensors   Sharded   Tensorflow

Llama3 German 8B 32K Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
๐ŸŒŸ Advertise your project ๐Ÿš€

Llama3 German 8B 32K Parameters and Internals

Model Type 
text generation, language model
Use Cases 
Areas:
research, commercial applications
Primary Use Cases:
German language tasks
Additional Notes 
Has a long-context version capable of handling up to 32k context length and an instruction-tuned variant.
Supported Languages 
German (high), English (moderate)
Training Details 
Data Sources:
occiglot-fineweb-0.5, LLM-Datasets, Common-Crawl
Data Volume:
65 billion German tokens
Methodology:
Continuous pretraining with intelligent document packing
Context Length:
8192
Training Time:
~60 hours
Hardware Used:
128 GPUs at Hessian.AI 42 Supercomputer
Model Architecture:
Improved on Llama3 architecture
Input Output 
Input Format:
Tokenized text
Accepted Modalities:
text
Output Format:
Generated text
Performance Tips:
Consider fine-tuning for specific use cases.
LLM NameLlama3 German 8B 32K
Repository ๐Ÿค—https://huggingface.co/DiscoResearch/Llama3-German-8B-32k 
Model Size8b
Required VRAM16.1 GB
Updated2025-06-09
MaintainerDiscoResearch
Model Typellama
Model Files  5.0 GB: 1-of-4   5.0 GB: 2-of-4   4.9 GB: 3-of-4   1.2 GB: 4-of-4
Supported Languagesde
Model ArchitectureLlamaForCausalLM
Licensellama3
Context Length65536
Model Max Length65536
Transformers Version4.40.2
Tokenizer ClassPreTrainedTokenizerFast
Vocabulary Size128256
Torch Data Typebfloat16
Llama3 German 8B 32K (DiscoResearch/Llama3-German-8B-32k)

Quantized Models of the Llama3 German 8B 32K

Model
Likes
Downloads
VRAM
Llama3 German 8B 32K GPTQ0155 GB

Best Alternatives to Llama3 German 8B 32K

Best Alternatives
Context / RAM
Downloads
Likes
...otron 8B UltraLong 4M Instruct4192K / 32.1 GB3284108
UltraLong Thinking4192K / 16.1 GB3672
...a 3.1 8B UltraLong 4M Instruct4192K / 32.1 GB17624
...a 3.1 8B UltraLong 2M Instruct2096K / 32.1 GB8759
...otron 8B UltraLong 2M Instruct2096K / 32.1 GB52615
Zero Llama 3.1 8B Beta61048K / 16.1 GB9581
...otron 8B UltraLong 1M Instruct1048K / 32.1 GB180845
...a 3.1 8B UltraLong 1M Instruct1048K / 32.1 GB138729
...xis Bookwriter Llama3.1 8B Sft1048K / 16.1 GB634
....1 1million Ctx Dark Planet 8B1048K / 32.3 GB902
Note: green Score (e.g. "73.2") means that the model is better than DiscoResearch/Llama3-German-8B-32k.

Rank the Llama3 German 8B 32K 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? 48046 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124