DeepSeek V3 by unsloth

 ยป  All LLMs  ยป  unsloth  ยป  DeepSeek V3   URL Share it on

  Arxiv:2412.19437   Autotrain compatible Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-...   Conversational   Custom code   Deepseek   Deepseek v3   En   Fp8   Region:us   Safetensors   Sharded   Tensorflow   Unsloth
Model Card on HF ๐Ÿค—: https://huggingface.co/unsloth/DeepSeek-V3 

DeepSeek V3 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
DeepSeek V3 (unsloth/DeepSeek-V3)

DeepSeek V3 Parameters and Internals

LLM NameDeepSeek V3
Repository ๐Ÿค—https://huggingface.co/unsloth/DeepSeek-V3 
Base Model(s)  DeepSeek V3   deepseek-ai/DeepSeek-V3
Model Size684.5b
Required VRAM167.5 GB
Updated2025-01-21
Maintainerunsloth
Model Typedeepseek_v3
Model Files  5.2 GB: 1-of-163   4.3 GB: 2-of-163   4.3 GB: 3-of-163   4.3 GB: 4-of-163   4.3 GB: 5-of-163   4.4 GB: 6-of-163   4.3 GB: 7-of-163   4.3 GB: 8-of-163   4.3 GB: 9-of-163   4.3 GB: 10-of-163   4.3 GB: 11-of-163   1.3 GB: 12-of-163   4.3 GB: 13-of-163   4.3 GB: 14-of-163   4.3 GB: 15-of-163   4.3 GB: 16-of-163   4.3 GB: 17-of-163   4.3 GB: 18-of-163   4.3 GB: 19-of-163   4.3 GB: 20-of-163   4.3 GB: 21-of-163   4.3 GB: 22-of-163   4.3 GB: 23-of-163   4.3 GB: 24-of-163   4.3 GB: 25-of-163   4.3 GB: 26-of-163   4.3 GB: 27-of-163   4.3 GB: 28-of-163   4.3 GB: 29-of-163   4.3 GB: 30-of-163   4.3 GB: 31-of-163   4.3 GB: 32-of-163   4.3 GB: 33-of-163   1.8 GB: 34-of-163   4.3 GB: 35-of-163   4.3 GB: 36-of-163   4.3 GB: 37-of-163   4.3 GB: 38-of-163   4.3 GB: 39-of-163   4.3 GB: 40-of-163
Supported Languagesen
Model ArchitectureDeepseekV3ForCausalLM
Licensemit
Context Length163840
Model Max Length163840
Transformers Version4.33.1
Vocabulary Size129280
Torch Data Typebfloat16

Best Alternatives to DeepSeek V3

Best Alternatives
Context / RAM
Downloads
Likes
DeepSeek V3160K / 171.8 GB1619172092
DeepSeek V3 Base160K / 171.8 GB173991282
DeepSeek R1160K / 184.7 GB0750
DeepSeek R1 Zero160K / 184.7 GB0231
DeepSeek R1160K / 193.3 GB9014
DeepSeek V3 Bf16160K / 358.8 GB8379
DeepSeek R1 Bf16160K / 367.4 GB111
Note: green Score (e.g. "73.2") means that the model is better than unsloth/DeepSeek-V3.

Rank the DeepSeek V3 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? 41727 in total.

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