LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

Sailor 7B by sail

 »  All LLMs  »  sail  »  Sailor 7B   URL Share it on

  269   
  Arxiv:2404.03608   Autotrain compatible Base model:finetune:qwen/qwen1...   Base model:qwen/qwen1.5-7b   Conversational   Dataset:allenai/madlad-400   Dataset:cc100 Dataset:cerebras/slimpajama-62...   Dataset:skywork/skypile-150b   En   Endpoints compatible   Id   Lo   Model-index   Ms   Multilingual   Qwen2   Region:us   Safetensors   Sailor   Sea   Th   Vi   Zh
Model Card on HF 🤗: https://huggingface.co/sail/Sailor-7B 

Sailor 7B Benchmarks

ARC: 49.83 vs 96.7 (so35)-48.5%
HellaSwag: 76.21 vs 95.3 (gpt4)-20%
MMLU: 54.84 vs 88.3 (so35)-37.9%
TruthfulQA: 40.12 vs 59 (gpt4)-32%
WinoGrande: 69.38 vs 87.5 (gpt4)-20.7%
GSM8K: 32.52 vs 96.4 (so35)-66.3%
LLME Score: 0.20393
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Sailor 7B (sail/Sailor-7B)

Sailor 7B Parameters and Internals

Model Type 
text-generation
Use Cases 
Areas:
research, commercial applications
Primary Use Cases:
question answering, commonsense reasoning in SEA languages
Supported Languages 
en (proficient), zh (proficient), id (proficient), th (proficient), vi (proficient), ms (proficient), lo (proficient)
Training Details 
Data Sources:
cerebras/SlimPajama-627B, Skywork/SkyPile-150B, allenai/MADLAD-400, cc100
Data Volume:
200B to 400B tokens depending on model sizes
Methodology:
Aggressive data deduplication and systematic experiments for language weighting
Input Output 
Input Format:
text
Accepted Modalities:
text
Output Format:
text
LLM NameSailor 7B
Repository 🤗https://huggingface.co/sail/Sailor-7B 
Base Model(s)  Qwen/Qwen1.5-7B   Qwen/Qwen1.5-7B
Model Size7b
Required VRAM15.4 GB
Updated2025-02-22
Maintainersail
Model Typeqwen2
Model Files  15.4 GB
Supported Languagesen zh id th vi ms lo
Model ArchitectureQwen2ForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.37.2
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Sailor 7B

Best Alternatives
Context / RAM
Downloads
Likes
Qwen2.5 7B Instruct 1M986K / 15.4 GB289038236
Qwen2.5 7B MixStock V0.1986K / 15.2 GB6823
Qwen2.5 7B RRP 1M986K / 15.2 GB2944
Qwen2.5 7B CelestialHarmony 1M986K / 14.8 GB1535
Qwen 2.5 7B Exp Sce986K / 15.2 GB282
COCO 7B Instruct 1M986K / 15.2 GB1059
SJT 7B V1.1986K / 14.8 GB1521
Q2.5 Instruct 1M Harmony986K / 15.2 GB611
Impish QWEN 7B 1M986K / 15.2 GB701
Qwen 2.5 7B Deep Stock V5986K / 15.2 GB302
Note: green Score (e.g. "73.2") means that the model is better than sail/Sailor-7B.

Rank the Sailor 7B Capabilities

🆘 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! 🌟

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? 43470 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227
  Support LLM Explorer