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

Swallow 70B Instruct Hf by tokyotech-llm

 »  All LLMs  »  tokyotech-llm  »  Swallow 70B Instruct Hf   URL Share it on

  1952   
  Arxiv:2404.17733   Arxiv:2404.17790   Autotrain compatible   En   Endpoints compatible   Instruct   Ja   Llama   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf 

Swallow 70B Instruct Hf Benchmarks

ARC: 66.21 vs 96.7 (so35)-31.5%
HellaSwag: 85.14 vs 95.3 (gpt4)-10.7%
MMLU: 67.08 vs 88.3 (so35)-24%
TruthfulQA: 48 vs 59 (gpt4)-18.6%
WinoGrande: 82.08 vs 87.5 (gpt4)-6.2%
GSM8K: 45.94 vs 96.4 (so35)-52.3%
LLME Score: 0.20569
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Swallow 70B Instruct Hf (tokyotech-llm/Swallow-70b-instruct-hf)

Swallow 70B Instruct Hf Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
Research, Commercial applications
Limitations:
Models not tuned to ensure outputs align with human intent and safety considerations.
Additional Notes 
The models are continually pre-trained and instruction-tuned, emphasizing Japanese language capabilities.
Supported Languages 
supported_languages_list (Japanese, English), languages_details (The Swallow model has undergone continual pre-training with the addition of Japanese language data.)
Training Details 
Data Sources:
Japanese Wikipedia, RefinedWeb, Swallow Corpus, The Pile
Methodology:
Supervised fine-tuning (SFT) and instruction tuning using Anthropic HH-RLHF, Databricks Dolly 15-k, and OpenAssistant Conversations Dataset.
Model Architecture:
Refer to LLaMA-2 technical report for details on the model architecture.
Input Output 
Accepted Modalities:
text
Output Format:
Text
Performance Tips:
Model employs a tokenizer with a broadened vocabulary based on Japanese data, offering efficient text representation and faster inference.
Release Notes 
Version:
0.1
Date:
April 26, 2024
Notes:
Release of Swallow-7b-instruct-v0.1, Swallow-13b-instruct-v0.1, and Swallow-70b-instruct-v0.1.
Version:
N/A
Date:
March 2, 2024
Notes:
Release of Swallow-7b-plus-hf with twice as many Japanese tokens as Swallow-7b-hf.
Version:
N/A
Date:
February 4, 2024
Notes:
Release of Swallow-13b-NVE-hf.
Version:
N/A
Date:
January 26, 2024
Notes:
Release of Swallow-7b-NVE-hf, Swallow-7b-NVE-instruct-hf, Swallow-70b-NVE-hf, Swallow-70b-NVE-instruct-hf.
Version:
N/A
Date:
December 19, 2023
Notes:
Release of Swallow-7b-hf, Swallow-7b-instruct-hf, Swallow-13b-hf, Swallow-13b-instruct-hf, Swallow-70b-hf, Swallow-70b-instruct-hf.
LLM NameSwallow 70B Instruct Hf
Repository 🤗https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf 
Model Size70b
Required VRAM139.1 GB
Updated2024-12-22
Maintainertokyotech-llm
Model Typellama
Instruction-BasedYes
Model Files  4.9 GB: 1-of-29   4.7 GB: 2-of-29   5.0 GB: 3-of-29   5.0 GB: 4-of-29   4.7 GB: 5-of-29   4.7 GB: 6-of-29   4.7 GB: 7-of-29   5.0 GB: 8-of-29   5.0 GB: 9-of-29   4.7 GB: 10-of-29   4.7 GB: 11-of-29   4.7 GB: 12-of-29   5.0 GB: 13-of-29   5.0 GB: 14-of-29   4.7 GB: 15-of-29   4.7 GB: 16-of-29   4.7 GB: 17-of-29   5.0 GB: 18-of-29   5.0 GB: 19-of-29   4.7 GB: 20-of-29   4.7 GB: 21-of-29   4.7 GB: 22-of-29   5.0 GB: 23-of-29   5.0 GB: 24-of-29   4.7 GB: 25-of-29   4.7 GB: 26-of-29   4.7 GB: 27-of-29   5.0 GB: 28-of-29   4.0 GB: 29-of-29
Supported Languagesen ja
Model ArchitectureLlamaForCausalLM
Licensellama2
Context Length4096
Model Max Length4096
Transformers Version4.33.2
Tokenizer ClassLlamaTokenizer
Beginning of Sentence Token<s>
End of Sentence Token</s>
Unk Token<unk>
Vocabulary Size43176
Torch Data Typebfloat16

Quantized Models of the Swallow 70B Instruct Hf

Model
Likes
Downloads
VRAM
Swallow 70B Instruct GGUF88629 GB
Swallow 70B Instruct GPTQ34335 GB
Swallow 70B Instruct AWQ14037 GB

Best Alternatives to Swallow 70B Instruct Hf

Best Alternatives
Context / RAM
Downloads
Likes
... Chat 1048K Chinese Llama3 70B1024K / 141.9 GB32225
... 3 70B Instruct Gradient 1048K1024K / 141.9 GB285121
Llama3 Function Calling 1048K1024K / 141.9 GB31
...a 3 70B Instruct Gradient 524K512K / 141.9 GB5123
...a 3 70B Instruct Gradient 262K256K / 141.9 GB20255
...ama 3 70B Arimas Story RP V2.0256K / 141.1 GB503
...ama 3 70B Arimas Story RP V1.6256K / 141.2 GB150
...ama 3 70B Arimas Story RP V1.5256K / 141.2 GB302
...a 3.1 Nemotron 70B Instruct HF128K / 141.9 GB1534001927
Meta Llama 3.1 70B Instruct128K / 141.9 GB521937513
Note: green Score (e.g. "73.2") means that the model is better than tokyotech-llm/Swallow-70b-instruct-hf.

Rank the Swallow 70B Instruct Hf 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? 40123 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 v20241217
  Support LLM Explorer