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

ALMA 7B Ja V2 by webbigdata

 »  All LLMs  »  webbigdata  »  ALMA 7B Ja V2   URL Share it on

  797   

ALMA 7B Ja V2 is an open-source language model by webbigdata. Features: 7b LLM, VRAM: 13.5GB, Context: 4K, License: llama2, HF Score: 47.9, LLM Explorer Score: 0.14, Arc: 52.4, HellaSwag: 77.9, MMLU: 44.7, TruthfulQA: 38.7, WinoGrande: 73.4.

  Arxiv:2309.11674   Cs   De   Deploy:azure   En   Is   Ja   Llama   Pytorch   Region:us   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/webbigdata/ALMA-7B-Ja-V2 

ALMA 7B Ja V2 Benchmarks

ARC: 52.39 vs 96.7 (so35)-45.8%
HellaSwag: 77.92 vs 95.3 (gpt4)-18.2%
MMLU: 44.72 vs 88.3 (so35)-49.4%
TruthfulQA: 38.66 vs 59 (gpt4)-34.5%
WinoGrande: 73.4 vs 87.5 (gpt4)-16.1%
LLME Score: 0.1417
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
ALMA 7B Ja V2 (webbigdata/ALMA-7B-Ja-V2)
🌟 Advertise your project 🚀

ALMA 7B Ja V2 Parameters and Internals

Model Type 
machine translation
Use Cases 
Primary Use Cases:
Japanese-English translation, English-Japanese translation
Limitations:
Performance slightly reduced in GPTQ version, Reduced ability for languages other than Japanese and English
Additional Notes 
C3TR-Adapter and GPTQ versions available with different performance and resource usages.
Supported Languages 
ja (Japanese), en (English), de (German), zh (Chinese), is (Icelandic), cs (Czech)
Training Details 
Data Sources:
monolingual data, high-quality parallel data
Methodology:
Two-step fine-tuning on monolingual and parallel data
LLM NameALMA 7B Ja V2
Repository 🤗https://huggingface.co/webbigdata/ALMA-7B-Ja-V2 
Model Size7b
Required VRAM13.5 GB
Updated2026-03-29
Maintainerwebbigdata
Model Typellama
Model Files  10.0 GB: 1-of-2   3.5 GB: 2-of-2   10.0 GB: 1-of-2   3.5 GB: 2-of-2
Supported Languagesja en de is zh cs
Model ArchitectureLlamaForCausalLM
Licensellama2
Context Length4096
Model Max Length4096
Transformers Version4.34.0
Tokenizer ClassLlamaTokenizer
Padding Token<unk>
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to ALMA 7B Ja V2

Best Alternatives
Context / RAM
Downloads
Likes
1241024K / 16.1 GB930
1621024K / 16.1 GB600
1571024K / 16.1 GB1010
1181024K / 16.1 GB150
A5.41024K / 16.1 GB120
A6 L1024K / 16.1 GB2010
A3.41024K / 16.1 GB130
A2.41024K / 16.1 GB120
M1024K / 16.1 GB1270
2 Very Sci Fi1024K / 16.1 GB3170
Note: green Score (e.g. "73.2") means that the model is better than webbigdata/ALMA-7B-Ja-V2.

Rank the ALMA 7B Ja V2 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? 52721 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.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a
  Support LLM Explorer