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

ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K by mmnga

 »  All LLMs  »  mmnga  »  ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K   URL Share it on

  36   
  Arxiv:2307.09288   Arxiv:2308.12950   4bit   Autotrain compatible   Codegen   Endpoints compatible   Gptq   Instruct   Ja   Llama   Llama2   Quantized   Region:us
Model Card on HF 🤗: https://huggingface.co/mmnga/ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k 

ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K Benchmarks

LLME Score: 0.15317
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K (mmnga/ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k)

ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K Parameters and Internals

Model Type 
Text Generation, Multimodal
Use Cases 
Areas:
Research, Commercial applications
Additional Notes 
Includes calibration with specific Japanese datasets to enhance performance on Japanese tasks.
Supported Languages 
Japanese (high)
Training Details 
Data Sources:
izumi-lab/wikipedia-ja-20230720, ELYZA-tasks-100
Data Volume:
1k samples and 200 additional samples
Methodology:
Fine-tuning with Japanese calibration set
Context Length:
1024
Model Architecture:
CodeLlama
Input Output 
Input Format:
Textual prompts in Japanese
Accepted Modalities:
Text
Output Format:
Generated Japanese text and code
LLM NameELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K
Repository 🤗https://huggingface.co/mmnga/ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k 
Model Size7b
Required VRAM3.9 GB
Updated2024-12-22
Maintainermmnga
Model Typellama
Instruction-BasedYes
Model Files  3.9 GB
Supported Languagesja
GPTQ QuantizationYes
Quantization Typegptq|4bit
Generates CodeYes
Model ArchitectureLlamaForCausalLM
Licensellama2
Context Length16384
Model Max Length16384
Transformers Version4.35.1
Tokenizer ClassCodeLlamaTokenizer
Vocabulary Size32016
Torch Data Typebfloat16

Best Alternatives to ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K

Best Alternatives
Context / RAM
Downloads
Likes
CodeLlama 7B Instruct GPTQ16K / 3.9 GB55046
...ruct Solidity Bnb 4bit Smashed16K / 4.2 GB120
...B Instruct Hf Bnb 4bit Smashed16K / 4.2 GB220
CodelLama7B Inst DPO 7K Mlx16K / 4.2 GB82
...eLlama 7B Instruct Hf 4bit MLX16K / 4.2 GB121
...6.7B Instruct 8.0bpw H8 EXL2 216K / 6.8 GB92
...6.7B Instruct 3.0bpw H6 EXL2 216K / 2.8 GB91
... 7B Instruct Nf4 Fp16 Upscaled16K / 13.5 GB160
CodeLlama 7B Instruct Fp1616K / 13.5 GB398
...Llama 7B Instruct Bf16 Sharded16K / 13.5 GB161
Note: green Score (e.g. "73.2") means that the model is better than mmnga/ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k.

Rank the ELYZA Japanese CodeLlama 7B Instruct GPTQ Calib Ja 1K 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? 40066 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