merged_ja_slerp_Llama_3_ELYZA_JP_8B_v0_20250605

This model was created by merging multilingual language models using the SLERP method.

Model Details

• Method: SLERP
• Language: ja
• Base Models:
  • Language Model: unknown
  • Knowledge Model: unknown
• Created: 2025-06-07

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("NTIS/merged_ja_slerp_Llama_3_ELYZA_JP_8B_v0_20250605")
tokenizer = AutoTokenizer.from_pretrained("NTIS/merged_ja_slerp_Llama_3_ELYZA_JP_8B_v0_20250605")

Training Details

This model was created using model merging techniques to combine the language capabilities of specialized multilingual models with general knowledge models.

Limitations

• This model inherits limitations from its base models
• Performance may vary across different languages and tasks
• Requires evaluation for specific use cases

Citation

@misc{merged_ja_slerp_Llama_3_ELYZA_JP_8B_v0_20250605,
  title={Multilingual Model Merging: merged_ja_slerp_Llama_3_ELYZA_JP_8B_v0_20250605},
  year={2025},
  url={https://huggingface.co/NTIS/merged_ja_slerp_Llama_3_ELYZA_JP_8B_v0_20250605}
}