DeepMount00/Mistral-Ita-7b

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Nov 8, 2023License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

DeepMount00/Mistral-Ita-7b is a 7 billion parameter language model developed by DeepMount00, based on the Mistral-7B-v0.1 architecture with an 8192-token context length. This model is specifically specialized and fine-tuned for the Italian language, excelling in Italian text generation tasks. It offers a quantized 4-bit version for efficient deployment on resource-constrained devices, making it suitable for Italian-centric applications requiring optimized performance.

Loading preview...

Overview

DeepMount00/Mistral-Ita-7b is a 7 billion parameter language model built upon the Mistral-7B-v0.1 base architecture. Its primary specialization is the Italian language, making it highly effective for tasks involving Italian text generation and understanding. The model maintains an 8192-token context length, suitable for processing moderately long Italian texts.

Key Capabilities

  • Italian Language Specialization: Fine-tuned specifically for Italian, enhancing its performance and fluency in the language.
  • Efficient Deployment: A quantized 4-bit GGUF version is available, significantly reducing memory usage and potentially increasing inference speed, which is beneficial for deployment on devices with limited computational resources.
  • Competitive Performance: Evaluated on Italian-specific benchmarks, achieving an average accuracy normalized score of 0.5866 across hellaswag_it, arc_it, and m_mmlu_it datasets.

Good For

  • Applications requiring high-quality Italian text generation.
  • Deploying LLMs on edge devices or environments with limited memory due to the availability of a 4-bit quantized version.
  • Research and development focused on Italian natural language processing.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p