heavytail/kullm-mistral-S

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kLicense:apache-2.0Architecture:Transformer0.0K Open Weights Cold

heavytail/kullm-mistral-S is a causal language model developed by heavytail, fine-tuned from the Mistral-7B-Instruct-v0.2 base model. This model is specifically trained on the KULLM dataset and additional hand-crafted instruction data. It is designed for instruction-following tasks, leveraging its 7 billion parameters for general-purpose conversational AI.

Loading preview...

Model Overview

heavytail/kullm-mistral-S is an instruction-tuned causal language model developed by heavytail. It is built upon the robust Mistral-7B-Instruct-v0.2 base model, inheriting its 7 billion parameters and foundational capabilities.

Training and Data

This model has undergone fine-tuning using a specialized combination of datasets:

  • KULLM dataset: A key component of its training, contributing to its unique instruction-following characteristics.
  • Hand-crafted instruction data: Additional curated data designed to enhance its ability to understand and execute diverse instructions.

Key Capabilities

  • Instruction Following: Optimized for responding to and executing user instructions effectively.
  • General-Purpose Conversational AI: Suitable for a wide range of dialogue-based applications due to its instruction-tuned nature.

Usage

Developers can easily integrate kullm-mistral-S into their projects using the Hugging Face transformers library, leveraging AutoModelForCausalLM and AutoTokenizer for inference. The model supports torch.float16 for efficient computation and device_map='auto' for flexible deployment across available hardware.

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