alignment-handbook/zephyr-7b-sft-full

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Nov 9, 2023License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

The alignment-handbook/zephyr-7b-sft-full is a 7 billion parameter language model fine-tuned from Mistral-7B-v0.1. Developed by alignment-handbook, this model is specifically trained on the HuggingFaceH4/ultrachat_200k dataset. It is optimized for supervised fine-tuning tasks, demonstrating a validation loss of 0.9353. This model is suitable for applications requiring a robust base model with enhanced conversational capabilities.

Loading preview...

Model Overview

The alignment-handbook/zephyr-7b-sft-full is a 7 billion parameter language model derived from the Mistral-7B-v0.1 architecture. This model has undergone supervised fine-tuning (SFT) using the comprehensive HuggingFaceH4/ultrachat_200k dataset, which is designed to enhance conversational abilities and alignment.

Key Training Details

  • Base Model: mistralai/Mistral-7B-v0.1
  • Dataset: HuggingFaceH4/ultrachat_200k
  • Training Objective: Supervised Fine-Tuning (SFT)
  • Validation Loss: Achieved 0.9353 on the evaluation set.
  • Hyperparameters: Trained with a learning rate of 2e-05, a total batch size of 128, and 1 epoch using an Adam optimizer with cosine learning rate scheduling.

Intended Use Cases

This model is primarily intended for applications that benefit from a fine-tuned Mistral-7B variant with improved conversational understanding and generation. Its training on a large-scale chat dataset suggests suitability for:

  • Chatbots and Conversational AI: Engaging in dialogue and responding to user queries.
  • Instruction Following: Executing tasks based on explicit instructions.
  • General Text Generation: Producing coherent and contextually relevant text in various formats.

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