Liangmingxin/ThetaWave-7B-sft

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Jan 24, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

Liangmingxin/ThetaWave-7B-sft is a 7 billion parameter language model, fine-tuned from freecs/ThetaWave-7B using Supervised Fine-Tuning (SFT) on the Open-Orca/SlimOrca datasets. This model is designed for general conversational tasks, leveraging its SFT training for improved instruction following. It currently uses Mistral's chat template and does not natively support system prompts without performance degradation.

Loading preview...

Liangmingxin/ThetaWave-7B-sft Overview

Liangmingxin/ThetaWave-7B-sft is a 7 billion parameter language model derived from the freecs/ThetaWave-7B base model. It has undergone Supervised Fine-Tuning (SFT) using the Open-Orca/SlimOrca datasets, aiming to enhance its instruction-following capabilities and general conversational performance.

Key Characteristics

  • Base Model: Fine-tuned from freecs/ThetaWave-7B.
  • Training Data: Utilizes Open-Orca/SlimOrca datasets for SFT.
  • Chat Template: Employs Mistral's chat template.
  • System Prompt Support: Currently, the model does not natively support system_prompt within its default chat template. While manual modification is possible, it may lead to degraded performance. Future releases are planned to switch to the ChatML template to address this limitation.

Potential Use Cases

  • General Conversational AI: Suitable for chatbots and interactive applications requiring instruction-tuned responses.
  • Instruction Following: Designed to respond effectively to user instructions due to its SFT training.

Further details regarding the model's specifics and performance are anticipated in future updates.

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