MINZIK77/lm-sft-ultrachat-3b-ckpts
MINZIK77/lm-sft-ultrachat-3b-ckpts is a 3.1 billion parameter causal language model fine-tuned from Qwen/Qwen2.5-3B-Instruct. Developed by MINZIK77, this model leverages Supervised Fine-Tuning (SFT) using the TRL framework. It is designed for general text generation tasks, building upon the base capabilities of the Qwen2.5 architecture with a 32768 token context length.
Loading preview...
Model Overview
MINZIK77/lm-sft-ultrachat-3b-ckpts is a 3.1 billion parameter language model, fine-tuned from the base model Qwen/Qwen2.5-3B-Instruct. This model was developed by MINZIK77 and utilizes the TRL (Transformers Reinforcement Learning) framework for its training process, specifically employing Supervised Fine-Tuning (SFT).
Key Capabilities
- Instruction Following: Inherits and enhances instruction-following capabilities from its base Qwen2.5-3B-Instruct model.
- Text Generation: Capable of generating coherent and contextually relevant text based on user prompts.
- Context Length: Supports a substantial context window of 32768 tokens, allowing for processing longer inputs and generating more extended responses.
Training Details
The model underwent Supervised Fine-Tuning (SFT) using the TRL framework (version 1.7.1). The training environment included Transformers version 4.57.6, Pytorch 2.10.0, Datasets 4.7.0, and Tokenizers 0.22.2.
Use Cases
This model is suitable for various text generation tasks where a compact yet capable model is required. Its fine-tuning process aims to improve its performance on conversational and instruction-based prompts, making it a good candidate for:
- General-purpose chatbots
- Content creation
- Question answering
- Summarization