ChuGyouk/DR-Tulu-SFT-4B
ChuGyouk/DR-Tulu-SFT-4B is a 4 billion parameter instruction-tuned causal language model, fine-tuned from Qwen/Qwen3-4B. This model is specifically adapted using the rl-research/dr-tulu-sft-data dataset, offering a 32768 token context length. It is designed for general-purpose conversational AI tasks, leveraging its fine-tuning for improved instruction following.
Loading preview...
Model Overview
ChuGyouk/DR-Tulu-SFT-4B is a 4 billion parameter language model, fine-tuned from the base Qwen/Qwen3-4B architecture. This model has been specifically adapted through supervised fine-tuning (SFT) using the rl-research/dr-tulu-sft-data dataset, aiming to enhance its instruction-following capabilities and general conversational performance. It supports a substantial context length of 32768 tokens, allowing for processing and generating longer sequences of text.
Training Details
The model was trained with a learning rate of 5e-05 over 5 epochs, utilizing a distributed multi-GPU setup with a total batch size of 128 (achieved via gradient accumulation steps of 128). The AdamW optimizer with standard betas and a cosine learning rate scheduler with a 0.1 warmup ratio were employed. The training leveraged Transformers 4.52.4, Pytorch 2.13.0+cu130, Datasets 3.6.0, and Tokenizers 0.21.1.
Potential Use Cases
- General-purpose conversational AI: Responding to user queries and engaging in dialogue.
- Instruction following: Executing specific commands or generating text based on detailed prompts.
- Text generation: Creating coherent and contextually relevant text for various applications.