HillPhelmuth/Qwen3-4B-GRPO-MathsFT
TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:May 18, 2025License:apache-2.0Architecture:Transformer Open Weights Warm
HillPhelmuth/Qwen3-4B-GRPO-MathsFT is a 2 billion parameter Qwen3 model developed by HillPhelmuth, fine-tuned from unsloth/Qwen3-1.7B-Base. This model was trained using Unsloth and Huggingface's TRL library, enabling 2x faster training. Its primary differentiator and strength are not explicitly detailed in the provided README, but its Qwen3 base suggests general language understanding capabilities.
Loading preview...