POLARIS-Project/Polaris-4B-Preview
Full fine-tune vs Qwen/Qwen3-4B. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 101/145, MLP 108/108, Attention 144/144
9 derivatives checked
- dzur658/Polaris-4B-Preview-IQ-GGUF: Same weights, converted
- bartowski/POLARIS-Project_Polaris-4B-Preview-GGUF: Same weights, converted
- gabriellarson/Polaris-4B-Preview-GGUF: Same weights, converted
- prithivMLmods/Polaris-4B-Preview-F32-GGUF: Same weights, converted
- mradermacher/Polaris-4B-Preview-GGUF: Same weights, converted
- ertghiu256/Qwen3-4b-tcomanr-merge: Full fine-tune
- ertghiu256/Qwen3-4b-tcomanr-merge-v2.1: Full fine-tune
- ertghiu256/Qwen3-4b-tcomanr-merge-v2: Full fine-tune
- ZeroXClem/Qwen3-4B-Valiant-Polaris: Full fine-tune