NovaSky-AI/Sky-T1-32B-Preview
Full fine-tune vs Qwen/Qwen2.5-32B-Instruct. Fine-tune: 98% of the weights changed, across 64 of 64 layers.
Tensors changed: Embeddings 0/1, Norms 127/129, MLP 192/192, Attention 448/448, lm_head 1/1
10 derivatives checked
- bartowski/Sky-T1-32B-Preview-GGUF: Same weights, converted
- gaianet/Sky-T1-32B-Preview-GGUF: Same weights, converted
- mradermacher/Sky-T1-32B-Preview-i1-GGUF: Quantization
- featherless-ai-quants/NovaSky-AI-Sky-T1-32B-Preview-GGUF: Quantization
- mradermacher/Sky-T1-32B-Preview-GGUF: Quantization
- second-state/Sky-T1-32B-Preview-GGUF: Same weights, converted
- NovaSky-AI/Sky-T1-32B-Flash: Full fine-tune
- radna/S1.1-Deepseek-R1-T1-32B: Full fine-tune
- ltn18/QwQ-SkyT1-32B-Preview: Full fine-tune
- Sorawiz/KunouSky-32B: Full fine-tune
1 other repos hold byte-for-byte identical weight files.