agentica-org/DeepSWE-Preview
Full fine-tune vs Qwen/Qwen3-32B. Fine-tune: 96% of the weights changed, across 64 of 64 layers.
Tensors changed: Embeddings 0/1, Norms 158/257, MLP 189/192, Attention 256/256, lm_head 1/1
5 derivatives checked
- bartowski/agentica-org_DeepSWE-Preview-GGUF: Same weights, converted
- mradermacher/DeepSWE-Preview-i1-GGUF: Quantization
- mradermacher/DeepSWE-Preview-GGUF: Quantization
- lmstudio-community/DeepSWE-Preview-GGUF: Quantization
- DevQuasar/agentica-org.DeepSWE-Preview-GGUF: Same weights, converted
4 other repos hold byte-for-byte identical weight files.