SteelStorage/LD-Zephyria-37b
Partial fine-tune (fewer layers) vs unsloth/Mistral-Small-Instruct-2409. unsloth/Mistral-Small-Instruct-2409 matches part of SteelStorage/LD-Zephyria-37b (507 of its 858 tensors; the rest of SteelStorage/LD-Zephyria-37b has no counterpart). On the shared part: Partial change: 14% of the weights changed (attention, mlp, norms), in 8 of 56 layers.
Tensors changed: Embeddings 0/1, Norms 16/113, MLP 24/168, Attention 32/224, lm_head 0/1
3 derivatives checked
- mradermacher/LD-Zephyria-37b-i1-GGUF: Quantization
- mradermacher/LD-Zephyria-37b-GGUF: Quantization
- athirdpath/LD-Zephyria-37b_HEALED: LoRA-style