bamec66557/MISCHIEVOUS-12B-Mix_0.5v
Partial fine-tune vs bamec66557/MISCHIEVOUS-12B-Mix_0.3v. Partial change: 40% of the weights changed (attention, embeddings, mlp, norms), in 19 of 40 layers.
Tensors changed: lm_head 0/1, Embeddings 1/1, Norms 16/81, MLP 45/120, Attention 72/160
3 derivatives checked
- mradermacher/MISCHIEVOUS-12B-Mix_0.5v-i1-GGUF: Quantization
- mradermacher/MISCHIEVOUS-12B-Mix_0.5v-GGUF: Quantization
- MaziyarPanahi/MISCHIEVOUS-12B-Mix_0.5v-GGUF: Same weights, converted