TheDrummer/Big-Alice-28B-v1
Full fine-tune (fewer layers) vs SillyTilly/ServiceNow-AI-Apriel-Nemotron-15b-Thinker-Chatml. SillyTilly/ServiceNow-AI-Apriel-Nemotron-15b-Thinker-Chatml matches part of TheDrummer/Big-Alice-28B-v1 (453 of its 903 tensors; the rest of TheDrummer/Big-Alice-28B-v1 has no counterpart). On the shared part: Fine-tune: 91% of the weights changed, across 50 of 50 layers.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 60/101, MLP 150/150, Attention 200/200
3 derivatives checked
- bartowski/TheDrummer_Big-Alice-28B-v1-GGUF: Same weights, converted
- mradermacher/Big-Alice-28B-v1-i1-GGUF: Quantization
- mradermacher/Big-Alice-28B-v1-GGUF: Quantization