QuantFactory/arco-reflection-v4-GGUF
Full fine-tune vs h2oai/h2o-danube3-500m-base. QuantFactory/arco-reflection-v4-GGUF is a GGUF (Q8_0) of different weights. Fine-tune: 100% of the weights changed, across 16 of 16 layers.
Tensors changed: lm_head 1/1, Embeddings 1/1, Norms 0/33, MLP 48/48, Attention 64/64