thinletter/harrier-0.6b-query-clients
Full fine-tune vs microsoft/harrier-oss-v1-0.6b. thinletter/harrier-0.6b-query-clients is a GGUF (Q2_K/F32) of different weights. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 0/113, MLP 84/84, Attention 112/112