newmindai/Mursit-Large-TR-Retrieval
Full fine-tune vs newmindai/Mursit-Large. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 2/2, Norms 56/56, Attention 56/56, MLP 56/56
Full fine-tune vs newmindai/Mursit-Large. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 2/2, Norms 56/56, Attention 56/56, MLP 56/56