ConvexAI/Pelican-9b-v0.1
Full fine-tune vs flemmingmiguel/MBX-7B. Fine-tune: 95% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 64/65, MLP 93/96, Attention 124/128, lm_head 0/1
Full fine-tune vs flemmingmiguel/MBX-7B. Fine-tune: 95% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 64/65, MLP 93/96, Attention 124/128, lm_head 0/1