ICTNLP/Llama-3.1-8B-Omni
Full fine-tune (fewer layers) vs meta-llama/Llama-3.1-8B-Instruct. meta-llama/Llama-3.1-8B-Instruct matches part of ICTNLP/Llama-3.1-8B-Omni (291 of its 804 tensors; the rest of ICTNLP/Llama-3.1-8B-Omni has no counterpart). On the shared part: Fine-tune: 93% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 12/65, MLP 96/96, Attention 128/128, lm_head 1/1
2 other repos hold byte-for-byte identical weight files.