mergekit-community/LLaMa-3-8B-First-8-Layers

Same weights, repacked (fewer layers) vs NousResearch/Meta-Llama-3-8B. mergekit-community/LLaMa-3-8B-First-8-Layers matches part of NousResearch/Meta-Llama-3-8B (75 of its 291 tensors; the rest of NousResearch/Meta-Llama-3-8B has no counterpart). On the shared part: Same weights in different files (re-saved, re-sharded or new metadata).

Tensors changed: Embeddings 0/1, Norms 0/17, MLP 0/24, Attention 0/32, lm_head 0/1

On the Hub · Downloads on Model Pulse