MaziyarPanahi/Llama-3-13B-Instruct-v0.1
Partial fine-tune (fewer layers) vs meta-llama/Meta-Llama-3-8B-Instruct. meta-llama/Meta-Llama-3-8B-Instruct matches part of MaziyarPanahi/Llama-3-13B-Instruct-v0.1 (291 of its 507 tensors; the rest of MaziyarPanahi/Llama-3-13B-Instruct-v0.1 has no counterpart). On the shared part: Partial change: 65% of the weights changed (attention, mlp, norms), in 24 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 48/65, MLP 72/96, Attention 96/128, lm_head 0/1
3 derivatives checked
- MaziyarPanahi/Llama-3-13B-Instruct-v0.1-GGUF: Same weights, converted
- mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF: Quantization
- mradermacher/Llama-3-13B-Instruct-v0.1-GGUF: Quantization