Issues

• Support for EmbeddingGemma 2 (gemma-embedding2 architecture)

  AmpereComputingAI/llama.cpp#22 · Filter by author dhaern opened on Oct 6, 2026

• Illegal instruction on Neoverse-N1 with latest Docker image; 3.4.2 works

  AmpereComputingAI/llama.cpp#21 · Filter by author BitRevenant opened on Sep 19, 2026

• /start.sh: line 3: 10 Illegal instruction (core dumped) /llm/llama-server "$@"

  AmpereComputingAI/llama.cpp#20 · Filter by author raphaelallain opened on Jul 1, 2026
  3

• cache-type-k/v not support Q4_K_4 and Q8R16

  AmpereComputingAI/llama.cpp#19 · Filter by author ed-hch opened on May 20, 2026

• Not Issue. To whom may concern. You can quantize granite 4.1 on this version

  AmpereComputingAI/llama.cpp#18 · Filter by author ed-hch opened on May 5, 2026

• UPDATE THE LLAMA.CPP

  AmpereComputingAI/llama.cpp#17 · Filter by author raphaelallain opened on Apr 22, 2026
  1

• Add support Gemma 3 models

  AmpereComputingAI/llama.cpp#12 · Filter by author semae opened on Apr 16, 2025
  2

• Publish readme details how to run an online serving llama.server

  AmpereComputingAI/llama.cpp#8 · Filter by author amgowda-oci opened on Sep 24, 2024

• Source code?

  AmpereComputingAI/llama.cpp#6 · Filter by author nilleb opened on Aug 31, 2024
  3

• Code on github

  AmpereComputingAI/llama.cpp#3 · Filter by author cduk opened on Aug 12, 2024

• llama_model_load: error loading model: done_getting_tensors: wrong number of tensors; expected 292, got 291

  AmpereComputingAI/llama.cpp#2 · Filter by author cduk opened on Aug 12, 2024