Repository Issue Activity (beta)

#2231 Request for Gemma 3 4B & Gemma 4 Integration Support with llama-cpp-python

#1813 [Feature request] High-level API support for DRY and XTC samplers

#1495 _LlamaModel.metadata() does not return `tokenizer.ggml.tokens`

#1360 [REQUEST] Accept raw token IDs in `stop` parameter

#1335 KV cache quantization fails with GGML_ASSERT

#878 Possible bug with greedy sampling code: Llama._sample()

#2362 Long whitespace tokens detokenize to empty bytes, breaking tokenize/detokenize round-trip

#1366 Pip Install faling in colab

#2361 [Bug] DFlash 2 speculative decoding crashes and fails in Python due to missing ctx_other and non-autoregressive graph execution

#2342 Llama.close() doesn't clean up chat_handler causing null pointer on second load of a vision model

#2013 Can't install with GPU support with Cuda toolkit 12.9 and Cuda 12.9

#2352 uv add llama-cpp-python wheels fails for versions above 0.3.30

#1169 Docker llama-cpp libcuda.so.1: cannot open shared object file: No such file or directory

#2341 Support sm100 and sm120 in CUDA pre‑built wheels

#2029 Access Violation issue facing for exe created using pyinstaller

#2211 Llama.embed() calls LlamaBatch.add_sequence with old 3-arg signature; missing logits_array

#1290 OpenAI compatible API not able to connect to llama.cpp web server

#247 No CUDA toolset found.

#2323 I am passionate for Edge AI and would like to try out Qwen 3.5 2B model using llama-cpp-python. But as of now, support for Qwen 3.5 is pending. I saw there's a PR (https://github.com/abetlen/llama-cpp-python/pull/2132) adding support. Any idea when will this be merged?

#635 Documentation of server command line parameters.

#2175 How to use "Gemma-4-E4B-it-heretic-GGUF" ?

#1095 Output the final settings used to load the llm

#1938 Specifying additional_files for model files in directory adds additional copy of directory to download URL

#2103 Pre-built wheels for Python 3.14 and 3.14 free-threaded

#2224 Feature Request: Gemma Multimodal Support