ggml-org/llama.cpp

Issues

• Misc. bug: The KV cache is sometimes truncated incorrectly when making v1/chat/completions API calls

  #11970 opened 18 days ago by vnicolici
  41

• Misc. bug: convert_hf_to_gguf failed

  #11991 opened 11 days ago by JSXGQ
  3

• Regression. Unable to run any model. CRASH!!!

  #12075 opened 11 days ago by acbits
  9

• Eval bug: does llama.cpp support Intel AMX instruction? how to enable it

  #12003 opened 17 days ago by montagetao
  6

• Worse performance on newer Linux kernel version

  #12086 opened 11 days ago by cxxr2
  1

• Misc. bug: llama-cli llama_backend_free may not free all the gpu memory

  #12057 opened 13 days ago by GaoXiangYa
  3

• Compile bug: How to build llama.android example with -DGGML_VULKAN=ON through android studio.

  #12085 opened 12 days ago by gaykawadpk
  0

• Misc. bug: Loop range computation question of Vulkan matmul shaders

  #12082 opened 12 days ago by blurSong
  1

• Eval bug: MUSA error: operation not supported

  #12077 opened 12 days ago by yeungtuzi
  3

• Eval bug: getting assertion error when trying to use a gguf quantized model at inference "GGML_ASSERT(n_outputs_enc > 0 && "call llama_encode() first") failed"

  #12080 opened 12 days ago by Vedapani0402
  0

• Misc. bug: Concurrency Limitation: Only 6 Inferences Run Simultaneously When Setting `--parallel` > 6

  #12013 opened 12 days ago by karanotsingyu
  6

• Eval bug: TikTokenTokenizer has no attribute vocab

  #12044 opened 14 days ago by zhanghui-china
  5

• Eval bug: Several models producing gibberish

  #12012 opened 12 days ago by iamangus
  15

• Eval bug: granite-vision-3.1-2b-preview ERROR:hf-to-gguf:Model LlavaNextForConditionalGeneration is not supported

  #12053 opened 13 days ago by gnusupport
  2

• Misc. bug: Web-UI now unusably slow - over network or locally.

  #12026 opened 16 days ago by clort81
  5

• Eval bug: llama.cpp:8910: GGML_ASSERT(strcmp(embd->name, "result_norm") == 0 && "missing result_output tensor") failed

  #12074 opened 12 days ago by 79154gb
  0

• Eval bug: std::filesystem::__cxx11::filesystem_error

  #11962 opened 18 days ago by gnusupport
  2

• Eval bug: CANNOT LINK EXECUTABLE "./llama-cli": library "libomp.so" not found: needed by main executable

  #11979 opened 17 days ago by Krallbe68
  6

• Misc. bug: ggml-backend.cpp:746: pre-allocated tensor (cache_k_l0 (view) (copy of cache_k_l0 (view))) in a buffer (Vulkan0) that cannot run the operation (CPY)

  #12045 opened 14 days ago by simonchen
  1

• Eval bug: GGML_ASSERT(hparams.n_embd_head_k % ggml_blck_size(type_k) == 0) failed

  #12033 opened 13 days ago by AbdullahMPrograms
  2

• Misc. bug: Rpc-server does not use opencl backend on Android.

  #11957 opened 19 days ago by belog2867
  6

• Misc. bug: cannot scroll to right side when input too long

  #12054 opened 13 days ago by gnusupport
  0

• Eval bug: Error when converting moonlight from bf16 to q4km

  #12040 opened 15 days ago by qiyuxinlin
  2

• Misc. bug: --no-context-shift OR --context-shift ?

  #12038 opened 15 days ago by simonchen
  1

• Compile bug: llama.cpp-b4749/ggml/src/ggml-cpu/ggml-cpu-quants.c:5141:26: error: initialization of ‘uint32_t *’ {aka ‘unsigned int *’} from incompatible pointer type ‘const uint8_t (*)[12]’ {aka ‘const unsigned char (*)[12]’} [-Wincompatible-pointer-types]

  #12050 opened 14 days ago by Arniiiii
  0

• Eval bug: context shift is disabled

  #11974 opened 18 days ago by deific
  3

• Eval bug: CPU usage is abnormal when running deepseek-r1-671B-Q4_0 weights in Atlas 800T a2 and NPU device。

  #11966 opened 18 days ago by woshidahunzi1
  2

• llama-cli misbehaving (changed?)

  #12036 opened 15 days ago by 0wwafa
  2

• Feature Request: 推理minicpmv时，encoding_image_with_clip耗时很久

  #11941 opened 20 days ago by EnzhiZhou
  4

• Misc. bug: llama-cli '--log-disable' parameter omits response

  #11983 opened 17 days ago by nmandic78
  2

• Misc. bug: add tool_calls id in response in server

  #11992 opened 17 days ago by henryclw
  1

• Eval bug: unknown pre-tokenizer type: 'deepseek-r1-qwen'

  #12021 opened 16 days ago by wr131
  4

• Misc. bug: llama-run segmentation fault

  #12022 opened 15 days ago by benoitf
  8

• [Feature]: SOC_VERSION ascend310b1 does not support

  #11978 opened 15 days ago by Cikaros
  0

• Add option to build CUDA backend without Flash attention

  #11946 opened 15 days ago by slaren
  2

• [CANN] Compile bug: no matching function for call to 'CastIntrinsicsImpl' Ascend NPU issues specific to Ascend NPUs

  #12010 opened 16 days ago by Cikaros
  3

• GGML to GGUF FAIL Quantized tensor bytes per row (5120) is not a multiple of Q2_K type size (84)

  #11976 opened 18 days ago by chokoon123
  0

• Maybe it would better to have a diagram to show how llama.cpp process inferences

  #11967 opened 17 days ago by yinuu
  2

• Misc. bug: `json_schema` under `response_format` is not working on OpenAI compatible API endpoint `v1/chat/completions`

  #11988 opened 17 days ago by henryclw
  1

• Feature Request: add Kernel level verbose option

  #11985 opened 17 days ago by 0400H
  0

• Eval bug: Unexpected empty grammar stack after accepting piece: <｜tool_calls_begin｜> on DeepSeek-R1-Distill-Qwen-32B

  #11938 opened 20 days ago by chgjin
  1

• tensor 'blk.25.ffn_down.weight' has invalid ggml type 42 (NONE)

  #11975 opened 18 days ago by evaninf
  0

• Misc. bug: Sporadic MUL_MAT Failures in test-backend-ops for Nvidia backend

  #11972 opened 18 days ago by ShanoToni
  1

• Eval bug: Ram boom after using llama-bench with cuda12.8 and deepseekr1q6

  #11965 opened 18 days ago by Xxianna
  0

• bamba

  #11955 opened 19 days ago by werruww
  0

• Misc. bug: Segmentation fault when importing model to opencl buffer

  #11953 opened 19 days ago by zhouzengming
  0

• Eval bug: llama.cpp Incorrectly Parses and Reports sprintf Calls in C++ Code

  #11951 opened 19 days ago by perdubug
  0

• Misc. bug: hipGraph causes a crash in hipGraphDestroy

  #11949 opened 19 days ago by IMbackK
  0

• Eval bug: Segmentation fault with Docker ROCm image "full-rocm"

  #11947 opened 19 days ago by JFingerle
  0

• Feature Request: dynamic speculation (i.e. dynamic draft-max)

  #11933 opened 20 days ago by fredlas
  2