ggerganov/ggml

Issues

• User defined operation

  #836 opened 9 days ago by Francis235
  0

• issues about YaRN

  #835 opened 9 days ago by foldl
  0

• Completion of error handling

  #834 opened 10 days ago by elfring
  0

• ggml how to compute depthwise conv

  #833 opened 12 days ago by Francis235
  0

• GGML Fragmentation Issue

  #830 opened 14 days ago by zhouwg
  0

• `ggml_flip` or `ggml_pad_reflect`?

  #819 opened 21 days ago by PABannier
  1

• GGML_MAX_NAME is too small.

  #825 opened 17 days ago by IntptrMax
  1

• Is there interest in `ggml_upscale_to_shape` supporting non integer scaling factors?

  #812 opened 17 days ago by balisujohn
  0

• Proposing To Add Naming Convention For GGUF files in documents

  #820 opened 20 days ago by mofosyne
  1

• ggml vs onnxruntime on SOC chip

  #810 opened a month ago by Francis235
  0

• ggml vs Qualcomm SNPE inference engine on qualcomm soc

  #809 opened a month ago by Francis235
  0

• Behaviour Mismatch between ggml_opt in Native Program and WASM

  #807 opened a month ago by saraalrawi
  0

• Add Qualcomm mobile SoC native backend for GGML

  #771 opened 2 months ago by zhouwg
  1

• License for Python-based GGUF parser with NumPy vectorization

  #802 opened a month ago by 99991
  2

• Behavior mismatch between PyTorch `GroupNorm` and `ggml_group_norm`

  #803 opened a month ago by balisujohn
  5

• Is there interest in implementations of something analogous to `torch.Tensor.scatter_` and `torch.gather`?

  #718 opened a month ago by balisujohn
  1

• Is there interest in a groupnorm operation being added?

  #800 opened a month ago by balisujohn
  1

• Is there interest in a cuda implementation for ggml_conv_1d

  #788 opened a month ago by balisujohn
  3

• The library crashes with illegal instructions when the F16C extension isn't supported, please use google's cpu_features library and fail gracefully in such case

  #799 opened a month ago by yurivict
  0

• GGUF quantization meta-data format

  #797 opened 2 months ago by mobicham
  2

• any support for MUL_MAT operation in unsupported GPU feature? (macos intel/amd gpu)

  #796 opened 2 months ago by posojeg
  2

• is there any plan to complete the implementation of use supports_op to check if the backend supports the op?

  #795 opened 2 months ago by zhouwg
  0

• Add dsp backend in ggml.h

  #794 opened 2 months ago by zhouwg
  2

• Magic number in example

  #791 opened 2 months ago by wilderfield
  7

• doc: lack of official logo of GGML

  #776 opened 2 months ago by zhouwg
  4

• Setting tensor's backend

  #779 opened 2 months ago by rgerganov
  2

• Standardized prompting metadata

  #774 opened 2 months ago by MoonRide303
  3

• `test-timestep_embedding` _sometimes_ fails with `ptrace: Operation not permitted.` [opencl-clover, gfx1103] (with `gdb` backtrace.)

  #772 opened 2 months ago by dreirund
  0

• Quant and Dequant operators

  #769 opened 2 months ago by sankalpdayal
  0

• Why perform such operations on k and v as shown in the above diagram?

  #766 opened 3 months ago by EveningLin
  1

• How to deploy my own model using ggml framework

  #762 opened 3 months ago by Francis235
  3

• quantization func test failed with GGML_QKK_64

  #739 opened 3 months ago by winice-test
  2

• Automatically convert pytorch model to ggml

  #756 opened 3 months ago by Maknee
  4

• Incorrect Error Handling in ggml_backend_graph_compute Function in example/magika/main.cpp

  #760 opened 3 months ago by charloco
  1

• Replit code completion example not working

  #758 opened 3 months ago by hassan404
  1

• Error: Unsupported op TIMESTEP_EMBEDDING

  #751 opened 3 months ago by paulocoutinhox
  0

• In gpt-2/convert-h5-to-ggml.py : size mismatch for wpe.weight ... torch.Size([50255, 1024]) ...

  #745 opened 3 months ago by Twenkid
  0

• ggml : add Magika inference

  #734 opened 3 months ago by ggerganov
  0

• ggml : make ggml_fp16_t private

  #720 opened 3 months ago by ggerganov
  2

• mnist convert from tensorflow to ggml

  #744 opened 3 months ago by datduonguva
  0

• ggml : simplify the ggml_compute_forward_ calls

  #724 opened 3 months ago by ggerganov
  2

• Question about documentation with encoder-decoder models

  #740 opened 3 months ago by NatanFreeman
  0

• Does ggml_cont really work?

  #735 opened 4 months ago by BayRanger
  0

• How do pixel unshuffle in ggml ?

  #732 opened 4 months ago by delldu
  2

• Division by zero error

  #723 opened 4 months ago by cpopescu
  2

• ggml : add optional CPU backend context, support reusing threads, async compute

  #721 opened 4 months ago by slaren
  2

• Error when trying to run GPT inference

  #717 opened 4 months ago by thekevinscott
  0

• "array size is too large" on model load

  #714 opened 4 months ago by iamlemec
  1

• some notes about how ggml works using the GPT-2 example

  #711 opened 4 months ago by chunhualiao
  2

• Is a ggml tensor's dimension order reversed?

  #710 opened 4 months ago by chunhualiao
  1