pytorch/executorch

Issues

• [Segmentation fault] python3 torchchat.py export stories15M --dtype fp32 --quantize '{"embedding": {"bitwidth": 4, "groupsize":32}, "linear:a8w4dq": {"groupsize" : 256}}' --output-pte-path stories15M.pte

  #3588 opened a month ago by mikekgfb
  0

• Evaluation results of llama2 with exetorch

  #3568 opened a month ago by l2002924700
  2

• Does llama2 example on Android utilize HTP?

  #3586 opened a month ago by CHNtentes
  0

• gpu

  #3584 opened a month ago by chenbd
  1

• How can I use ExecuTorch to deploy a model to a MicroController,such as Infineon TC3xxx ?

  #3585 opened a month ago by AlexLuya
  0

• Is Qwen in the roadmap?

  #3583 opened a month ago by DzAvril
  0

• ExecuTorch Build Problem

  #3561 opened a month ago by emreaniloguz
  7

• XNNPack Fails For `nn.MaxPool2d`

  #3567 opened a month ago by kinghchan
  4

• Operator torch._ops.aten.linalg_vector_norm.default is not Aten Canonical

  #3566 opened a month ago by nbansal90
  2

• [method.cpp:825] Error setting input 0: 0x10

  #3572 opened a month ago by mikekgfb
  1

• Exporting Llama3's tokenizer

  #3555 opened a month ago by vifi2021
  6

• what's the meaning of "Groupwise 4-bit (128)"

  #3559 opened a month ago by l2002924700
  4

• package download facing 'type' object is not subscriptable

  #3570 opened a month ago by Ciao-Wen-Chen
  0

• Can I run ExecuTorch on ARM Cortex-A53 processor?

  #3541 opened a month ago by neverparadise
  3

• kv cache manipulation?

  #3518 opened a month ago by l3utterfly
  10

• Memory Planning Issue/Optimization (for "concat" style operations)

  #3210 opened 2 months ago by rvijayc
  4

• memory issue during export_llama?

  #3480 opened a month ago by antmikinka
  5

• WebAssembly / Web runtime (both for wasm-simd and WebGPU)

  #3497 opened a month ago by vadimkantorov
  5

• Issue running XNN-pack unittests

  #3311 opened a month ago by freddan80
  2

• converting llama3 models with added tokens

  #3519 opened a month ago by l3utterfly
  3

• Can it run in python virtual environment?

  #3200 opened 2 months ago by tayloryoung-o
  2

• load_method with only method name

  #3198 opened 2 months ago by victoriapoghosian
  3

• ERROR: Overriding output data pointer allocated by memory plan is not allowed.

  #3528 opened a month ago by sunqijie0350
  5

• "Error creating cell resolver" buck2 failure while building wheel

  #3524 opened a month ago by nbansal90
  2

• to edge IR from transformers library model

  #3540 opened a month ago by mhs4670go
  2

• Quantize Llava encoder

  #3557 opened a month ago by iseeyuan
  0

• Support Phi 3 model

  #3550 opened a month ago by iseeyuan
  0

• How can I convert llama3 safetensors to the pth file needed to use with executorch?

  #3303 opened a month ago by l3utterfly
  16

• Executorch exported model produces gibberish: stories15M --dtype fp32 --quantize '{"embedding": {"bitwidth": 4, "groupsize":32}, "linear:a8w4dq": {"groupsize" : 256}}'

  #3542 opened a month ago by mikekgfb
  0

• Buck 2 Error on running ./install_requirements.sh

  #3502 opened a month ago by gochaudhari
  12

• Why is `torch.min` not ATen canonical?

  #3517 opened a month ago by kinghchan
  2

• Executorch reports a bug for pages and pages: [method.cpp:939] Overriding output data pointer allocated by memory plan is not allowed.

  #3515 opened a month ago by mikekgfb
  1

• `torch.max(input)` fails at XNNPACK runtime

  #3516 opened a month ago by kinghchan
  0

• Error when running inference for nanoGPT LLM example

  #3465 opened a month ago by bryangarza
  9

• [v0.2.1] Release Tracker

  #3409 opened a month ago by dbort
  6

• checkpoint str has no attribute 'get'

  #3444 opened a month ago by antmikinka
  5

• Downstream users have dependences on cmake variables and internals, making cmake a compatibility surface

  #3501 opened a month ago by mikekgfb
  0

• exir "missing out vars"

  #3443 opened a month ago by antmikinka
  5

• Add bf16 kernel support

  #3488 opened a month ago by lucylq
  0

• Duplicate registration of quantiation operators, e.g. quantized_decomposed::embedding_byte.out

  #3370 opened a month ago by robell
  7

• error while Building an ExecuTorch Android Demo App

  #3463 opened a month ago by tggmbi
  1

• UNSTABLE Android / test-llama-app / mobile-job (android)

  #3344 opened a month ago by huydhn
  2

• Tagging ConstantArgument in delegation

  #3278 opened a month ago by mhs4670go
  4

• missing packages & incorrect package versions

  #3430 opened a month ago by antmikinka
  6

• Memory planner errors?

  #3425 opened a month ago by mikekgfb
  3

• UserWarning: Attempted to insert a get_attr Node .. when `to_backend` is called

  #3276 opened a month ago by mhs4670go
  4

• UNSTABLE trunk / test-coreml-delegate / macos-job

  #3264 opened a month ago by huydhn
  2

• Remove dump of model IR

  #3280 opened 2 months ago by mikekgfb
  1

• build of executorch triggers a warning related to too many arguments provided for format string

  #3189 opened 2 months ago by mikekgfb
  0

• [QNN] Support non-Android aarch64 Linux targets for QNN backend?

  #3132 opened 2 months ago by tc-wolf
  0