4 Releases published by 2 people

• v6.0.1

  published Mar 5, 2025

• v6.0.2

  published Mar 8, 2025

• v6.0.3

  published Mar 21, 2025

• v6.1.1

  published Mar 29, 2025

6 Pull requests merged by 3 people

• Adopt new RAI whl naming convention

  #300 merged Mar 29, 2025

• Release lemonade 6.1.0

  #298 merged Mar 28, 2025

• Release v6.0.3

  #295 merged Mar 20, 2025

• Added mmlu accuracy test command format

  #293 merged Mar 12, 2025

• Release v6.0.2

  #294 merged Mar 8, 2025

• Lemonade release v6.0.1: Extended server support, increased reliability

  #292 merged Mar 5, 2025

31 Issues closed by 2 people

• What is the best way to use turnkeyml to sniff the inputs to a pytorch model?

  #296 closed Mar 25, 2025

• TurnkeyML shows an unhelpful error message when docker daemon is not running

  #30 closed Mar 25, 2025

• Proposal: Enable setting all optional Turnkey CLI arguments using a config file

  #29 closed Mar 25, 2025

• Proposal: Add --max-throughput to CLI and expose it to runtimes

  #28 closed Mar 25, 2025

• Analysis of model scripts finds two models

  #22 closed Mar 25, 2025

• Models that require CUDA don't work by default

  #17 closed Mar 25, 2025

• midas_v3_large.py in torch_hub export failed

  #16 closed Mar 25, 2025

• New feature: Model acceptance test

  #9 closed Mar 25, 2025

• Every model in the transformers corpus uses the task::Generative_AI label

  #7 closed Mar 25, 2025

• Leverage the new "dynamo" torch onnx export logic in the discover tool

  #235 closed Mar 25, 2025

• OEM info prompts user for sudo password

  #230 closed Mar 25, 2025

• TKML tool not showing global args in help menu

  #218 closed Mar 25, 2025

• Increase the default ONNX opset to 18

  #164 closed Mar 25, 2025

• Refresh: Remove all benchmarking code

  #156 closed Mar 25, 2025

• `dummy_inputs` should make `onnx_session` an optinal input

  #217 closed Mar 25, 2025

• Refresh: Delete the ONNX model analysis code

  #154 closed Mar 25, 2025

• Handle inconsistent opset placement in ONNX models

  #150 closed Mar 25, 2025

• Proposal: check model path in `--model_path` scripts

  #131 closed Mar 25, 2025

• Input size is only saved to stats in ONNX tool flows

  #129 closed Mar 25, 2025

• Enhance conda creation methodology for onnxrt

  #132 closed Mar 25, 2025

• Validate `evaluate_files()` arguments

  #117 closed Mar 25, 2025

• Create an ORT base class and refactor run/onnxrt on to it

  #110 closed Mar 25, 2025

• Proposal: A way to skip LLMs

  #93 closed Mar 25, 2025

• Support for PyTorch Lightning

  #71 closed Mar 25, 2025

• Add support for pipelines of models

  #88 closed Mar 25, 2025

• Make sure that every public API has a nice docstring

  #63 closed Mar 25, 2025

• Create unit tests for each public API

  #62 closed Mar 25, 2025

• Features Request: Signing Models with sigstore?

  #52 closed Mar 25, 2025

• ONNX model size on disk is not captured for models >2GB in size

  #41 closed Mar 25, 2025

• Help users who forgot to install the models requirements

  #40 closed Mar 25, 2025

• Proposal: Set batch size as a first-class argument in the CLI

  #31 closed Mar 25, 2025

2 Issues opened by 2 people

• Add GIFs to ReadMe to Improve UX

  #299 opened Mar 27, 2025

• Selected Models not working

  #297 opened Mar 22, 2025

2 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Fix the gibberish output from llm-prompt

  #286 commented on Mar 25, 2025 • 1 new comment

• Knowledge Base: How to apply `black` formatting

  #281 commented on Mar 25, 2025 • 0 new comments