Commits

Branch selector

 main

User selector

RyanJDick

All time

Commit history

Commits on Oct 10, 2023

• Fix bug in skip_torch_weight_init() where the original behavior of torch.nn.Conv*d modules wasn't being restored correctly.

  RyanJDick

  committed
  61242bf

  Copy full SHA for 61242bf

Commits on Oct 9, 2023

• Add a skip_torch_weight_init() context manager to improve model load times (from disk).

  RyanJDick

  committed
  58b56e9

  Copy full SHA for 58b56e9

Commits on Oct 7, 2023

• Bump DenoiseLatentsInvocation version.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  9c720da

  Copy full SHA for 9c720da

• yarn build

  

  RyanJDick

  authored and

  hipsterusername

  committed
  e1b576c

  Copy full SHA for e1b576c

• Refactor multi-IP-Adapter to clean up the interface around changing scales.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  971ccfb

  Copy full SHA for 971ccfb

• Fix typo in setting IP-Adapter scales.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  43a3c3c

  Copy full SHA for 43a3c3c

• Tidy _prepare_attention_processors(...) logic.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  4df1cdb

  Copy full SHA for 4df1cdb

• Fixup IP-Adapter locale strings.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  3f860c3

  Copy full SHA for 3f860c3

• Fix handling of scales with multiple IP-Adapters.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  d8d0c9a

  Copy full SHA for d8d0c9a

• Bugfix for multi-ip-adapter in DenoiseLatentsInvocation.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  9403672

  Copy full SHA for 9403672

• Frontend changes to enable multiple IP-Adapters in the workflow editor.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  9459184

  Copy full SHA for 9459184

• Fixes to get IP-Adapter tests working with new multi-IP-Adapter support.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  26b91a5

  Copy full SHA for 26b91a5

• Update IP-Adapter model to enable running multiple IP-Adapters at once. (Not tested yet.)

  

  RyanJDick

  authored and

  hipsterusername

  committed
  7ca456d

  Copy full SHA for 7ca456d

• WIP - Accept a list of IPAdapterFields in DenoiseLatents.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  78828b6

  Copy full SHA for 78828b6

Commits on Oct 6, 2023

• Proposal: Support slow tests that depend on models (#4813)

  Show description for 166ff9d
  RyanJDick

  authored
  166ff9d

  Copy full SHA for 166ff9d

• Merge branch 'main' into ryan/model-tests

  RyanJDick

  authored
  4f97bd4

  Copy full SHA for 4f97bd4

• Remove @slow decorator in favor of @pytest.mark.slow.

  RyanJDick

  committed
  e0e0017

  Copy full SHA for e0e0017

• Improve model cache debug logging (#4784)

  Show description for c188713
  RyanJDick

  authored
  c188713

  Copy full SHA for c188713

• Merge branch 'main' into ryan/model-cache-logging-only

  RyanJDick

  authored
  096d195

  Copy full SHA for 096d195

Commits on Oct 5, 2023

• Add TESTS.md documentation.

  RyanJDick

  committed
  7870b90

  Copy full SHA for 7870b90

• Fix Flake8 errors by using a pytest conftest.py file.

  RyanJDick

  committed
  9854b24

  Copy full SHA for 9854b24

• Fix broken link in documentation to 'Frontend Documentation'.

  RyanJDick

  committed
  7d800e1

  Copy full SHA for 7d800e1

• POC of a test that depends on models.

  RyanJDick

  committed
  1c8b1fb

  Copy full SHA for 1c8b1fb

• Set MALLOC_MMAP_THRESHOLD_=1048576 by default in invoke.sh. And add it to the manual installation docs.

  RyanJDick

  committed
  594a3ae

  Copy full SHA for 594a3ae

• Add support for T2I-Adapter in node workflows (#4612)

  Show description for 7837746
  

  3 peopleauthored
  7837746

  Copy full SHA for 7837746

• Add support for IPAdapterPlusXL based on https://github.com/tencent-ailab/IP-Adapter/commit/6219530507cb4696a0496f10c0e5a4f1dbdc7672.

  

  RyanJDick

  authored and

  hipsterusername

  committed
  fbe6452

  Copy full SHA for fbe6452

Commits on Oct 3, 2023

• Re-enable garbage collection in model cache MemorySnapshots.

  RyanJDick

  committed
  2479a59

  Copy full SHA for 2479a59

• (minor) clean up typos.

  RyanJDick

  committed
  7d0ac2c

  Copy full SHA for 7d0ac2c

• Add unit test for Struct_mallinfo2.__str__()

  RyanJDick

  committed
  519b892

  Copy full SHA for 519b892

• Add unit test for get_pretty_snapshot_diff(...).

  RyanJDick

  committed
  763dcac

  Copy full SHA for 763dcac

• Add unit test for LibcUtil().mallinfo2().

  RyanJDick

  committed
  3599d54

  Copy full SHA for 3599d54

• Disable garbage collection in ModelCache calls to MemorySnapshot in order minimize snapshot overhead.

  RyanJDick

  committed
  22a8493

  Copy full SHA for 22a8493

• Add README with info about glib memory fragmentation caused by the model cache.

  RyanJDick

  committed
  d64e17e

  Copy full SHA for d64e17e

• Catch a more specific exception in environments that do not have a libc shared library.

  RyanJDick

  committed
  ba54277

  Copy full SHA for ba54277

• Minor fixes.

  RyanJDick

  committed
  5915a4a

  Copy full SHA for 5915a4a