Commits

Branch selector

 master

User selector

james77777778

Datepicker

All time

Commit History

Commits on Nov 7, 2024

• Add support for direct tensor as initializer (#20457)

  Show description for ccb07df
  james77777778

  authored
  ccb07df

  Copy full SHA for ccb07df

Commits on Nov 6, 2024

• Suppress warnings for mismatched tuples and lists in functional models. (#20456)

  james77777778

  authored
  4e71b2e

  Copy full SHA for 4e71b2e

Commits on Nov 3, 2024

• Fix encoding issue (#20443)

  Show description for 0adc924
  james77777778

  authored
  0adc924

  Copy full SHA for 0adc924

Commits on Oct 26, 2024

• Add `mean_with_sample_weight` reduction to `Loss` (#20410)

  Show description for 093b055
  james77777778

  authored
  093b055

  Copy full SHA for 093b055

Commits on Oct 22, 2024

• Use `IndexError` instead of `ValueError` in `deserialize_node` (#20389)

  james77777778

  authored
  1cc1eb5

  Copy full SHA for 1cc1eb5

Commits on Oct 18, 2024

• Refactor `pythonify_logs` and add custom fit integration tests. (#20380)

  james77777778

  authored
  d6dfba5

  Copy full SHA for d6dfba5

Commits on Sep 27, 2024

• Fix `_filter_unused_inputs` (#20296)

  james77777778

  authored
  93cb954

  Copy full SHA for 93cb954

Commits on Sep 25, 2024

• Add ops.nn.dot_product_attention (#20286)

  james77777778

  authored
  3bc3aef

  Copy full SHA for 3bc3aef

Commits on Sep 13, 2024

• Fix numeric instability in LayerNormalization and BatchNormalization. (#20257)

  james77777778

  authored
  85e3681

  Copy full SHA for 85e3681

Commits on Sep 8, 2024

• Add `ops.trunc` (#20232)

  Show description for e7731ce
  james77777778

  authored
  e7731ce

  Copy full SHA for e7731ce

Commits on Sep 4, 2024

• Replace `ValueError` with `AttributeError` in `Sequential`'s property (#20206)

  Show description for dbad6b0
  james77777778

  authored
  dbad6b0

  Copy full SHA for dbad6b0

• Fix and improve the mechansim for removing metrics in Trainer (#20205)

  james77777778

  authored
  7267cd5

  Copy full SHA for 7267cd5

Commits on Sep 2, 2024

• Fix nested trainer metrics (#20197)

  james77777778

  authored
  18ee0a6

  Copy full SHA for 18ee0a6

Commits on Aug 28, 2024

• Mark layers with no state as built. (#20175)

  Show description for 19ba3d1
  james77777778

  authored
  19ba3d1

  Copy full SHA for 19ba3d1

Commits on Aug 27, 2024

• Add warning for mismatched inputs structure for `Functional` (#20170)

  james77777778

  authored
  cb233fa

  Copy full SHA for cb233fa

Commits on Aug 22, 2024

• Fix typing.cast when subclassing (#20149)

  james77777778

  authored
  f4a4725

  Copy full SHA for f4a4725

Commits on Aug 17, 2024

• Fix CI (#20131)

  james77777778

  authored
  6499f91

  Copy full SHA for 6499f91

Commits on Aug 16, 2024

• Fix JAX's bitwise_*_shift and bitwise_invert (#20129)

  james77777778

  authored
  7150675

  Copy full SHA for 7150675

Commits on Aug 15, 2024

• Add bitwise ops (#20126)

  james77777778

  authored
  d72a0ea

  Copy full SHA for d72a0ea

Commits on Aug 14, 2024

• Fix svd (#20122)

  james77777778

  authored
  6fd4873

  Copy full SHA for 6fd4873

Commits on Aug 9, 2024

• Fix summary(show_trainable=True) (#20102)

  james77777778

  authored
  933579d

  Copy full SHA for 933579d

Commits on Aug 2, 2024

• Add numpy `scale` and `offset` support to `Rescaling` (#20079)

  Show description for 16dd8b8
  james77777778

  authored
  16dd8b8

  Copy full SHA for 16dd8b8

Commits on Aug 1, 2024

• Fix `input_shape` in Sequential model (#20076)

  james77777778

  authored
  d8b3314

  Copy full SHA for d8b3314

Commits on Jul 30, 2024

• Fix dtype for `abs_max_quantize` (#20065)

  james77777778

  authored
  c1a00e6

  Copy full SHA for c1a00e6

Commits on Jul 24, 2024

• Bring back loss information for multiple outputs (#20023)

  Show description for f6a81cc
  james77777778

  authored
  f6a81cc

  Copy full SHA for f6a81cc

Commits on Jul 23, 2024

• Simplify the implementation of quantization-related methods (#19954)

  Show description for 3ac43b1
  james77777778

  authored
  3ac43b1

  Copy full SHA for 3ac43b1

• Fix jax's `in_top_k` and numpy's `top_k` (#20033)

  Show description for 9ebc65a
  james77777778

  authored
  9ebc65a

  Copy full SHA for 9ebc65a

• Fix QLoRA saving/loading (#20029)

  james77777778

  authored
  7c6d501

  Copy full SHA for 7c6d501

Commits on Jul 19, 2024

• Simplify and improve `CompileLoss` (#20015)

  james77777778

  authored
  8f5ad1e

  Copy full SHA for 8f5ad1e

• Revert to the original default behavior of `dtype` (#20014)

  james77777778

  authored
  8ca7cd8

  Copy full SHA for 8ca7cd8

• Avoid exceptions in `save_model` and `load_model` (#19989)

  Show description for dfc4f78
  james77777778

  authored
  dfc4f78

  Copy full SHA for dfc4f78

Commits on Jul 17, 2024

• Extend support to`Sequential._maybe_build` for functional/sequential model as first layer and fix `.summary()` (#20002)

  james77777778

  authored
  902f9da

  Copy full SHA for 902f9da

Commits on Jul 12, 2024

• Fix dtype inconsistency in TF (#19978)

  james77777778

  authored
  de066e5

  Copy full SHA for de066e5

Commits on Jul 11, 2024

• Fix GPU CI (#19975)

  james77777778

  authored
  ed1b3fe

  Copy full SHA for ed1b3fe

Commits on Jul 5, 2024

• Remove upperbound for torch version (#19956)

  james77777778

  authored
  58c3ee5

  Copy full SHA for 58c3ee5

Pagination

Previous

Next