Nikita Savelyev
Nikita Savelyev
- [x] I've checked [docs](https://rich.readthedocs.io/en/latest/introduction.html) and [closed issues](https://github.com/Textualize/rich/issues?q=is%3Aissue+is%3Aclosed) for possible solutions. - [x] I can't find my issue in the [FAQ](https://github.com/Textualize/rich/blob/master/FAQ.md). **Describe the bug** Setting `style` to `TaskProgressColumn` does not...
### Changes When compression is applied to a model saved with FP32 weights, the resulting graph is different compared to the case when an input model is saved with FP16...
### Issue type Bug ### Have you reproduced the bug with TensorFlow Nightly? Yes ### Source binary ### TensorFlow version 2.13.0 ### Custom code No ### OS platform and distribution...
### Changes - Unpatch torch during `torch.compile()` call - Unpatch torch during forward call of compiled model ### Reason for changes PyTorch dynamo compilation conflicts with nncf patching of PyTorch....
### Changes ### Reason for changes ### Related tickets ### Tests
# What does this PR do? Changes: - Tweak some reference values - Update model ids for `test_exporters_cli_int4_with_local_model_and_default_config` ## Before submitting - [ ] This PR fixes a typo or...
**Changes** - For `4BIT_DEFAULT` option always import compression config from optimum-intel - Remove all default configs from openvino.genai, rely only on default configs in optimum-intel - Extend compression parameters parsing...
**Changes** Fix how max token slicing is applied inside data-aware compression `transform_fn`. **Reason for change** `input_ids` and `attention_mask` have the shape of `(B, N)`, but max token slicing is applied...
### Changes Extending #2683 with implementation for OV backend | Model | Backend | Compression | Sparsity Parameters | Wikitext word perplexity | | ------------------------ | ------- | ---------------------------------- |...