Hafedh Hichri
Hafedh Hichri
# What does this PR do? fixes configuration file not pointing at a remote repo with custom pipeline architecture Fixes #28907 ## Before submitting - [ ] This PR fixes...
# What does this PR do? this will add `push_to_hub` method to the pipelines allowing people to push their custom pipelines to the huggingface hub easily Fixes #28857 #28983 ##...
### System Info - `transformers` version: 4.38.2 - Platform: Linux-6.1.58+-x86_64-with-glibc2.35 - Python version: 3.10.12 - Huggingface_hub version: 0.22.2 - Safetensors version: 0.4.3 - Accelerate version: not installed - Accelerate config:...
not an issue but rather an improvement. the trainer API offers a set of easy-to-implement functionalities plus it's compatible with the multi-GPU training by default. I described the process of...
given a repo with a custom pipeline and not a custom model * repo : https://huggingface.co/sgugger/test-dynamic-pipeline * docs : https://huggingface.co/docs/transformers/main/en/add_new_pipeline#add-the-pipeline-to--transformers the `pipelineSnippet` is wrongly annotated ## what is present ...
### Feature request when working with a really large dataset it would save us a lot of time (and compute resources) to use either with_transform or the set_transform from the...
This pr is about allowing for huggingface integration, I made this [notebook](https://colab.research.google.com/drive/1y_QeQCjP7BZP9vOW1GJTOA1rTeW7Oq5_?usp=sharing) using `PytorchModelHubMixin` allowing us to download, load and initialize the model in 1 line ```python from mobileclip import...
this pr will mainly add 3 methods to the uniformer model * `save_pretrained` * `push_to_hub` * `from_pretrained` : initialize and load the model weights allowing your model to be easily...
## this pr will mainly add 3 methods to michelangelo * `save_pretrained` * `push_to_hub` * `from_pretrained`: initialize and load the model weights allowing your models to be easily integrated with...