huggingfaceR
huggingfaceR copied to clipboard
farach
Hugging Face state-of-the-art models in R
hf_load_pipeline() from zzz.R is changed to hf_import_pipeline() to better reflect what it does. hf_load_model() changed to hf_load_pipeline() to better reflect what it does. hf_load_model() introduced - loads a model. hf_load_autotokenizer...
Hi all! I'm Amit from R4ds :) Anyway, was giving this repo a go, and trying to run the example on the readme, and hitting this bug: ``` > emotion
Necessary ahead of release I feel. README could perhaps do with a rejig too + potentially a package intro vignette separate to introduction to tokenizers, models, pipelines etc.
introduction_to_piplines --> introduction_to_pipelines Will need to fix the merging issue, should be straightforward as it's just a change of name.
``` huggingfaceR:::hf_search_models("facebook") ``` results in `Cannot convert object to an environment: [type=list; target=ENVSXP].` while ``` reticulate::py$hf_api$list_models(filter = reticulate::py$ModelFilter(author = "facebook")) ``` behaves as expected. It looks like reticulate::iterate is being...
There are some slick ways to pull only subsets of splits you can read up on [here](https://huggingface.co/docs/datasets/loading). Current implementation doesn't allow for these. They used to work by being passed...
Some datasets don't have labels defined but are still valid datasets (one I'm using for the ez functions vignette). `hf_load_dataset("sentiment140", split = "test")` results in `KeyError: 'label'`.
Function + arguments break too easily for datasets other than emotions e.g. `x
Hello. Interesting stuff! Thanks for it. Is it possible to install _huggingfaceR_ via devtools::install_github without also installing a miniconda environment? I have existing Python virtual environments that I use via...
