Fooling around with Copilot
ML Engineer @ Hugging Face
- Bern, Switzerland
- https://lewtun.github.io/blog/
- @_lewtun
Highlights
- Pro
Block or Report
Block or report lewtun
Report abuse
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abusePinned
-
Jupyter notebooks for the Natural Language Processing with Transformers book
-
-
-
-
giotto-tda Public
Forked from giotto-ai/giotto-tda
A high performance topological machine learning toolbox in Python
Python
2,537 contributions in the last year
Less
More
Activity overview
Contributed to
huggingface/course,
huggingface/setfit,
huggingface/model-evaluator
and 41 other
repositories
Contribution activity
November 2022
Created 35 commits in 9 repositories
Created 1 repository
- lewtun/course Python
Created a pull request in huggingface/hub-docs that received 16 comments
Add note about embedding Spaces with iframe tags
Adds a note about how to embed Streamlit / Gradio Spaces in other webpages with <iframe> tags.
Related Slack thread: https://huggingface.slack.com/…
+16
−0
•
16
comments
Opened 24 other pull requests in 7 repositories
huggingface/course
1
closed
5
merged
1
open
huggingface/diffusion-models-class
5
merged
huggingface/setfit
1
open
4
merged
huggingface/blog
4
merged
huggingface/huggingface_hub
1
merged
huggingface/transformers
1
merged
huggingface/hf_benchmarks
1
merged
Reviewed 55 pull requests in 7 repositories
huggingface/setfit
23 pull requests
-
add option to set
samples_per_label -
Throw clear
ValueErrorwhen neglecting to passtrain_datasettoDistillationSetFitTrainer - Various cleanups; type hint fixes incl. corresponding to PEP 484
- Fix/input type hint
- delete duplicated code
- add option to normalize embeddings
- Allow setting max length
- Bugfix/body and head on different devices
- Use scikit-learn rather than sklearn in requirements files
- Update evaluate
- Distill trainer
- Added initial onnx export function
- Fixed typo in model head from predict_prob to predict_proba
- add more loss function options
- Fix typos: image(s) -> sentence(s)
-
fix non default
loss_classissue to use triplet loss - Add the usage and relevant info. of the differentiable head to README
-
Fix typo:
temerature=>temperature - Small fix on hyperparameter search
-
Add support to kwargs in
compute()method called bytrainer.evaluate() - fix: templated examples copy empty vector
- redirect __call__ to predict
- Support for the differentiable head
huggingface/course
21 pull requests
- Add zh-CN machine translations.
- Add accurate English subtitles.
- [i18n-EN] same links across languages: step 1
- Migrate Spaces URLs to new domain
- [i18n-KO] Translated chapter 2-1 page to Korean
- Custom English anchor for non-alphabet headings in chapter0
- Chapter 9 - Italian
- Typo
- Typo in Chapter 2, Section 2
- Adding Portuguese Translation to Chapter3
- Add Japanese trasnlation of chapter 1/ 7 to 10
- Spanish translation of Chapter 5
- [RU] Chapter 6 (1/2) finished
- [FR] Add FAQ and more
- Add Python syntax highlighting
- Update pr docs actions
- Add Japanese trasnlation of chapter 1 / 2 to 3
- [typo] Update 1.mdx
- [typo] Update 1.mdx
- [de] Translation Chapter 1
- [Audio] Tour of applications
huggingface/huggingface_hub
3 pull requests
huggingface/transformers
3 pull requests
huggingface/optimum
3 pull requests
huggingface/hub-docs
1 pull request
huggingface/hf_benchmarks
1 pull request
Created an issue in huggingface/setfit that received 5 comments
Refactor training hyperparameters in SetFitTrainer to dedicated class
The current implementation of SetFitTrainer has become a bit bloated because it accepts all the hyperparameters for training as separate arguments.…
2 tasks
Opened 7 other issues in 4 repositories
huggingface/huggingface_hub
2
closed
1
open
huggingface/setfit
1
open
1
closed
UKPLab/sentence-transformers
1
open
huggingface/datasets
1
closed
6
contributions
in private repositories
Nov 1 – Nov 29





