metadata
base_model: BAAI/bge-base-en-v1.5
library_name: setfit
metrics:
- accuracy
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
widget:
- text: >-
###Instruction: Multi-class classification, answer with one of the labels:
[delete, keep, redirect, merge, no_consensus] : ###Input: Music groups: —
Jeffq 10:52, 13 May 2005 (UTC) [ reply ] Music groups [ edit ] This was
nominated for VfD but not listed here. All existing articles in this list
are now tagged with Category:Bands or Category:Musicians as appropriate.
It could still be used as a "requested" list for musician quote
articles. — Jeff Q (talk) 22:26, 19 Apr 2005 (UTC) VOTE CLOSED. Result:
[MASK] (3 Deletes; 1 [MASK]/Rename). — Jeff Q (talk) 10:52, 13 May 2005
(UTC) [ reply ] [MASK] . Jeff Q (talk) 22:26, 19 Apr 2005 (UTC) [MASK] .
We don't have a tradition of red link lists on this project but only lists
of actual articles. Rmhermen 02:42, 20 Apr 2005 (UTC) [MASK] and rename
to List of music groups or List of bands . Lists and categories are
complementary, not mutually exclusive. jni 10:52, 21 Apr 2005 (UTC) That
is only true when a list provides information or formatting that a
category cannot. Other than providing a "requested bands" list, are there
some additional benefits to this list? — Jeff Q (talk) 02:10, 22 Apr 2005
(UTC) [MASK] . We are enough with categories. -- Aphaia 00:04, 5 May
2005 (UTC) [ reply ] The above discussion is preserved as an archive of
the debate. Please do not modify it. Subsequent comments should be made
on the appropriate discussion page (such as the article's talk page or in
a deletion review ). No further edits should be made to this page.
###Output:
- text: >-
###Instruction: Multi-class classification, answer with one of the labels:
[delete, keep, redirect, merge, no_consensus] : ###Input: Kent Hovind: —
Jeff Q (talk) 15:50, 5 August 2006 (UTC) [ reply ] Kent Hovind [ edit ]
This is a Wikipedia article, and Wikipedia already has one . This one
appears to be an essay on Hovind's views, which is not the purpose of
Wikiquote. I don't know if it is material considered undesirable or too
detailed for the WP article, but that's irrelevant. We need quotes and
only quotes. ~ Jeff Q (talk) 20:07, 21 July 2006 (UTC) [ reply ] Vote
closed. Result: [MASK] (4 keeps; no dissent; article signficantly
improved). ~ Jeff Q (talk) 15:50, 5 August 2006 (UTC) [ reply ] [MASK]
unless all text (except a 1-paragraph intro) replaced with actual quotes
(preferably sourced). I'd recommend transwiki, except that the sole
editor is already actively editing the WP article and can add this
material to it if they wish. ~ Jeff Q (talk) 20:07, 21 July 2006 (UTC) [
reply ] [MASK] now that C56C has done considerable work to convert this to
a proper quote article. Some issues remain, including a few
not-really-quote items and a need for better sourcing, but I think it's
mostly cleanup at this point. ~ Jeff Q (talk) 00:49, 22 July 2006 (UTC) [
reply ] [MASK] per Jeffq. — LrdChaos 20:15, 21 July 2006 (UTC) [ reply ]
[MASK] now that it's actually a quote article. — LrdChaos 12:24, 22 July
2006 (UTC) [ reply ] [MASK] I cleaned it up and added quotes and their
sources. C56C 23:22, 21 July 2006 (UTC) [ reply ] [MASK]. - InvisibleSun
01:34, 22 July 2006 (UTC) [ reply ] The above discussion is preserved as
an archive of the debate. Please do not modify it. Subsequent comments
should be made on the appropriate discussion page (such as the article's
talk page or in a deletion review ). No further edits should be made to
this page. ###Output:
- text: >-
###Instruction: Multi-class classification, answer with one of the labels:
[delete, keep, redirect, merge, no_consensus] : ###Input: But Chi Huen: —
Jeffq 22:50, 8 May 2006 (UTC) [ reply ] But Chi Huen [ edit ] Not English
Might be Vanity Rather nonsensical Sydneyfong 15:11, 23 April 2006 (UTC) [
reply ] Vote closed. Result: [MASK] (2 Deletes, 1 [MASK]/Transwiki w/
emphasis on former, 1 implicit [MASK]; no dissent; no evidence provided
[at least in English]). ~ Jeff Q (talk) 22:50, 8 May 2006 (UTC) [ reply ]
[MASK] . This looks like a typical article about a professor from an
admiring college student. Google suggests this person is real, but is
unlikely to rise to a wiki notability level. (I've been wrong before,
however, so evidence is requested.) The WP link in the article points to
a non-existent w:But Sir , and this name is not explained in the largely
irrelevant WQ intro text. (The author seems to have been trying to do
both a WP stub and a WQ article here, and talks more about the class than
the quotee.) The intro itself is extremely POV and is unsourced.
Finally, these quotes are likely all unverifiable, and seem to be the
usual stuff picked up by students in class. (I admit some are
entertaining; they remind me of a computer professor I had who would
always say "that take cares [sic] of that".) I'm sure the instructor is
interesting and honorable, but that isn't sufficient for a WQ article. ~
Jeff Q (talk) 19:33, 23 April 2006 (UTC) [ reply ] [MASK] , not notable.
~ UDScott 11:45, 24 April 2006 (UTC) [ reply ] [MASK] or Transwiki to ZH
Wikiquote . I prefer to [MASK] it, since all quotes including Chinese
ones seem not so significant, but rather "favorite criches of Prof But".
Even this professor is wiki-notable, the current content isn't in my
humble opinion. -- Aphaia 10:44, 1 May 2006 (UTC) [ reply ] The above
discussion is preserved as an archive of the debate. Please do not modify
it. Subsequent comments should be made on the appropriate discussion page
(such as the article's talk page or in a deletion review ). No further
edits should be made to this page. ###Output:
- text: >-
###Instruction: Multi-class classification, answer with one of the labels:
[delete, keep, redirect, merge, no_consensus] : ###Input: Faith (Buffy
the Vampire Slayer): -- Aphaia 03:08, 23 Jun 2005 (UTC) Faith (Buffy the
Vampire Slayer) [ edit ] We do not have character pages for any other
characters on Buffy, most quotes would be dialogues anyway. I've already
added a few "five-by-five" themed quotes to the Buffy the Vampire Slayer
page. MosheZadka 06:36, 8 Jun 2005 (UTC) Vote closed: Deleted. (3
deletes, no dissent; for "expand" vote, see below). -- Aphaia 03:08, 23
Jun 2005 (UTC) [MASK] MosheZadka 06:36, 8 Jun 2005 (UTC) [MASK] . The
Buffy page is very thorough, so Faith quotes have a good forum already. --
RPickman 20:50, 12 Jun 2005 (UTC) [MASK] , but with some reservation.
This is a recurring issue and will only become more visible as Wikiquote
grows. Just as individual's quotes are duplicated in theme pages, it can
be useful to have some character quotes in their own pages as well as show
pages, especially for show articles as large and as heavily formatted as
Buffy … but only if the character has a large number of pithy quotes
listed . That is not currently the case for Faith, but it is for Darth
Vader and other Star Wars characters. — Jeff Q (talk) 05:27, 15 Jun 2005
(UTC) Expand I suggest someone make more quotes from her charicter. --
Admiral Roo 18:38, 22 Jun 2005 (UTC) I don't count this vote, because it
was voted after the deadline. -- Aphaia 03:08, 23 Jun 2005 (UTC) The above
discussion is preserved as an archive of the debate. Please do not modify
it. Subsequent comments should be made on the appropriate discussion page
(such as the article's talk page or in a deletion review ). No further
edits should be made to this page. ###Output:
- text: >-
###Instruction: Multi-class classification, answer with one of the labels:
[delete, keep, redirect, merge, no_consensus] : ###Input: Christopher
Chippindale: Aphaia 23:30, 29 June 2005 (UTC) [ reply ] Christopher
Chippindale [ edit ] Wikipedia is a link to a non-existing article and no
quotes. MosheZadka 06:38, 12 Jun 2005 (UTC) Vote closed : Deleted (3
deletes, no dissent) —The preceding unsigned comment was added by Aphaia (
talk • contribs ) 23:30, 29 June 2005 (UTC) [MASK] MosheZadka 06:38, 12
Jun 2005 (UTC) Comment: I've added a link to provide confirmation of this
person's existence, for what it's worth. — Jeff Q (talk) 11:22, 12 Jun
2005 (UTC) Comment: If someone were to find a half-way verifiable quote
and put it there, I would change my vote. But as is, I couldn't find
anything via a google search or otherwise. MosheZadka 13:29, 12 Jun 2005
(UTC) [MASK] . -- RPickman 19:33, 19 Jun 2005 (UTC) [MASK] . No
expectation of any quotes. — Jeff Q (talk) 23:38, 24 Jun 2005 (UTC) The
above discussion is preserved as an archive of the debate. Please do not
modify it. Subsequent comments should be made on the appropriate
discussion page (such as the article's talk page or in a deletion review
). No further edits should be made to this page. ###Output:
inference: true
SetFit with BAAI/bge-base-en-v1.5
This is a SetFit model that can be used for Text Classification. This SetFit model uses BAAI/bge-base-en-v1.5 as the Sentence Transformer embedding model. A LogisticRegression instance is used for classification.
The model has been trained using an efficient few-shot learning technique that involves:
- Fine-tuning a Sentence Transformer with contrastive learning.
- Training a classification head with features from the fine-tuned Sentence Transformer.
Model Details
Model Description
- Model Type: SetFit
- Sentence Transformer body: BAAI/bge-base-en-v1.5
- Classification head: a LogisticRegression instance
- Maximum Sequence Length: 512 tokens
- Number of Classes: 5 classes
Model Sources
- Repository: SetFit on GitHub
- Paper: Efficient Few-Shot Learning Without Prompts
- Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts
Model Labels
Label | Examples |
---|---|
0 |
|
1 |
|
4 |
|
2 |
|
3 |
|
Uses
Direct Use for Inference
First install the SetFit library:
pip install setfit
Then you can load this model and run inference.
from setfit import SetFitModel
# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("research-dump/bge-base-en-v1.5_wikiquote_masked_outcome_prediction_masked")
# Run inference
preds = model("###Instruction: Multi-class classification, answer with one of the labels: [delete, keep, redirect, merge, no_consensus] : ###Input: Christopher Chippindale: Aphaia 23:30, 29 June 2005 (UTC) [ reply ] Christopher Chippindale [ edit ] Wikipedia is a link to a non-existing article and no quotes. MosheZadka 06:38, 12 Jun 2005 (UTC) Vote closed : Deleted (3 deletes, no dissent) —The preceding unsigned comment was added by Aphaia ( talk • contribs ) 23:30, 29 June 2005 (UTC) [MASK] MosheZadka 06:38, 12 Jun 2005 (UTC) Comment: I've added a link to provide confirmation of this person's existence, for what it's worth. — Jeff Q (talk) 11:22, 12 Jun 2005 (UTC) Comment: If someone were to find a half-way verifiable quote and put it there, I would change my vote. But as is, I couldn't find anything via a google search or otherwise. MosheZadka 13:29, 12 Jun 2005 (UTC) [MASK] . -- RPickman 19:33, 19 Jun 2005 (UTC) [MASK] . No expectation of any quotes. — Jeff Q (talk) 23:38, 24 Jun 2005 (UTC) The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page. ###Output: ")
Training Details
Training Set Metrics
Training set | Min | Median | Max |
---|---|---|---|
Word count | 101 | 313.5930 | 3998 |
Label | Training Sample Count |
---|---|
0 | 402 |
1 | 46 |
2 | 5 |
3 | 12 |
4 | 19 |
Training Hyperparameters
- batch_size: (8, 2)
- num_epochs: (2, 2)
- max_steps: -1
- sampling_strategy: oversampling
- num_iterations: 10
- body_learning_rate: (1e-05, 1e-05)
- head_learning_rate: 5e-05
- loss: CosineSimilarityLoss
- distance_metric: cosine_distance
- margin: 0.25
- end_to_end: True
- use_amp: True
- warmup_proportion: 0.1
- l2_weight: 0.01
- seed: 42
- eval_max_steps: -1
- load_best_model_at_end: False
Training Results
Epoch | Step | Training Loss | Validation Loss |
---|---|---|---|
0.0008 | 1 | 0.2005 | - |
0.4132 | 500 | 0.118 | 0.1672 |
0.8264 | 1000 | 0.0064 | 0.2326 |
1.2397 | 1500 | 0.0052 | 0.1932 |
1.6529 | 2000 | 0.0037 | 0.2023 |
Framework Versions
- Python: 3.10.12
- SetFit: 1.1.0
- Sentence Transformers: 3.3.1
- Transformers: 4.44.1
- PyTorch: 2.2.1+cu121
- Datasets: 2.21.0
- Tokenizers: 0.19.1
Citation
BibTeX
@article{https://doi.org/10.48550/arxiv.2209.11055,
doi = {10.48550/ARXIV.2209.11055},
url = {https://arxiv.org/abs/2209.11055},
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Efficient Few-Shot Learning Without Prompts},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}