No more text2text
#1590
No more text2text
#1590
Conversation
Heads up this is pure vibe-coding _pre-LLM_, i.e. I'm not sure what I'm doing but I'm still doing it, manually (though I tried to take inspiration from #457) The goal is to address https://discuss.huggingface.co/t/no-0-models-returned-by-text2text-search-filter/161546 following huggingface-internal/moon-landing#14258
Pierrci
requested review from
SBrandeis,
gary149,
Wauplin,
julien-c,
pcuenca,
ngxson and
hanouticelina
as code owners
|
just dropping this here, anyone feel free to push on top of it x) |
| @@ -1,7 +1,6 @@ | |||
| import type { TaskDataCustom } from "../index.js"; | |||
|
|
|||
| const taskData: TaskDataCustom = { | |||
| canonicalId: "text2text-generation", | |||
There was a problem hiding this comment.
shouldn't we replace by
canonicalId: "text-generation",?
There was a problem hiding this comment.
well no, because since those examples were created, we added dedicated pipelines for both summarization and translation, iiuc, cc @SBrandeis
or is it not how it works? ^^'
There was a problem hiding this comment.
ok so this canonical thing was added in #449 actually, so quite recently, so ok with text-generation, even though the whole thing is still confusing to me since it seems we have top-level pipelines for summarization and translation
| @@ -1,7 +1,6 @@ | |||
| import type { TaskDataCustom } from "../index.js"; | |||
|
|
|||
| const taskData: TaskDataCustom = { | |||
| canonicalId: "text2text-generation", | |||
|
Nice! |
| [Text-to-Text generation models](https://huggingface.co/models?pipeline_tag=text2text-generation&sort=downloads) have a separate pipeline called `text2text-generation`. This pipeline takes an input containing the sentence including the task and returns the output of the accomplished task. | ||
|
|
||
| ```python | ||
| from transformers import pipeline | ||
|
|
||
| text2text_generator = pipeline("text2text-generation") | ||
| text2text_generator("question: What is 42 ? context: 42 is the answer to life, the universe and everything") | ||
| [{'generated_text': 'the answer to life, the universe and everything'}] | ||
|
|
||
| text2text_generator("translate from English to French: I'm very happy") | ||
| [{'generated_text': 'Je suis très heureux'}] | ||
| ``` | ||
|
|
There was a problem hiding this comment.
This still exists in transformers, so I'm not entirely sure we should remove it
There was a problem hiding this comment.
I'd suggest we keep it here but add a "historical note" comment. cc @LysandreJik @ArthurZucker for viz
There was a problem hiding this comment.
switched all the links to other instead of pipeline_tag then
|
let's go? |
There was a problem hiding this comment.
did we want to redirect https://huggingface.co/models?pipeline_tag=text2text-generation or we don't care?
| @@ -14,7 +14,7 @@ const taskData: TaskDataCustom = { | |||
| widgetModels: [], | |||
| youtubeId: undefined, | |||
| /// If this is a subtask, link to the most general task ID | |||
| /// (eg, text2text-generation is the canonical ID of translation) | |||
| /// (eg, text-generation is the canonical ID of text-simplification) | |||
Heads up this is pure vibe-coding pre-LLM, i.e. I'm not sure what I'm doing but I'm still doing it, manually (though I tried to take inspiration from #457)
The goal is to address https://discuss.huggingface.co/t/no-0-models-returned-by-text2text-search-filter/161546 following https://github.com/huggingface-internal/moon-landing/pull/14258