-
Notifications
You must be signed in to change notification settings - Fork 25.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add LlamaForQuestionAnswering
#28265
Comments
This represents the resolution for the repository issue on GitHub with the reference number huggingface#28265. huggingface#28265
Hey @NielsRogge I would like to work on this issue . |
@ArthurZucker @NielsRogge Is this feature still requested? |
Hey @nakranivaibhav , as you can see, @Tanmaypatil123 has already started working on it, let's not duplicate work ! 🤗 Unless the PR is not updated in a week or so, feel free to take over, starting from the review I did 😉 |
@ArthurZucker Alright, I'll keep an 👀 on it. |
@ArthurZucker Can i take the issue now? |
Sure, just feel free to open a PR and take into account my reviews! |
can i have that |
Looks like the change has been merged. Can it be closed? |
Feature request
Add a
LlamaForQuestionAnswering
class to themodeling_llama.py
so Llama models haveAutoModelForQuestionAnswering
support (by also adding Llama-style models to theMODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES
in themodeling_auto.py
file.Motivation
1 - Evaluation benchmarks like Squad or FaQUAD are commonly used to evaluate language models.
2 - Many decoder-only transformers (BLOOM, Falcon, OpenAI GPT-2, GPT Neo, GPT NeoX, GPT-J, etc.) have support for the
AutoModelForQuestionAnswering
.3 - Creating a fine-tuning/evaluation procedure using things like
AutoModelForQuestionAnswering
andevaluate.load('squad')
is very simple, making these features very helpful and desirable.4 - On the contrary, if one cannot use
AutoModelForQuestionAnswering
, like in the Llama style models, everything becomes more difficult.Hence, I would like to request the addition of a
LlamaForQuestionAnswering
class to themodeling_llama.py
file. Hence, we could all easily perform experiments with Llama models and squad-style Q&A benchmarks:Your contribution
I think, as suggested by nielsr in the forum, we can use the
GptjForQuestionAnswering
as a starting point, adding aLlamaForQuestionAnswering
to themodeling_llama.py
file:and then, we add the
Llama
models to theMODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES
in themodeling_auto.py
file:I can try to make these changes if no one more qualified wants to take the job 😅.
The text was updated successfully, but these errors were encountered: