Instructions to use webbigdata/ALMA-7B-Ja-V2-GPTQ-Ja-En with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use webbigdata/ALMA-7B-Ja-V2-GPTQ-Ja-En with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "translation" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # pip install "transformers<5.0.0" from transformers import pipeline pipe = pipeline("translation", model="webbigdata/ALMA-7B-Ja-V2-GPTQ-Ja-En")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("webbigdata/ALMA-7B-Ja-V2-GPTQ-Ja-En") model = AutoModelForCausalLM.from_pretrained("webbigdata/ALMA-7B-Ja-V2-GPTQ-Ja-En", device_map="auto") - Notebooks
- Google Colab
- Kaggle
お疲れ様です translates to Good evening
Thank you for this model.
I was evaluating it for a personal project and it is really good. However, something that bothers me is that it translates お疲れ様です as good evening. I would like to know the reason behind this behaviour.
Hello.
In Japan, "お疲れ様です." is used in a 2 way.
Originally, it was meant to express appreciation for the work of others.
However, it is often used to greet colleagues (especially at the end of a workday).
In the latter sense, good evening or Have a nice weekend! is not a mistake.
In what situations is this translation causing trouble?
Thanks for the response.
In my case, I am translating work emails which I receive during the day which contains sentences such as お疲れ様です。今日、資料を送ります。
I think in this case, it should not translate to good evening.
I see.
Thanks. I'll try to be aware of that in the next version (which I'm currently struggling with quite a bit).
Thank you. Good luck with the next version.
Wow. That is awesome. Looks good to me.
It will take a little more time to finish.
No problem. Thanks for considering my request.
Hello.
I have published as a QLoRA adapter for gemma.
Despite the 4-bit quantization, the memory requirement has increased to 8.1 GB. However, it is possible to run it with the free version of Colab, so I made it a priority to release it to the public first!
