Instructions to use Unbabel/TowerInstruct-Mistral-7B-v0.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Unbabel/TowerInstruct-Mistral-7B-v0.2 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="Unbabel/TowerInstruct-Mistral-7B-v0.2")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Unbabel/TowerInstruct-Mistral-7B-v0.2") model = AutoModelForCausalLM.from_pretrained("Unbabel/TowerInstruct-Mistral-7B-v0.2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
capricious model
Hi,
This is a very capricious model. I made a request to translate the subtitles, but it responded in my language and didn't give me a translation. The model said that it went to work on the translation. I wasn't able to get any results from it.
I noticed that models based on Mistral-Nemo-12B are smarter than many others. I would like to ask you to make an option based on this model. I am interested in a literary translation of the subtitles. This document format is quite difficult for many models: they either destroy the document by giving out continuous text or make literal and inappropriate translations.
If you or someone else decides to refine your model, I would ask you to inform it about what subtitles are and what Netflix and HBO standards are.
And I would prefer to work with a model trained on the literature corpora of the target translation language.
Regards
- I would also ask you to add glossaries of terms from the fields of science and technology: history, archeology, zoology, aviation, etc.