Update README.md

README.md

@@ -1,3 +1,71 @@
----
-license: mit
----
+---
+license: mit
+datasets:
+- allenai/MADLAD-400
+language:
+- en
+- sw
+- id
+- ht
+- ko
+- el
+- ru
+- bg
+- uk
+- kk
+base_model:
+- mistralai/Mistral-7B-v0.1
+---
+VocADT is a solution for vocabulary adaptation using adapter modules that are trained to learn the optimal linear combination of existing embeddings while keeping the model’s weights fixed.
+VocADT offers a flexible and scalable solution without requiring external resources or language constraints.
+
+
+## New Vocabulary Adapted Models
+Only the input/output embeddings are replaced, while all other original weights of base model remain fixed.
+These are the merged version: after training the adapters, we merge the original embeddings with the adapter to generate the new embeddings.
+| Name | Adapted Model | Base Model | New Vocab Size | Focused Languages |
+|---|---|---|---|---|
+| VocADT-Latin-Mistral | [h-j-han/Mistral-7B-VocADT-50k-Latin](https://huggingface.co/h-j-han/Mistral-7B-VocADT-50k-Latin) | [Mistral](https://huggingface.co/mistralai/Mistral-7B-v0.1) | 50k | Swahili (sw), Indonesian (id), Estonian (et), Haitian Creole (ht), English (en)|
+| VocADT-Mixed-Mistral | [h-j-han/Mistral-7B-VocADT-50k-Mixed](https://huggingface.co/h-j-han/Mistral-7B-VocADT-50k-Mixed) | [Mistral](https://huggingface.co/mistralai/Mistral-7B-v0.1) | 50k | Korean (ko), Greek (el), Russian (ru), Bulgarian (bg), English (en) |
+| VocADT-Cyrillic-Mistral | [h-j-han/Mistral-7B-VocADT-50k-Cyrillic](https://huggingface.co/h-j-han/Mistral-7B-VocADT-50k-Cyrillic) | [Mistral](https://huggingface.co/mistralai/Mistral-7B-v0.1) | 50k | Russian (ru), Bulgarian (bg), Ukrainian (uk), Kazakh (kk), English (en) |
+| VocADT-All-Mistral | [h-j-han/Mistral-7B-VocADT-50k-All](https://huggingface.co/h-j-han/Mistral-7B-VocADT-50k-All) | [Mistral](https://huggingface.co/mistralai/Mistral-7B-v0.1) | 50k | Swahili (sw), Indonesian (id), Estonian (et), Haitian Creole (ht), Korean (ko), Greek (el), Russian (ru), Bulgarian (bg), Ukrainian (uk), Kazakh (kk), English (en) |
+|||||
+| VocADT-Latin-LLama | [h-j-han/Llama2-7B-VocADT-50k-Latin](https://huggingface.co/h-j-han/Llama2-7B-VocADT-50k-Latin) | [Llama](https://huggingface.co/meta-llama/Llama-2-7b-hf) | 50k | Swahili (sw), Indonesian (id), Estonian (et), Haitian Creole (ht), English (en)|
+| VocADT-Mixed-LLama | [h-j-han/Llama2-7B-VocADT-50k-Mixed](https://huggingface.co/h-j-han/Llama2-7B-VocADT-50k-Mixed) | [Llama](https://huggingface.co/meta-llama/Llama-2-7b-hf) | 50k | Korean (ko), Greek (el), Russian (ru), Bulgarian (bg), English (en) |
+| VocADT-Cyrillic-LLama | [h-j-han/Llama2-7B-VocADT-50k-Cyrillic](https://huggingface.co/h-j-han/Llama2-7B-VocADT-50k-Cyrillic) | [Llama](https://huggingface.co/meta-llama/Llama-2-7b-hf) | 50k | Russian (ru), Bulgarian (bg), Ukrainian (uk), Kazakh (kk), English (en) |
+
+
+## Quick Start
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+# model_name = "mistralai/Mistral-7B-v0.1" # Base Model
+model_name = "h-j-han/Mistral-7B-VocADT-50k-Mixed" # Vocabulary Adapted Model
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")
+prefix = "\nEnglish: Hello \nKorean: 안녕하세요 \nEnglish: Thank you\nKorean: 고맙습니다\nEnglish: "
+line = "I'm a student."
+suffix = f"\nKorean:"
+prompt = prefix + line + suffix
+inputs = tokenizer(prompt, return_tensors="pt")
+for item in inputs:
+    inputs[item] = inputs[item].cuda()
+outputs = model.generate(**inputs, max_new_tokens=88)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+# Base Model Output: "나는 학" # This short incomplete phrase in Korean is 5 tokens for the base model.
+# VocADT Output: "저는 학생입니다." # Complete and good output within 5 tokens
+```
+
+## Reference
+We provide code in Github repo : https://github.com/h-j-han/VocADT  
+Also, please find details in this paper :
+```
+@misc{han2024vocadt,
+      title={Adapters for Altering LLM Vocabularies: What Languages Benefit the Most?}, 
+      author={HyoJung Han and Akiko Eriguchi and Haoran Xu and Hieu Hoang and Marine Carpuat and Huda Khayrallah},
+      year={2024},
+      eprint={2410.09644},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL},
+      url={https://arxiv.org/abs/2410.09644}, 
+}
+```