Text Generation
Transformers
Safetensors
Bulgarian
English
llama
conversational
text-generation-inference
Inference Endpoints
zolicsaki's picture
Update README.md
b868179 verified
metadata
license: llama2
datasets:
  - HuggingFaceH4/ultrachat_200k
  - HuggingFaceH4/ultrafeedback_binarized
  - HuggingFaceH4/cai-conversation-harmless
language:
  - bg
  - en

SambaLingo-Bulgarian-Chat

SambaLingo-Bulgarian-Chat is a human aligned chat model trained in Bulgarian and English. It is trained using direct preference optimization on top the base model SambaLingo-Bulgarian-Base. The base model adapts Llama 2 to Bulgarian by training on 63 billion tokens from the Bulgarian split of the Cultura-X dataset. Try this model at SambaLingo-chat-space.

Model Description

Getting Started

Loading Model With Hugging Face

Please make sure to set use_fast=False when loading the tokenizer.

from transformers import AutoModelForCausalLM, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("sambanovasystems/SambaLingo-Bulgarian-Chat", use_fast=False)
model = AutoModelForCausalLM.from_pretrained("sambanovasystems/SambaLingo-Bulgarian-Chat", device_map="auto", torch_dtype="auto")

Interacting With Model Pipeline

Please make sure to set use_fast=False when loading the tokenizer.

from transformers import pipeline
pipe = pipeline("text-generation", model="sambanovasystems/SambaLingo-Bulgarian-Chat", device_map="auto", use_fast=False)
messages = [
                {"role": "user", "content": {YOUR_QUESTION}},
]
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipe(prompt)[0]
outputs = outputs["generated_text"]

Suggested Inference Parameters

  • Temperature: 0.8
  • Repetition penalty: 1.0
  • Top-p: 0.9

Prompting Guidelines

To prompt this model, please use the following chat template:

<|user|>\n{question}</s>\n<|assistant|>\n

Example Prompts and Generations

<|user|>
Какъв извод можеш да извадиш за чувството, което изпитват героите от следната кратка история:

Мария чакаше с трепет да дойде 27-ми. Още само три дни и корабът щеше да акостира на пристанището. След 6 дълги месеца тя най-после щеше да види Иван. Когато денят най-после дойде, тя отиде на пристанището заедно с десетки други жени. Моряците един по един започнаха да слизат. Най-накрая тя съзря погледа на Иван, и двамата се прегърнаха силно.</s>
<|assistant|>
Мария изпитва силно чувство на очакване и нетърпение да види Иван след шестмесечна раздяла. Тя е развълнувана и щастлива да го види отново и прегръдката им показва, че и двамата са били много развълнувани да се съберат отново.

Training Details

The alignment phase follows the recipe for Zephyr-7B, and comprises two stages: supervised fine-tuning (SFT) and Direct Performance Optimization (DPO).

The SFT phase was done on the ultrachat_200k dataset mixed with the Google translated version of the ultrachat_200k dataset. It was trained for one epoch with global batch size 512 and max sequence length 2048 tokens. We used a linear decay learning rate of 2e-5 and 10% warmup.

The DPO phase was done on the ultrafeedback dataset and cai-conversation-harmless dataset, mixed with 10% of the data Google translated. It was trained with global batch size 32 and for three epochs. We used a linear decay learning rate of 5e-7, 10% warmup and β=0.1 as the regularization factor for DPO.

Tokenizer Details

We extended the vocabulary of the base llama model from 32,000 tokens to 57,000 tokens by adding up to 25,000 non-overlapping tokens from the new language.

Evaluation

For evaluation results see our paper: SambaLingo: Teaching Large Language Models New Languages

Uses

Direct Use

Use of this model is governed by the Meta’s Llama 2 Community License Agreement. Please review and accept the license before downloading the model weights.

Out-of-Scope Use

SambaLingo should NOT be used for:

  • Mission-critical applications
  • Applications that involve the safety of others
  • Making highly important decisions

Bias, Risks, and Limitations

Like all LLMs, SambaLingo has certain limitations:

  • Hallucination: Model may sometimes generate responses that contain plausible-sounding but factually incorrect or irrelevant information.
  • Code Switching: The model might unintentionally switch between languages or dialects within a single response, affecting the coherence and understandability of the output.
  • Repetition: The Model may produce repetitive phrases or sentences, leading to less engaging and informative responses.
  • Coding and Math: The model's performance in generating accurate code or solving complex mathematical problems may be limited.
  • Toxicity: The model could inadvertently generate responses containing inappropriate or harmful content.

Acknowledgments

We extend our heartfelt gratitude to the open-source AI community; this endeavor would not have been possible without open source. SambaNova embraces the open-source community and aspires to actively contribute to this initiative.

We would like to give a special thanks to the following groups:

  • Meta for open sourcing LLama 2 and open sourcing FLORES-200 dataset
  • Nguyen et al for open sourcing CulturaX dataset
  • CohereAI for releasing AYA-101 and open sourcing a multilingual instruction tuning dataset
  • EleutherAI for their open source evaluation framework
  • Hugging Face-H4 team for open source the zephyr training recipe and alignment handbook repo

Cite SambaLingo

@misc{csaki2024sambalingo,
      title={SambaLingo: Teaching Large Language Models New Languages}, 
      author={Zoltan Csaki and Bo Li and Jonathan Li and Qiantong Xu and Pian Pawakapan and Leon Zhang and Yun Du and Hengyu Zhao and Changran Hu and Urmish Thakker},
      year={2024},
      eprint={2404.05829},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}