Table of Contents
Model Summary
We present BLOOMZ & mT0, a family of models capable of following human instructions in dozens of languages zero-shot. We finetune BLOOM & mT5 pretrained multilingual language models on our crosslingual task mixture (xP3) and find the resulting models capable of crosslingual generalization to unseen tasks & languages.
- Repository: bigscience-workshop/xmtf
- Paper: Crosslingual Generalization through Multitask Finetuning
- Point of Contact: Niklas Muennighoff
- Languages: Refer to bloom for pretraining & xP3 for finetuning language proportions. It understands both pretraining & finetuning languages.
- BLOOMZ & mT0 Model Family:
Use
Intended use
We recommend using the model to perform tasks expressed in natural language. For example, given the prompt "Translate to English: Je t’aime.", the model will most likely answer "I love you.". Some prompt ideas from our paper:
- 一个传奇的开端，一个不灭的神话，这不仅仅是一部电影，而是作为一个走进新时代的标签，永远彪炳史册。你认为这句话的立场是赞扬、中立还是批评?
- Suggest at least five related search terms to "Mạng neural nhân tạo".
- Write a fairy tale about a troll saving a princess from a dangerous dragon. The fairy tale is a masterpiece that has achieved praise worldwide and its moral is "Heroes Come in All Shapes and Sizes". Story (in Spanish):
- Explain in a sentence in Telugu what is backpropagation in neural networks.
Feel free to share your generations in the Community tab!
How to use
CPU
# pip install -q transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "bigscience/bloomz-7b1-p3"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForCausalLM.from_pretrained(checkpoint)
inputs = tokenizer.encode("Translate to English: Je t’aime.", return_tensors="pt")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
GPU
# pip install -q transformers accelerate
from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "bigscience/bloomz-7b1-p3"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForCausalLM.from_pretrained(checkpoint, torch_dtype="auto", device_map="auto")
inputs = tokenizer.encode("Translate to English: Je t’aime.", return_tensors="pt").to("cuda")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
GPU in 8bit
# pip install -q transformers accelerate bitsandbytes
from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "bigscience/bloomz-7b1-p3"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto", load_in_8bit=True)
inputs = tokenizer.encode("Translate to English: Je t’aime.", return_tensors="pt").to("cuda")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
Limitations
Prompt Engineering: The performance may vary depending on the prompt. For BLOOMZ models, we recommend making it very clear when the input stops to avoid the model trying to continue it. For example, the prompt "Translate to English: Je t'aime" without the full stop (.) at the end, may result in the model trying to continue the French sentence. Better prompts are e.g. "Translate to English: Je t'aime.", "Translate to English: Je t'aime. Translation:" "What is "Je t'aime." in English?", where it is clear for the model when it should answer. Further, we recommend providing the model as much context as possible. For example, if you want it to answer in Telugu, then tell the model, e.g. "Explain in a sentence in Telugu what is backpropagation in neural networks.".
Training
Model
- Architecture: Same as bloom-7b1, also refer to the
config.jsonfile - Finetuning steps: 1000
- Finetuning tokens: 4.19 billion
- Finetuning layout: 1x pipeline parallel, 1x tensor parallel, 64x data parallel
- Precision: float16
Hardware
- CPUs: AMD CPUs with 512GB memory per node
- GPUs: 64 A100 80GB GPUs with 8 GPUs per node (8 nodes) using NVLink 4 inter-gpu connects, 4 OmniPath links
- Communication: NCCL-communications network with a fully dedicated subnet
Software
- Orchestration: Megatron-DeepSpeed
- Optimizer & parallelism: DeepSpeed
- Neural networks: PyTorch (pytorch-1.11 w/ CUDA-11.5)
- FP16 if applicable: apex
Evaluation
We refer to Table 7 from our paper & bigscience/evaluation-results for zero-shot results on unseen tasks. The sidebar reports zero-shot performance of the best prompt per dataset config.
Citation
@misc{muennighoff2022crosslingual,
title={Crosslingual Generalization through Multitask Finetuning},
author={Niklas Muennighoff and Thomas Wang and Lintang Sutawika and Adam Roberts and Stella Biderman and Teven Le Scao and M Saiful Bari and Sheng Shen and Zheng-Xin Yong and Hailey Schoelkopf and Xiangru Tang and Dragomir Radev and Alham Fikri Aji and Khalid Almubarak and Samuel Albanie and Zaid Alyafeai and Albert Webson and Edward Raff and Colin Raffel},
year={2022},
eprint={2211.01786},
archivePrefix={arXiv},
primaryClass={cs.CL}
}