license: apache-2.0
language:
- en
- ja
programming_language:
- C
- C++
- C#
- Go
- Java
- JavaScript
- Lua
- PHP
- Python
- Ruby
- Rust
- Scala
- TypeScript
library_name: transformers
pipeline_tag: text-generation
inference: false
llm-jp-13b-v1.0
This repository provides large language models developed by LLM-jp, a collaborative project launched in Japan.
| Pre-trained models |
| llm-jp-13b-v1.0 |
| llm-jp-1.3b-v1.0 |
Checkpoints format: transformers (Megatron-DeepSpeed format available here) |
Required Libraries and Their Versions
- torch>=2.0.0
- transformers>=4.34.0
- tokenizers>=0.14.0
Usage
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("llm-jp/llm-jp-13b-v1.0")
model = AutoModelForCausalLM.from_pretrained("llm-jp/llm-jp-13b-v1.0", torch_dtype=torch.float16)
text = "自然言語処理とは何か"
tokenized_input = tokenizer.encode(text, add_special_tokens=False, return_tensors="pt").to(model.device)
with torch.no_grad():
output = model.generate(
tokenized_input,
max_new_tokens=100,
do_sample=True,
top_p=0.95,
temperature=0.7,
)[0]
print(tokenizer.decode(output))
Model Details
- Model type: Transformer-based Language Model
- Total seen tokens: 300B
| Model | Params | Layers | Hidden size | Heads | Context length |
|---|---|---|---|---|---|
| 13b model | 13b | 40 | 5120 | 40 | 2048 |
| 1.3b model | 1.3b | 24 | 2048 | 16 | 2048 |
Training
Pre-training:
- Hardware: 96 A100 40GB GPUs (mdx cluster)
- Software: Megatron-DeepSpeed
Instruction tuning:
- Hardware: 8 A100 40GB GPUs (mdx cluster)
- Software: TRL, PEFT, and DeepSpeed
Tokenizer
The tokenizer of this model is based on huggingface/tokenizers Unigram byte-fallback model.
The vocab entries were converted from llm-jp-tokenizer v2.1 (50k).
Please refer to README.md of llm-ja-tokenizer for the details of vocab constuction steps.
- Model: Hugging Face Fast Tokenizer using Unigram byte-fallback model which requires
tokenizers>=0.14.0 - Training algorithm: SentencePiece Unigram byte-fallback
- Training data: A subset of the datasets for model pre-training
- Vocabulary size: 50,570 (mixed vocabulary of Japanese, English, and source code)
Datasets
Pre-training
The models have been pre-trained on approximately 287.5B tokens, sourced from a blend of the following datasets.
| Language | Dataset | Tokens |
|---|---|---|
| Japanese | Wikipedia | 1.5B |
| mC4 | 136B | |
| English | Wikipedia | 5B |
| The Pile | 135B | |
| Codes | The Stack | 10B |
Pretraining was done by 10-hold shards that consists approx. 27-28B tokens. We further finalized the pretraining with additional cleaned 27B tokens data.
Instruction tuning
The models have been fine-tuned on the following datasets.
| Language | Dataset | description |
|---|---|---|
| Japanese | jaster | An automatically transformed data from the existing Japanese NLP datasets |
| databricks-dolly-15k | A translated one by DeepL in LLM-jp | |
| OpenAssistant Conversations Dataset | A translated one by DeepL in LLM-jp |
Evaluation
You can view the evaluation results of several LLMs on this leaderboard. We used llm-jp-eval for the evaluation.
Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
Send Questions to
llm-jp(at)nii.ac.jp
License
Model Card Authors
The names are listed in alphabetical order.
Namgi Han, Hirokazu Kiyomaru, Hiroshi Matsuda, Shota Sasaki, Shuhei Kurita, Taishi Nakamura, Takumi Okamoto.