Japanese-StableLM-Base-Alpha-7B

japanese-stablelm-icon

"A parrot able to speak Japanese, ukiyoe, edo period" โ€” Stable Diffusion XL

Model Description

japanese-stablelm-base-alpha-7b is a 7B-parameter decoder-only language model pre-trained on a diverse collection of Japanese and English datasets which focus on maximizing Japanese language modeling performance and Japanese downstream task performance.

For an instruction-following model, check Japanese-StableLM-Instruct-Alpha-7B and get access by accepting the terms and conditions.

Usage

First install additional dependencies in requirements.txt:

pip install sentencepiece einops

Then start generating text with japanese-stablelm-base-alpha-7b by using the following code snippet:

import torch
from transformers import LlamaTokenizer, AutoModelForCausalLM

tokenizer = LlamaTokenizer.from_pretrained("novelai/nerdstash-tokenizer-v1", additional_special_tokens=['โ–โ–'])

model = AutoModelForCausalLM.from_pretrained(
    "stabilityai/japanese-stablelm-base-alpha-7b",
    trust_remote_code=True,
)
model.half()
model.eval()

if torch.cuda.is_available():
    model = model.to("cuda")

prompt = """
AI ใง็ง‘ๅญฆ็ ”็ฉถใ‚’ๅŠ ้€Ÿใ™ใ‚‹ใซใฏใ€
""".strip()

input_ids = tokenizer.encode(
    prompt,
    add_special_tokens=False,
    return_tensors="pt"
)

# this is for reproducibility.
# feel free to change to get different result
seed = 23  
torch.manual_seed(seed)

tokens = model.generate(
    input_ids.to(device=model.device),
    max_new_tokens=128,
    temperature=1,
    top_p=0.95,
    do_sample=True,
)

out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
"""
AI ใง็ง‘ๅญฆ็ ”็ฉถใ‚’ๅŠ ้€Ÿใ™ใ‚‹ใซใฏใ€ใƒ‡ใƒผใ‚ฟ้ง†ๅ‹•ๅž‹ๆ–‡ๅŒ–ใŒๅฟ…่ฆใงใ‚ใ‚‹ใ“ใจใ‚‚ๆ˜Žใ‚‰ใ‹ใซใชใฃใฆใใฆใ„ใพใ™ใ€‚็ ”็ฉถใฎใ‚ใ‚‰ใ‚†ใ‚‹ๅด้ขใงใ€ใƒ‡ใƒผใ‚ฟใŒใ‚ˆใ‚Š้‡่ฆใซใชใฃใฆใ„ใ‚‹ใฎใงใ™ใ€‚
20 ไธ–็ด€ใฎ็ง‘ๅญฆใฏใ€็ ”็ฉถ่€…ใŒ็›ดๆŽฅ็ ”็ฉถใ‚’่กŒใ†ใ“ใจใงใ€็ ”็ฉถใƒ‡ใƒผใ‚ฟใ‚’ๆดป็”จใ—ใฆใใพใ—ใŸใ€‚ใใฎๅพŒใ€ๅคšใใฎ็ง‘ๅญฆๅˆ†้‡Žใงใฏใƒ‡ใƒผใ‚ฟใฏๆ‰‹ๅ‹•ใงๅˆ†ๆžใ•ใ‚Œใ‚‹ใ‚ˆใ†ใซใชใฃใŸใ‚‚ใฎใฎใ€ใ“ใ‚Œใ‚‰ใฎๆ–นๆณ•ใซใฏๅคšๅคงใชใ‚ณใ‚นใƒˆใจๅŠดๅŠ›ใŒใ‹ใ‹ใ‚‹ใ“ใจใŒๅˆ†ใ‹ใ‚Šใพใ—ใŸใ€‚ ใใ“ใงใ€ๅคšใใฎ็ ”็ฉถ่€…ใ‚„็ ”็ฉถ่€…ใ‚ฐใƒซใƒผใƒ—ใฏใ€ใ‚ˆใ‚ŠๅŠน็Ž‡็š„ใชๆ‰‹ๆณ•ใ‚’้–‹็™บใ—ใ€็ ”็ฉถใฎ่ฆๆจกใ‚’ๆ‹กๅคงใ—ใฆใใพใ—ใŸใ€‚21 ไธ–็ด€ใซใชใ‚‹ใจใ€็ ”็ฉถ่€…ใŒๆ‰‹ๅ‹•ใงๅฎŸๆ–ฝใ™ใ‚‹ๅฟ…่ฆใฎใ‚ใ‚‹็ ”็ฉถใฏใ€ใใฎๅคง้ƒจๅˆ†ใ‚’็ ”็ฉถ่€…ใŒ่‡ชๅ‹•ๅŒ–ใงใใ‚‹ใ‚ˆใ†ใซใชใ‚Šใพใ—ใŸใ€‚
"""

We suggest playing with different generation config (top_p, repetition_penalty etc) to find the best setup for your tasks. For example, use higher temperature for roleplay task, lower temperature for reasoning.

Model Details

  • Model type: japanese-stablelm-base-alpha-7b model is an auto-regressive language model based on the NeoX transformer architecture.
  • Language(s): Japanese
  • Library: GPT-NeoX
  • License: This model is licensed under Apache License, Version 2.0.

Training

Parameters Hidden Size Layers Heads Sequence Length
7B 4096 32 32 2048

Training Dataset

japanese-stablelm-base-alpha-7b is pre-trained on around 750B tokens from a mixture of the following corpora:

Use and Limitations

Intended Use

The model is intended to be used by all individuals as foundational models for application-specific fine-tuning without strict limitations on commercial use.

Limitations and bias

The pre-training dataset may have contained offensive or inappropriate content even after applying data cleansing filters which can be reflected in the model generated text. We recommend users exercise reasonable caution when using these models in production systems. Do not use the model for any applications that may cause harm or distress to individuals or groups.

Authors

Acknowledgements

We are utilizing the v1 version of the novelai-tokenizer, introduced by NovelAI, because it processes both Japanese and English text effectively and efficiently. We extend our gratitude to NovelAI for allowing us to use their remarkable work. For more details about the tokenizer, please refer to their blog post.

We are grateful for the contributions of the EleutherAI Polyglot-JA team in helping us to collect a large amount of pre-training data in Japanese. Polyglot-JA members includes Hyunwoong Ko (Project Lead), Fujiki Nakamura (originally started this project when he commited to the Polyglot team), Yunho Mo, Minji Jung, KeunSeok Im, and Su-Kyeong Jang.

We are also appreciative of AI Novelist/Sta (Bit192, Inc.) and the numerous contributors from Stable Community Japan for assisting us in gathering a large amount of high-quality Japanese textual data for model training.

How to cite

@misc{JapaneseStableLMBaseAlpha7B, 
      url={[https://huggingface.co/stabilityai/japanese-stablelm-base-alpha-7b](https://huggingface.co/stabilityai/japanese-stablelm-base-alpha-7b)}, 
      title={Japanese StableLM Base Alpha 7B}, 
      author={Lee, Meng and Nakamura, Fujiki and Shing, Makoto and McCann, Paul and Akiba, Takuya and Orii, Naoki}
}

Citations

@software{gpt-neox-library,
  title = {{GPT-NeoX: Large Scale Autoregressive Language Modeling in PyTorch}},
  author = {Andonian, Alex and Anthony, Quentin and Biderman, Stella and Black, Sid and Gali, Preetham and Gao, Leo and Hallahan, Eric and Levy-Kramer, Josh and Leahy, Connor and Nestler, Lucas and Parker, Kip and Pieler, Michael and Purohit, Shivanshu and Songz, Tri and Phil, Wang and Weinbach, Samuel},
  url = {https://www.github.com/eleutherai/gpt-neox},
  doi = {10.5281/zenodo.5879544},
  month = {8},
  year = {2021},
  version = {0.0.1},
}
Downloads last month
1,761
Inference Examples
Inference API (serverless) does not yet support model repos that contain custom code.

Model tree for stabilityai/japanese-stablelm-base-alpha-7b

Adapters
1 model

Datasets used to train stabilityai/japanese-stablelm-base-alpha-7b

Spaces using stabilityai/japanese-stablelm-base-alpha-7b 2

Collection including stabilityai/japanese-stablelm-base-alpha-7b