This repository is publicly accessible, but you have to accept the conditions to access its files and content.
To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging Face and click below. Requests are processed immediately.
Log in or Sign Up to review the conditions and access this model content.
This repository corresponds to the research Gemma PyTorch repository. If you're looking for the transformers implementation, visit this page
We are aware that there are compatibility issues with Torch v2.3.0. Please use Torch v2.1.0 in the interim.
Terms of Use: Terms
Authors: Google
Summary description and brief definition of inputs and outputs.
Gemma is a series of best-in-class open models and draws inspiration and technological lineage from the Gemini family of models. They are text-to-text, decoder-only large language models with open weights. Gemma models are well-suited for a variety of text generation tasks, including question answering, summarization, and reasoning.
Gemma-2-JPN is a Gemma 2 2B model fine-tuned on Japanese text. It supports the Japanese language with the same level of performance of English only queries on Gemma 2.
First, download and install the gemma_pytorch library:
git clone https://github.com/google/gemma_pytorch.git
cd gemma_pytorch
pip install .
You can then load the model and tokenizer using:
from huggingface_hub import snapshot_download
from gemma.config import GemmaConfig, get_model_config
from gemma.model import GemmaForCausalLM
from gemma.tokenizer import Tokenizer
snapshot_dir = snapshot_download(repo_id='google/gemma-2-2b-jpn-it-pytorch')
# Ensure that the tokenizer is present
tokenizer_path = os.path.join(snapshot_dir, 'tokenizer.model')
assert os.path.isfile(tokenizer_path), 'Tokenizer not found!'
# Ensure that the checkpoint is present
ckpt_path = os.path.join(snapshot_dir, f'model.ckpt')
assert os.path.isfile(ckpt_path), 'PyTorch checkpoint not found!'
model_config = get_model_config("2b-v2")
model_config.tokenizer = tokenizer_path
# Instantiate the model and load the weights.
torch.set_default_dtype(model_config.get_dtype())
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = GemmaForCausalLM(model_config)
model.load_weights(ckpt_path)
model = model.to(device).eval()
Finally, you can generate text as follows:
def format_gemma_instruction(instruction: str) -> str:
return f"<start_of_turn>user\n{instruction}<end_of_turn><start_of_turn>model\n"
input_text = "マシーンラーニングについての詩を書いてください。"
input_formatted = format_gemma_instruction(input_text)
output = model.generate(input_formatted, device=device, output_len=1024)
print(output)
Data used for model training and how the data was processed.
These models were trained on a dataset of text data that includes a wide variety of sources, totaling 8 trillion tokens. Here are the key components:
The combination of these diverse data sources is crucial for training a powerful language model that can handle a wide variety of different tasks and text formats.
Here are the key data cleaning and filtering methods applied to the training data:
Details about the model internals.
Gemma was trained using the latest generation of Tensor Processing Unit (TPU) hardware (TPUv5p).
Training large language models requires significant computational power. TPUs, designed specifically for matrix operations common in machine learning, offer several advantages in this domain:
These advantages are aligned with Google's commitments to operate sustainably.
Training was done using JAX and ML Pathways.
JAX allows researchers to take advantage of the latest generation of hardware, including TPUs, for faster and more efficient training of large models.
ML Pathways is Google's latest effort to build artificially intelligent systems capable of generalizing across multiple tasks. This is specially suitable for foundation models, including large language models like these ones.
Together, JAX and ML Pathways are used as described in the paper about the Gemini family of models; "the 'single controller' programming model of Jax and Pathways allows a single Python process to orchestrate the entire training run, dramatically simplifying the development workflow."
To assess the quality of this model, we collected a diverse set of Japanese prompts and evaluated performance using an LLM-as-a-judge approach against GPT-3.5. The rating system is based on a 7-scale assessments, which are MuchBetterThan, BetterThan, SlightlyBetterThan, AboutTheSame, SlightlyWorse, WorseThan, MuchWorseThan associated with the numerical scores 1.5, 1.0, 0.5, 0, -0.5, -1.0, -1.5 respectively. We also tracked the ability of the model to answer in the correct language: for a Japanese prompt, the model should typically answer in Japanese rather than defaulting to English.
Benchmark |
Gemma-2-IT |
Gemma-2-IT-JPN |
|
|---|---|---|---|
Preference vs GPT-3.5 |
-0.25 ± 0.05 |
0.03 ± 0.04 |
|
Language correctness |
86.47% |
98.24% |
Ethics and safety evaluation approach and results.
Our evaluation methods include structured evaluations and internal red-teaming testing of relevant content policies. Red-teaming was conducted by a number of different teams, each with different goals and human evaluation metrics. These models were evaluated against a number of different categories relevant to ethics and safety, including:
These models have certain limitations that users should be aware of.
Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development.
The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following:
Risks identified and mitigations:
At the time of release, this family of models provides high-performance open large language model implementations designed from the ground up for Responsible AI development compared to similarly sized models.