Granite-3.0-3B-A800M-Instruct

Model Summary

• Developer: IBM Research
• GitHub Repository: ibm-granite/granite-language-models
• Website: Granite Docs
• Paper: Granite Language Models
• Release Date: October 21st, 2024

Granite-3.0-3B-A800M-Instruct is a 3B parameter model finetuned from Granite-3.0-3B-A800M-Base-4K using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets. This model is developed using a diverse set of techniques with a structured chat format, including supervised finetuning, model alignment using reinforcement learning, and model merging.

Third-Party Community Consideration

This model is not owned or developed by NVIDIA. This model has been developed and built to a third-party’s requirements for this application and use case; see link to Non-NVIDIA *Granite-3.0-3B-A800M-Base-4K model card.

License/Terms of Use:

GOVERNING TERMS: The trial service is governed by the NVIDIA API Trial Terms of Service; and the use of this model is governed by the NVIDIA AI Foundation Models Community License Agreement. ADDITIONAL INFORMATION: Apache 2.0 License.

Model Architecture:

Architecture Type: Transformer
Network Architecture: [Other - MoE]

Usage

Intended use

The model is designed to respond to general instructions and can be used to build AI assistants for multiple domains, including bussiness applications.

Capabilities

• Summarization
• Text classification
• Text extraction
• Question-answering
• Retrieval Augmented Generation (RAG)
• Code related
• Function-calling
• Multilingual dialog use cases

Input:

Input Type(s): [ Text]
Input Format(s): [String]
Input Parameters: min_tokens, max_tokens, temperature, top_p, stop, frequency_penalty, presence_penalty
Other Properties Related to Input: Granite instruct models are primarily finetuned using instruction-response pairs mostly in English, but also in German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese (Simplified). As this model has been exposed to multilingual data, it can handle multilingual dialog use cases with a limited performance in non-English tasks. In such case, introducing a small number of examples (few-shot) can help the model in generating more accurate outputs.\

Output:

Output Type(s): Text
Output Format: String
Output Parameters: None
Other Properties Related to Output: None

Generation

This is a simple example of how to use Granite-3.0-3B-A800M-Instruct model.

Install the following libraries:

Copy Code
pip install torch torchvision torchaudio
pip install accelerate
pip install transformers

Then, copy the snippet from the section that is relevant for your usecase.

Copy Code
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

device = "auto"
model_path = "ibm-granite/granite-3.0-3b-a800m-instruct"
tokenizer = AutoTokenizer.from_pretrained(model_path)
# drop device_map if running on CPU
model = AutoModelForCausalLM.from_pretrained(model_path, device_map=device)
model.eval()
# change input text as desired
chat = [
    { "role": "user", "content": "Please list one IBM Research laboratory located in the United States. You should only output its name and location." },
]
chat = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
# tokenize the text
input_tokens = tokenizer(chat, return_tensors="pt").to(device)
# generate output tokens
output = model.generate(**input_tokens, 
                        max_new_tokens=100)
# decode output tokens into text
output = tokenizer.batch_decode(output)
# print output
print(output)

Model Architeture

Granite-3.0-3B-A800M-Instruct is based on a decoder-only sparse Mixture of Experts(MoE) transformer architecture. Core components of this architecture are: Fine-grained Experts, Dropless Token Routing, and Load Balancing Loss.

Model	2B Dense	8B Dense	1B MoE	3B MoE
Embedding size	2048	4096	1024	1536
Number of layers	40	40	24	32
Attention head size	64	128	64	64
Number of attention heads	32	32	16	24
Number of KV heads	8	8	8	8
MLP hidden size	8192	12800	512	512
MLP activation	SwiGLU	SwiGLU	SwiGLU	SwiGLU
Number of Experts	—	—	32	40
MoE TopK	—	—	8	8
Initialization std	0.1	0.1	0.1	0.1
Sequence Length	4096	4096	4096	4096
Position Embedding	RoPE	RoPE	RoPE	RoPE
# Paremeters	2.5B	8.1B	1.3B	3.3B
# Active Parameters	2.5B	8.1B	400M	800M
# Training tokens	12T	12T	10T	10T

Training Data

Granite Language Instruct models are trained on a selection of open-srouce instruction datasets with a non-restrictive license, as well as a collection of synthetic datasets created by IBM. Together, these instruction datasets are a solid representation of the following domains: English, multilingual, code, math, tools, and safety.

Infrastructure

We train the Granite Language models using IBM's super computing cluster, Blue Vela, which is outfitted with NVIDIA H100 GPUs. This cluster provides a scalable and efficient infrastructure for training our models over thousands of GPUs.

Model Version(s):

Granite-MoE-3.0-instruct

Ethical Considerations and Limitations

The model inherits ethical considerations and limitations from its base model.

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications.  When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.  

Please report security vulnerabilities or NVIDIA AI Concerns here.