Model Summary: Granite-3.1-8B-Instruct is a 8B parameter long-context instruct model finetuned from Granite-3.1-8B-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems. 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.
This model can be deployed efficiently on vLLM, Red Hat Enterprise Linux AI, and Openshift AI, as shown in the example below.
Deploy on vLLM
from vllm import LLM, SamplingParams
from transformers import AutoTokenizer
model_id = "RedHatAI/granite-3.1-8b-instruct"
number_gpus = 1
sampling_params = SamplingParams(temperature=0.7, top_p=0.8, max_tokens=256)
tokenizer = AutoTokenizer.from_pretrained(model_id)
prompt = "Give me a short introduction to large language model."
llm = LLM(model=model_id, tensor_parallel_size=number_gpus)
outputs = llm.generate(prompt, sampling_params)
generated_text = outputs[0].outputs[0].text
print(generated_text)
vLLM also supports OpenAI-compatible serving. See the documentation for more details.
podman run --rm -it --device nvidia.com/gpu=all -p 8000:8000 \
--ipc=host \
--env "HUGGING_FACE_HUB_TOKEN=$HF_TOKEN" \
--env "HF_HUB_OFFLINE=0" -v ~/.cache/vllm:/home/vllm/.cache \
--name=vllm \
registry.access.redhat.com/rhaiis/rh-vllm-cuda \
vllm serve \
--tensor-parallel-size 1 \
--max-model-len 32768 \
--enforce-eager --model RedHatAI/granite-3.1-8b-instruct
See Red Hat AI Inference Server documentation for more details.
# Download model from Red Hat Registry via docker
# Note: This downloads the model to ~/.cache/instructlab/models unless --model-dir is specified.
ilab model download --repository docker://registry.redhat.io/rhelai1/granite-3-1-8b-instruct:1.5
# Serve model via ilab
ilab model serve --model-path ~/.cache/instructlab/models/granite-3-1-8b-instruct -- --trust-remote-code
# Chat with model
ilab model chat --model ~/.cache/instructlab/models/granite-3-1-8b-instruct
See Red Hat Enterprise Linux AI documentation for more details.
# Setting up vllm server with ServingRuntime
# Save as: vllm-servingruntime.yaml
apiVersion: serving.kserve.io/v1alpha1
kind: ServingRuntime
metadata:
name: vllm-cuda-runtime # OPTIONAL CHANGE: set a unique name
annotations:
openshift.io/display-name: vLLM NVIDIA GPU ServingRuntime for KServe
opendatahub.io/recommended-accelerators: '["nvidia.com/gpu"]'
labels:
opendatahub.io/dashboard: 'true'
spec:
annotations:
prometheus.io/port: '8080'
prometheus.io/path: '/metrics'
multiModel: false
supportedModelFormats:
- autoSelect: true
name: vLLM
containers:
- name: kserve-container
image: quay.io/modh/vllm:rhoai-2.20-cuda # CHANGE if needed. If AMD: quay.io/modh/vllm:rhoai-2.20-rocm
command:
- python
- -m
- vllm.entrypoints.openai.api_server
args:
- "--port=8080"
- "--model=/mnt/models"
- "--served-model-name={{.Name}}"
env:
- name: HF_HOME
value: /tmp/hf_home
ports:
- containerPort: 8080
protocol: TCP
# Attach model to vllm server. This is an NVIDIA template
# Save as: inferenceservice.yaml
apiVersion: serving.kserve.io/v1beta1
kind: InferenceService
metadata:
annotations:
openshift.io/display-name: granite-3-1-8b-instruct # OPTIONAL CHANGE
serving.kserve.io/deploymentMode: RawDeployment
name: granite-3-1-8b-instruct # specify model name. This value will be used to invoke the model in the payload
labels:
opendatahub.io/dashboard: 'true'
spec:
predictor:
maxReplicas: 1
minReplicas: 1
model:
args:
- '--trust-remote-code'
modelFormat:
name: vLLM
name: ''
resources:
limits:
cpu: '2' # this is model specific
memory: 8Gi # this is model specific
nvidia.com/gpu: '1' # this is accelerator specific
requests: # same comment for this block
cpu: '1'
memory: 4Gi
nvidia.com/gpu: '1'
runtime: vllm-cuda-runtime # must match the ServingRuntime name above
storageUri: oci://registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-instruct:1.5
tolerations:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
# make sure first to be in the project where you want to deploy the model
# oc project <project-name>
# apply both resources to run model
# Apply the ServingRuntime
oc apply -f vllm-servingruntime.yaml
# Apply the InferenceService
oc apply -f qwen-inferenceservice.yaml
# Replace <inference-service-name> and <cluster-ingress-domain> below:
# - Run `oc get inferenceservice` to find your URL if unsure.
# Call the server using curl:
curl https://<inference-service-name>-predictor-default.<domain>/v1/chat/completions
-H "Content-Type: application/json" \
-d '{
"model": "granite-3-1-8b-instruct",
"stream": true,
"stream_options": {
"include_usage": true
},
"max_tokens": 1,
"messages": [
{
"role": "user",
"content": "How can a bee fly when its wings are so small?"
}
]
}'
See Red Hat Openshift AI documentation for more details.
Supported Languages: English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese. Users may finetune Granite 3.1 models for languages beyond these 12 languages.
Intended Use: The model is designed to respond to general instructions and can be used to build AI assistants for multiple domains, including business applications.
Capabilities
Generation: This is a simple example of how to use Granite-3.1-8B-Instruct model.
Install the following libraries:
pip install torch torchvision torchaudio
pip install accelerate
pip install transformers
Then, copy the snippet from the section that is relevant for your use case.
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "auto"
model_path = "ibm-granite/granite-3.1-8b-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)
Evaluation Results:
| Models | ARC-Challenge | Hellaswag | MMLU | TruthfulQA | Winogrande | GSM8K | Avg |
|---|---|---|---|---|---|---|---|
| Granite-3.1-8B-Instruct | 62.62 | 84.48 | 65.34 | 66.23 | 75.37 | 73.84 | 71.31 |
| Granite-3.1-2B-Instruct | 54.61 | 75.14 | 55.31 | 59.42 | 67.48 | 52.76 | 60.79 |
| Granite-3.1-3B-A800M-Instruct | 50.42 | 73.01 | 52.19 | 49.71 | 64.87 | 48.97 | 56.53 |
| Granite-3.1-1B-A400M-Instruct | 42.66 | 65.97 | 26.13 | 46.77 | 62.35 | 33.88 | 46.29 |
| Models | IFEval | BBH | MATH Lvl 5 | GPQA | MUSR | MMLU-Pro | Avg |
|---|---|---|---|---|---|---|---|
| Granite-3.1-8B-Instruct | 72.08 | 34.09 | 21.68 | 8.28 | 19.01 | 28.19 | 30.55 |
| Granite-3.1-2B-Instruct | 62.86 | 21.82 | 11.33 | 5.26 | 4.87 | 20.21 | 21.06 |
| Granite-3.1-3B-A800M-Instruct | 55.16 | 16.69 | 10.35 | 5.15 | 2.51 | 12.75 | 17.1 |
| Granite-3.1-1B-A400M-Instruct | 46.86 | 6.18 | 4.08 | 0 | 0.78 | 2.41 | 10.05 |
Model Architecture: Granite-3.1-8B-Instruct is based on a decoder-only dense transformer architecture. Core components of this architecture are: GQA and RoPE, MLP with SwiGLU, RMSNorm, and shared input/output embeddings.
| 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 | 128K | 128K | 128K | 128K |
| Position embedding | RoPE | RoPE | RoPE | RoPE |
| # Parameters | 2.5B | 8.1B | 1.3B | 3.3B |
| # Active parameters | 2.5B | 8.1B | 400M | 800M |
| # Training tokens | 12T | 12T | 10T | 10T |
Training Data: Overall, our SFT data is largely comprised of three key sources: (1) publicly available datasets with permissive license, (2) internal synthetic data targeting specific capabilities including long-context tasks, and (3) very small amounts of human-curated data. A detailed attribution of datasets can be found in the Granite 3.0 Technical Report, Granite 3.1 Technical Report (coming soon), and Accompanying Author List.
Infrastructure: We train Granite 3.1 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.
Ethical Considerations and Limitations: Granite 3.1 Instruct Models are primarily finetuned using instruction-response pairs mostly in English, but also multilingual data covering eleven languages. Although this model can handle multilingual dialog use cases, its performance might not be similar to English tasks. In such case, introducing a small number of examples (few-shot) can help the model in generating more accurate outputs. While this model has been aligned by keeping safety in consideration, the model may in some cases produce inaccurate, biased, or unsafe responses to user prompts. So we urge the community to use this model with proper safety testing and tuning tailored for their specific tasks.
Resources
Base model
ibm-granite/granite-3.1-8b-base