Instructions to use ibm-granite/granite-8b-code-instruct-4k with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ibm-granite/granite-8b-code-instruct-4k with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ibm-granite/granite-8b-code-instruct-4k")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ibm-granite/granite-8b-code-instruct-4k")
model = AutoModelForCausalLM.from_pretrained("ibm-granite/granite-8b-code-instruct-4k")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ibm-granite/granite-8b-code-instruct-4k with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ibm-granite/granite-8b-code-instruct-4k"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-8b-code-instruct-4k",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ibm-granite/granite-8b-code-instruct-4k

SGLang

How to use ibm-granite/granite-8b-code-instruct-4k with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ibm-granite/granite-8b-code-instruct-4k" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-8b-code-instruct-4k",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ibm-granite/granite-8b-code-instruct-4k" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-8b-code-instruct-4k",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ibm-granite/granite-8b-code-instruct-4k with Docker Model Runner:
```
docker model run hf.co/ibm-granite/granite-8b-code-instruct-4k
```

Response is not good as expected

by skumarai - opened May 14, 2024

Discussion

skumarai

May 14, 2024

I have tried both 3b and 8b models and getting below response.
I copied same code from model card and tested it. can anyone help why i am getting bad response .

3b Model response

This IS expected if you are initializing LlamaForCausalLM from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
This IS NOT expected if you are initializing LlamaForCausalLM from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
Question:
Write a code to find the maximum value in a list of numbers.

Answer:

def find_largest_even_odd_elements_list(arr):
    even_sum = even_list = odd_sum =0
        for i in range(0n2):
            if(arr[i)%2 ==0 and i<3:\
                even_list.append(arr[i])
            elif(arr[i)%2!=0 and odd_list.count(1)>=1)::\:\s.rindex(max_

8b model response
---------------------------
- This IS expected if you are initializing LlamaForCausalLM from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
- This IS NOT expected if you are initializing LlamaForCausalLM from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
Question:
Write a code to find the maximum value in a list of numbers.

Answer:
[jButton,jButton,njButton24[0jButton,0 of <<reement<,,,,,0 of
0":,,0,NuDreement,,,0 of
njButton,0 <<, <<,":": I3:":":000 of1 <<<<EFlags <<, <<, <<jButton",,<<,0 <<",1, of1 <<":":"<<_3 << <<,18.2 <<,1,<<

mayank-mishra

May 14, 2024

you need to install HF transformers from source for it to work correctly @skumarai .
Some changes required to run our models are currently not in the release and with the next release, a pip install should work.

Relevant PR: https://github.com/huggingface/transformers/pull/30031
git clone https://github.com/huggingface/transformers
cd transformers/
pip install ./
cd ..

skumarai

May 14, 2024

Thank you for quick response @mayank-mishra .

sure, I will try to update transformers to reflect above mentioned PR changes and test the model.

mayank-mishra

May 14, 2024

@skumarai were you able to get good outputs?
can we close this issue?

skumarai

May 15, 2024

Hi @mayank-mishra
yes, I got the good results.
we can close this.
Thanks.

skumarai changed discussion status to closed May 15, 2024

mayank-mishra

May 15, 2024

awesome.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment