Instructions to use microsoft/Phi-3-small-128k-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use microsoft/Phi-3-small-128k-instruct with Transformers:

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

pipe = pipeline("text-generation", model="microsoft/Phi-3-small-128k-instruct", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3-small-128k-instruct", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use microsoft/Phi-3-small-128k-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "microsoft/Phi-3-small-128k-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "microsoft/Phi-3-small-128k-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/microsoft/Phi-3-small-128k-instruct

SGLang

How to use microsoft/Phi-3-small-128k-instruct 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 "microsoft/Phi-3-small-128k-instruct" \
    --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": "microsoft/Phi-3-small-128k-instruct",
		"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 "microsoft/Phi-3-small-128k-instruct" \
        --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": "microsoft/Phi-3-small-128k-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use microsoft/Phi-3-small-128k-instruct with Docker Model Runner:
```
docker model run hf.co/microsoft/Phi-3-small-128k-instruct
```

Out of resource: shared memory

#16

by iszhaoxin - opened Jun 8, 2024

Discussion

iszhaoxin

Jun 8, 2024

•

edited Jun 8, 2024

I got the error message as below:

'triton.runtime.autotuner.OutOfResources: out of resource: shared memory, Required: 135200, Hardware limit: 101376. Reducing block sizes or num_stages may help.'

I tried on both RTX A6000 and RTX 6000.
I guess maybe it is because the model is only trained and tested on specific types GPUs, such as A100?

Satandon1999

Jun 10, 2024

Yes, in my experience as well this model works well only on the GPUs listed as 'tested' in the documentation.

LeeStott

Jul 19, 2024

The recommended adjustment layer is

"target_modules": [
"o_proj",
"qkv_proj"
]

joker26

Jul 21, 2024

@LeeStott how to achieve

ProfLinh

Sep 23, 2024

@LeeStott I'm running into this error as well. Can you show us how to adjust the layer?

Satandon1999

Sep 24, 2024

If using PEFT you can set this using the "target_modules" parameter in LoraConfig
https://huggingface.co/docs/peft/package_reference/lora#peft.LoraConfig

moidhassan

Sep 29, 2024

@LeeStott ,
Using these target modules with this model give the following error -
ValueError: Target modules {'o_proj', 'qkv_proj'} not found in the base model. Please check the target modules and try again.

I tried other target modules such as "all-linear" and [q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj,lm_head] but both these options are giving the same error which is mentioned in the main question of this thread.

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