Instructions to use RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8 with Transformers:

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

pipe = pipeline("text-generation", model="RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8", trust_remote_code=True)
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

vLLM

How to use RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8

SGLang

How to use RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8 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 "RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8" \
    --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": "RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8",
		"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 "RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8" \
        --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": "RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8 with Docker Model Runner:
```
docker model run hf.co/RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8
```

Model doesnt load in vllm version 0.5.0 : ValueError: `rope_scaling`'s type field must be one of ['su', 'yarn'], got longrope

by sujjosep - opened Jul 20, 2024

Discussion

sujjosep

Jul 20, 2024

llm = LLM(model=model_id, trust_remote_code=True, max_model_len=8196, tensor_parallel_size=number_gpus)

With trust remote code True

File ~/.local/lib/python3.11/site-packages/vllm/config.py:1287, in _get_and_verify_max_len(hf_config, max_model_len, disable_sliding_window, sliding_window_len)
   1280 if disable_sliding_window:
   1281     # TODO(robertgshaw): Find a model that supports rope_scaling
   1282     # with sliding window to see if this case should be allowed.
   1283     raise NotImplementedError(
   1284         "Disabling sliding window is not supported for models "
   1285         "with rope_scaling. Please raise an issue so we can "
   1286         "investigate.")
-> 1287 assert "factor" in rope_scaling
   1288 scaling_factor = rope_scaling["factor"]
   1289 if rope_scaling["type"] == "yarn":

AssertionError:

Without trust remote code True :

File ~/.local/lib/python3.11/site-packages/transformers/models/phi3/configuration_phi3.py:185, in Phi3Config._rope_scaling_validation(self)
    183 rope_scaling_long_factor = self.rope_scaling.get("long_factor", None)
    184 if rope_scaling_type is None or rope_scaling_type not in ["su", "yarn"]:
--> 185     raise ValueError(f"`rope_scaling`'s type field must be one of ['su', 'yarn'], got {rope_scaling_type}")
    186 if not (
    187     isinstance(rope_scaling_short_factor, list)
    188     and all(isinstance(x, (int, float)) for x in rope_scaling_short_factor)
    189 ):
    190     raise ValueError(
    191         f"`rope_scaling`'s short_factor field must be a list of numbers, got {rope_scaling_short_factor}"
    192     )

ValueError: `rope_scaling`'s type field must be one of ['su', 'yarn'], got longrope

mgoin

Red Hat AI org Jul 20, 2024

Please update your version of vLLM to 0.5.2

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