Instructions to use sjster/test_medium with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sjster/test_medium with Transformers:

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

pipe = pipeline("text-generation", model="sjster/test_medium", 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("sjster/test_medium", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("sjster/test_medium", 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 Settings

vLLM

How to use sjster/test_medium with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/sjster/test_medium

SGLang

How to use sjster/test_medium 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 "sjster/test_medium" \
    --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": "sjster/test_medium",
		"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 "sjster/test_medium" \
        --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": "sjster/test_medium",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use sjster/test_medium with Docker Model Runner:
```
docker model run hf.co/sjster/test_medium
```

test_medium / handler.py

Srijith Rajamohan

Added model inference code

1dcfe80 almost 2 years ago

Raw

History Blame Contribute Delete

2.43 kB

	from typing import Dict, List, Any
	from transformers import (
	AutoModelForCausalLM,
	AutoTokenizer)
	import torch

	model = AutoModelForCausalLM.from_pretrained(
	"sjster/test_medium",
	trust_remote_code=True,
	quantization_config=None,
	torch_dtype=torch.float, # data type is float
	device_map="auto",
	)

	class EndpointHandler():
	def __init__(self, path=""):
	# Preload all the elements you are going to need at inference.
	self.model = AutoModelForCausalLM.from_pretrained(
	path,
	trust_remote_code=True,
	quantization_config=None,
	torch_dtype=torch.float, # data type is float
	device_map="auto")
	self.tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
	self.tokenizer.padding_side = "left"
	self.tokenizer.pad_token = self.tokenizer.eos_token
	self.tokenizer.add_eos_token = True

	def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]:
	"""
	data args:
	inputs (:obj: `str` \| `PIL.Image` \| `np.array`)
	kwargs
	Return:
	A :obj:`list` \| `dict`: will be serialized and returned
	"""

	inputs = data.pop("inputs", data)
	messages = [
	{
	"role": "user",
	"content": ""
	+ inputs,
	},
	]
	encodeds = self.tokenizer.apply_chat_template(messages, return_tensors="pt")
	encoded_length = len(encodeds[0])
	model_inputs = encodeds.to('cuda')
	result = self.model.generate(model_inputs,
	do_sample=False,
	output_scores=True,
	return_dict_in_generate=True,
	output_attentions=True,
	output_hidden_states=True,
	#num_beams=3,
	#no_repeat_ngram_size=1,
	early_stopping = True,
	#top_k=0,
	max_new_tokens=400)
	x, logits_gen = result.sequences, result.scores
	x = x[:,encoded_length:]
	decoded = self.tokenizer.batch_decode(x)

	return [{"outputs": decoded[0]}]