Instructions to use OpceanAI/qwenti-inference with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpceanAI/qwenti-inference with Transformers:

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

pipe = pipeline("text-generation", model="OpceanAI/qwenti-inference")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OpceanAI/qwenti-inference")
model = AutoModelForCausalLM.from_pretrained("OpceanAI/qwenti-inference", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OpceanAI/qwenti-inference with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/OpceanAI/qwenti-inference

SGLang

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

Docker Model Runner
How to use OpceanAI/qwenti-inference with Docker Model Runner:
```
docker model run hf.co/OpceanAI/qwenti-inference
```

qwenti-inference / scripts /run_tl4.py

OpceanAI

Model card, paper, figures, logs and scripts

2049fc8 verified 1 day ago

Raw History Blame Contribute Delete

578 Bytes

import os

os.environ.setdefault("CHIPFORMER_WROM_HEADS", "320")
import sys
import time

sys.path.insert(0, "/tmp/opencode/qp/experimental/riscv")
from chipformer.rv import elf, ref

m = ref.Machine.from_program(elf.program(sys.argv[1]))
n = int(sys.argv[2])
t0 = time.time()
ntok = 0
while not m.halted:
    ntok += len(m.record())
d = lambda i: ref.sx(m.ram[0xE000 // 4 + i], 32)
print("instrucciones", m.steps, "trace tokens", ntok, f"{time.time() - t0:.0f}s")
print("kernel tokens :", [d(100 + p) for p in range(n)])
print("kernel logits :", [d(200 + p) for p in range(n)])

	import os

	os.environ.setdefault("CHIPFORMER_WROM_HEADS", "320")
	import sys
	import time

	sys.path.insert(0, "/tmp/opencode/qp/experimental/riscv")
	from chipformer.rv import elf, ref

	m = ref.Machine.from_program(elf.program(sys.argv[1]))
	n = int(sys.argv[2])
	t0 = time.time()
	ntok = 0
	while not m.halted:
	ntok += len(m.record())
	d = lambda i: ref.sx(m.ram[0xE000 // 4 + i], 32)
	print("instrucciones", m.steps, "trace tokens", ntok, f"{time.time() - t0:.0f}s")
	print("kernel tokens :", [d(100 + p) for p in range(n)])
	print("kernel logits :", [d(200 + p) for p in range(n)])