Instructions to use brianwoo/Orca-Mini-3B-Onnx-Quantized with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use brianwoo/Orca-Mini-3B-Onnx-Quantized with Transformers:

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

pipe = pipeline("text-generation", model="brianwoo/Orca-Mini-3B-Onnx-Quantized")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("brianwoo/Orca-Mini-3B-Onnx-Quantized")
model = AutoModelForCausalLM.from_pretrained("brianwoo/Orca-Mini-3B-Onnx-Quantized")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use brianwoo/Orca-Mini-3B-Onnx-Quantized with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "brianwoo/Orca-Mini-3B-Onnx-Quantized"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "brianwoo/Orca-Mini-3B-Onnx-Quantized",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/brianwoo/Orca-Mini-3B-Onnx-Quantized

SGLang

How to use brianwoo/Orca-Mini-3B-Onnx-Quantized 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 "brianwoo/Orca-Mini-3B-Onnx-Quantized" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "brianwoo/Orca-Mini-3B-Onnx-Quantized",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "brianwoo/Orca-Mini-3B-Onnx-Quantized" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "brianwoo/Orca-Mini-3B-Onnx-Quantized",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use brianwoo/Orca-Mini-3B-Onnx-Quantized with Docker Model Runner:
```
docker model run hf.co/brianwoo/Orca-Mini-3B-Onnx-Quantized
```

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Have external files can load directly from hub,must down all the yourself . Can run on mobile or mini PC,it's much more faster than the normal version. Sample code for using this model :

from transformers import AutoTokenizer, pipeline
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer,GenerationConfig,TextIteratorStreamer
from optimum.onnxruntime import ORTModelForCausalLM
from transformers import logging
logging.set_verbosity_error()
import time
model_id="pathto/Orca-Mini-3B-Onnx-Quantized"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = ORTModelForCausalLM.from_pretrained(model_id)
streamer = TextStreamer(tokenizer,skip_prompt=True, skip_special_tokens=True,return_text=True)

onnx_gen = pipeline("text-generation", model=model, tokenizer=tokenizer,streamer=streamer,return_text=True)

while True:
    text = input("\nBrian:")
    print("Bot:\n")
    t0=time.time()
    gen = onnx_gen(text)
    t=time.time()-t0
    text=gen[0]["generated_text"]
    print(t,len(text.split(" "))/t,"words /sec")

My sample code for quant modes on Colab :

from optimum.onnxruntime import ORTStableDiffusionXLImg2ImgPipeline , ORTStableDiffusionXLPipeline,ORTModelForCausalLM
from transformers import AutoTokenizer
from optimum.onnxruntime import ORTOptimizer
from optimum.onnxruntime.configuration import OptimizationConfig
from optimum.onnxruntime.configuration import AutoQuantizationConfig
from optimum.onnxruntime import ORTQuantizer
import os,torch
from pathlib import Path
Gbase="/gdrive/MyDrive/onnx/"

model_checkpoint = "gpt2"
save_directory = Gbase+"onnx/gpt2_arm64"

tasks=['TinyLlama/TinyLlama-1.1B-Chat-v0.6',
'pankajmathur/orca_mini_3b',
'Fredithefish/RedPajama-INCITE-Chat-3B-Instruction-Tuning-with-GPT-4',
'CobraMamba/mamba-gpt-3b-v4',
'WizardLM/WizardCoder-3B-V1.0',
'GeneZC/MiniChat-3B']




#TinyLlama/TinyLlama-1.1B-Chat-v0.6
#pankajmathur/orca_mini_3b
#Fredithefish/RedPajama-INCITE-Chat-3B-Instruction-Tuning-with-GPT-4
#CobraMamba/mamba-gpt-3b-v4
#WizardLM/WizardCoder-3B-V1.0
#GeneZC/MiniChat-3B

def quantModel(model_checkpoint=model_checkpoint,save_directory=save_directory):
    tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)
    tokenizer.save_pretrained(save_directory)
    ort_model = ORTModelForCausalLM.from_pretrained(model_checkpoint, export=True)

    optimizer = ORTOptimizer.from_pretrained(ort_model)
    optimization_config = OptimizationConfig(optimization_level=3)

    qconfig = AutoQuantizationConfig.arm64(is_static=False, per_channel=False)

    quantizer = ORTQuantizer.from_pretrained(ort_model)

    quantizer.quantize(save_dir=save_directory, quantization_config=qconfig,use_external_data_format=True)

def  doTasks(tasks=tasks,Gbase=Gbase):
    for model_checkpoint in tasks:
        save_directory=os.path.join(Gbase,Path(model_checkpoint).name)
        try:quantModel(model_checkpoint=model_checkpoint,save_directory=save_directory)
        except:
            import traceback
            traceback.print_exc()


doTasks()

Downloads last month: -