Run generative AI models with ONNX Runtime.

This library provides the generative AI loop for ONNX models run with ONNX Runtime, including logits processing, search and sampling, and KV cache management.

Users can call a high level generate() method, or provide their own customizations of the loop.

• Search techniques like greedy/beam search to generate token sequences
• Built in scoring tools like repetition penalties
• Easy custom scoring

std::vector<int32_t> input_ids{0, 0, 0, 52, 0, 0, 195, 731};
 
Generators::Model model(*ort_env, "models/gpt2_fp32.onnx");

Generators::SearchParams params{model};
params.batch_size = 2;
params.sequence_length = 4;
params.input_ids = input_ids;
params.max_length = max_length;
params.num_beams = 4;

auto search = params.CreateSearch();
auto state = model.CreateState{search->GetSequenceLengths(), params};

while (!search->IsDone()) {
  search->SetLogits(state->Run(search.GetNextTokens(), search.GetNextIndices(), search.GetSequenceLength());

  // Scoring
  search->Apply_MinLength(5);
  search->Apply_RepetitionPenalty(1.1f);

  search->SelectTop();
}

// Access resulting sequences of tokens
for(unsigned i=0;i<params.batch_size;i++) {
  auto result=search.GetSequence(0);
}

import onnxruntime_genai as og
import numpy as np
from transformers import GPT2Tokenizer

text = "The best hotel in bay area"

# Generate input tokens from the text prompt
tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
input_tokens = tokenizer.encode(text, return_tensors='np')

model=og.Model("../../python/onnx_models", og.DeviceType.CUDA)

params=og.SearchParams(model)
params.max_length = 64
params.input_ids = input_tokens

search=params.CreateSearch()
state=model.CreateState(model, search.GetSequenceLengths(), params)

print("Inputs:")
print(input_tokens)
print("Input prompt:", text)

print("Running greedy search loop...")
while not search.IsDone():
  search.SetLogits(state.Run(search.GetNextTokens(), search.GetSequenceLength())
  search.SelectTop();

print("Output:")
output_tokens=search.GetSequence(0).GetArray()
decoded_output=tokenizer.decode(output_tokens)
print(decoded_output)

• Built in Model Support:
  • GPT2
  • Llama2
• CPU & CUDA
• Beam & Greedy Searches
• C++ static library
• Python Bindings

• Make model code stateless, move state into search? This would allow for multiple searches with one model loaded
• Support more models built-in, T5/Whisper/Llama
• Tokenizer?

• Copy onnxruntime library into the ort/ folder
  • Can either build Onnxruntime from source in release mode, then copy the files specified in install_ort.bat
  • Or download a release from https://github.com/microsoft/onnxruntime/releases
  • Files in ort\ should be:
    • onnxruntime.dll
    • onnxruntime.lib
    • onnxruntime_providers_shared.dll (if using cuda)
    • onnxruntime_providers_cuda.dll (if using cuda)
    • onnxruntime_c_api.h
• Run the build.bat script to generate build files
• Open build\Generators.sln in visual studio

To run the python scripts, use PYTHONPATH: set PYTHONPATH=/path/to/onnxruntime-genai/build/Release/

• Copy onnxruntime library into the ort/ folder
  • Can either build Onnxruntime from source in release mode, then copy the files specified in install_ort.sh
  • Or download a release from https://github.com/microsoft/onnxruntime/releases
  • Files in ort\ should be:
    • libonnxruntime.so
    • libonnxruntime.so.(version #)
    • libonnxruntime_providers_shared.so (if using cuda)
    • libonnxruntime_providers_cuda.so (if using cuda)
    • onnxruntime_c_api.h
• Run the build.sh script to build

To run the python scripts, use PYTHONPATH: export PYTHONPATH=/path/to/onnxruntime-genai/build/

• Onnxruntime
• cmake

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.