PyPI page
Home page
Author:
icnatspell
Summary:
Retrieval-based speculative decoding for ONNX Runtime and Hugging Face transformers
Latest version:
0.2.0
Required dependencies:
huggingface-hub
|
loguru
|
numpy
|
onnxruntime
|
pydantic
|
rich
|
tqdm
|
transformers
|
typer
Optional dependencies:
datasets
|
dejavuu
|
onnx
|
onnx-ir
|
onnxruntime-genai
|
onnxscript
|
pillow
|
torch
|
torchvision
Downloads last day:
1
Downloads last week:
2
Downloads last month:
20
7-day smoothing