PyPI page
Home page
Summary:
Find and prove inference savings for agent workloads on your own GPUs: $/M tokens with 95% CIs, calibrated verdicts and verified-savings statements
Latest version:
0.6.2
Required dependencies:
fastapi
|
httpx
|
uvicorn
Optional dependencies:
guidellm
|
huggingface-hub
|
numpy
|
onnxruntime
|
pyyaml
|
tokenizers
Downloads last day:
6
Downloads last week:
135
Downloads last month:
330
7-day smoothing