v0.4.6

Fix [WinError 10049] by using ollama.Client and client.pull

v0.4.5

• upgrade h11 to 0.16.0
• upgrade setuptools to 78.1.1

v0.4.4

Add in users' provided own custom benchmark models selection

v0.4.2

make the ollama pull models consistent with README.md for v0.4.x

v0.4.1

Add in deepseek-r1 variations for a variety of ram sizes.
It' because deepseek-r1 models are good at the followings in Feb. 2025.

• Instruction-Following Tasks:
• Question-Answering:
• Code Generation and Debugging
  Add in ram-size-32gb to default running phi4:14b, deepseek-r1:14b, deepseek-r1:32b

v0.3.31

Fix passing None to ollama-bin on run_benchmark with 4 arguments in main function

v0.3.30

Get multiple gpu names in Linux if there are more than 1

v0.3.29

dependency
requests==2.32.3
setuptools | Upgrade to~> 70.0.0

v0.3.27

Add in phi4:14b

v0.3.22

• Update model to Llama3.1:8b