v0.4.1

• Fix minor bugs

v0.4.0

• Support vision model

v0.3.4

• Use vllm as the backend
• Make sure using local snapshots

v0.3.3

• Add %config show magic to show current configurations
• Improve help message

v0.3.2

• New config options: n_predict and n_new_tokens

v0.3.1

Fix bugs

v0.3.0

• Support %help and %config magics.

v0.2.1

• The first release which can be properly installed by pip
• Support %load, %new_chat, %hf_home, %model_list magics
• Support loading meta-llama/Meta-Llama-3-8B-Instruct, mistralai/Mistral-7B-Instruct-v0.3, unsloth/llama-3-8b-Instruct-bnb-4bit and similar models