changed - `LLM.prompt`, `LLM.ask`, and `LLM.chat` now accept extra `**kwargs` that are sent directly to model (54)
fixed: - N/A
0.0.35
new: - N/A
changed - Updates for `langchain>=0.1.0` (which is now minimum version)
fixed: - N/A
0.0.34
new: - Use [Zephyr-7B](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF) as default model in `webapp.yml`. (#52)
changed - Added `stop` parameter to `LLM.prompt` (overrides `stop` parameter supplied to constructor) (53)
fixed: - N/A
0.0.33
new: - N/A
changed - Added `prompt_template` parameter to `LLM` constructor (51) - Added `update_max_tokens` and `update_stop` methods to `LLM` for dynamic adjustments during prompt experiments
fixed: - Explicitly set `offload_kqv` to ensure GPUs are fully utilized (50)