Update docs

This commit is contained in:
oobabooga 2023-07-31 19:20:33 -07:00
parent f094330df0
commit b2207f123b
3 changed files with 0 additions and 38 deletions

View File

@ -1,35 +0,0 @@
# Generation parameters
For a description of the generation parameters provided by the transformers library, see this link: https://huggingface.co/docs/transformers/main_classes/text_generation#transformers.GenerationConfig
### llama.cpp
llama.cpp only uses the following parameters:
* temperature
* top_p
* top_k
* repetition_penalty
* tfs
* mirostat_mode
* mirostat_tau
* mirostat_eta
### ExLlama
ExLlama only uses the following parameters:
* temperature
* top_p
* top_k
* repetition_penalty
* repetition_penalty_range
* typical_p
### RWKV
RWKV only uses the following parameters when loaded through the old .pth weights:
* temperature
* top_p
* top_k

View File

@ -8,7 +8,6 @@
* [Docker](Docker.md)
* [ExLlama](ExLlama.md)
* [Extensions](Extensions.md)
* [Generation parameters](Generation-parameters.md)
* [GPTQ models (4 bit mode)](GPTQ-models-(4-bit-mode).md)
* [LLaMA model](LLaMA-model.md)
* [llama.cpp](llama.cpp.md)

View File

@ -357,8 +357,6 @@ def create_settings_menus(default_preset):
with gr.Accordion("Learn more", open=False):
gr.Markdown("""
Not all parameters are used by all loaders. See [this page](https://github.com/oobabooga/text-generation-webui/blob/main/docs/Generation-parameters.md) for details.
For a technical description of the parameters, the [transformers documentation](https://huggingface.co/docs/transformers/main_classes/text_generation#transformers.GenerationConfig) is a good reference.
The best presets, according to the [Preset Arena](https://github.com/oobabooga/oobabooga.github.io/blob/main/arena/results.md) experiment, are: