mmarch/text-generation-webui
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

dynatemp_low, dynatemp_high, dynatemp_exponent parameters (#5209)

Browse source
This commit is contained in:
oobabooga 2024-01-08 23:28:35 -03:00 • committed by GitHub
parent dc1df22a2b
commit 29c2693ea0
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
9 changed files with 51 additions and 27 deletions
Download patch file Download diff file Expand all files Collapse all files

3
docs/03 - Parameters Tab.md
View file

@ -54,8 +54,7 @@ For more information about the parameters, the [transformers documentation](http
* **mirostat_mode**: Activates the Mirostat sampling technique. It aims to control perplexity during sampling. See the [paper](https://arxiv.org/abs/2007.14966).
* **mirostat_tau**: No idea, see the paper for details. According to the Preset Arena, 8 is a good value.
* **mirostat_eta**: No idea, see the paper for details. According to the Preset Arena, 0.1 is a good value.
* **dynamic_temperature_low**: The lower bound for temperature in Dynamic Temperature. Only used when "dynamic_temperature" is checked.
* **dynamic_temperature**: Activates Dynamic Temperature. This modifies temperature to range between "dynamic_temperature_low" (minimum) and "temperature" (maximum), with an entropy-based scaling.
* **dynamic_temperature**: Activates Dynamic Temperature. This modifies temperature to range between "dynatemp_low" (minimum) and "dynatemp_high" (maximum), with an entropy-based scaling. The steepness of the curve is controlled by "dynatemp_exponent".
* **temperature_last**: Makes temperature the last sampler instead of the first. With this, you can remove low probability tokens with a sampler like min_p and then use a high temperature to make the model creative without losing coherency.
* **do_sample**: When unchecked, sampling is entirely disabled, and greedy decoding is used instead (the most likely token is always picked).
* **Seed**: Set the Pytorch seed to this number. Note that some loaders do not use Pytorch (notably llama.cpp), and others are not deterministic (notably ExLlama v1 and v2). For these loaders, the seed has no effect.