Add top n sigma sampler and other webui fix (#512)

Co-authored-by: firecoperana <firecoperana>
This commit is contained in:
firecoperana
2025-06-12 00:19:26 -05:00
committed by GitHub
parent baa412aa42
commit 07777dde1f
8 changed files with 36 additions and 26 deletions

View File

@@ -17,7 +17,7 @@ export const CONFIG_DEFAULT = {
showThoughtInProgress: false,
excludeThoughtOnReq: true,
// make sure these default values are in sync with `common.h`
samplers: 'dkypmxt',
samplers: 'dkypmxnt',
temperature: 0.8,
dynatemp_range: 0.0,
dynatemp_exponent: 1.0,
@@ -26,6 +26,7 @@ export const CONFIG_DEFAULT = {
min_p: 0.05,
xtc_probability: 0.0,
xtc_threshold: 0.1,
top_n_sigma: 0.0,
typical_p: 1.0,
repeat_last_n: 64,
repeat_penalty: 1.0,
@@ -44,7 +45,7 @@ export const CONFIG_INFO: Record<string, string> = {
apiKey: 'Set the API Key if you are using --api-key option for the server.',
systemMessage: 'The starting message that defines how model should behave.',
samplers:
'The order at which samplers are applied, in simplified way. Default is "dkypmxt": dry->top_k->typ_p->top_p->min_p->xtc->temperature',
'The order at which samplers are applied, in simplified way. Default is "dkypmxt": dry->top_k->typ_p->top_p->min_p->xtc->top_sigma->temperature',
temperature:
'Controls the randomness of the generated text by affecting the probability distribution of the output tokens. Higher = more random, lower = more focused.',
dynatemp_range:
@@ -60,6 +61,8 @@ export const CONFIG_INFO: Record<string, string> = {
'XTC sampler cuts out top tokens; this parameter controls the chance of cutting tokens at all. 0 disables XTC.',
xtc_threshold:
'XTC sampler cuts out top tokens; this parameter controls the token probability that is required to cut that token.',
top_n_sigma:
'Top-n-sigma sampling filters out low-value tokens by discarding tokens that fall more than n standard deviations below the maximum probability',
typical_p:
'Sorts and limits tokens based on the difference between log-probability and entropy.',
repeat_last_n: 'Last n tokens to consider for penalizing repetition',

View File

@@ -70,7 +70,7 @@ export default function Header() {
</svg>
</label>
<div className="grow text-2xl font-bold ml-2">llama.cpp</div>
<div className="grow text-2xl font-bold ml-2">ik_llama.cpp</div>
{/* action buttons (top right) */}
<div className="flex items-center">

View File

@@ -29,6 +29,7 @@ const SAMPLER_KEYS: SettKey[] = [
'typical_p',
'xtc_probability',
'xtc_threshold',
'top_n_sigma'
];
const PENALTY_KEYS: SettKey[] = [
'repeat_last_n',
@@ -196,7 +197,7 @@ const SETTING_SECTIONS: SettingSection[] = [
label: (
<>
Custom JSON config (For more info, refer to{' '}
<OpenInNewTab href="https://github.com/ggerganov/llama.cpp/blob/master/examples/server/README.md">
<OpenInNewTab href="https://github.com/ikawrakow/ik_llama.cpp/tree/main/examples/server/README.md">
server documentation
</OpenInNewTab>
)
@@ -224,7 +225,7 @@ const SETTING_SECTIONS: SettingSection[] = [
<br />
<br />
If you encounter any problems, create a{' '}
<OpenInNewTab href="https://github.com/ggerganov/llama.cpp/issues/new?template=019-bug-misc.yml">
<OpenInNewTab href="https://github.com/ikawrakow/ik_llama.cpp/issues/new?template=019-bug-misc.yml">
Bug (misc.)
</OpenInNewTab>{' '}
report on Github. Please also specify <b>webui/experimental</b> on
@@ -456,11 +457,11 @@ function SettingsModalLongInput({
<div className="label inline">{label || configKey}</div>
<textarea
className="textarea textarea-bordered h-24"
placeholder={`Default: ${CONFIG_DEFAULT[configKey] || 'none'}`}
value={value}
onChange={(e) => onChange(e.target.value)}
/>
</label>
placeholder={`Default: ${CONFIG_DEFAULT[configKey] || 'none'}`}
value={value}
onChange={(e) => onChange(e.target.value)}
/>
</label>
);
}
@@ -485,7 +486,7 @@ function SettingsModalShortInput({
<div className="block md:hidden mb-1">
<b>{label || configKey}</b>
<br />
<p className="text-xs">{helpMsg}</p>
<p className="text-xs whitespace-normal">{helpMsg}</p>
</div>
)}
<label className="input input-bordered join-item grow flex items-center gap-2 mb-2">
@@ -494,7 +495,7 @@ function SettingsModalShortInput({
{label || configKey}
</div>
{helpMsg && (
<div className="dropdown-content menu bg-base-100 rounded-box z-10 w-64 p-2 shadow mt-4">
<div className="dropdown-content menu bg-base-100 rounded-box z-10 w-64 p-2 shadow mt-4 whitespace-normal break-words">
{helpMsg}
</div>
)}

View File

@@ -200,6 +200,7 @@ export const AppContextProvider = ({
typical_p: config.typical_p,
xtc_probability: config.xtc_probability,
xtc_threshold: config.xtc_threshold,
top_n_sigma: config.top_n_sigma,
repeat_last_n: config.repeat_last_n,
repeat_penalty: config.repeat_penalty,
presence_penalty: config.presence_penalty,