Skip to content

Commit

Permalink
add more stop tokens for qwen and starcoder
Browse files Browse the repository at this point in the history
  • Loading branch information
rjmacarthy committed Sep 25, 2024
1 parent 0dba01f commit 1064518
Show file tree
Hide file tree
Showing 2 changed files with 33 additions and 14 deletions.
21 changes: 18 additions & 3 deletions src/common/constants.ts
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,7 @@ export const EVENT_NAME = {
twinnyStopGeneration: 'twinny-stop-generation',
twinnyStopSymmetryProvider: 'twinny-stop-symmetry-provider',
twinnySymmetryModeles: 'twinny-symmetry-models',
twinnyTextSelection: 'twinny-text-selection',
twinnyTextSelection: 'twinny-text-selection'
}

export const TWINNY_COMMAND_NAME = {
Expand Down Expand Up @@ -156,7 +156,7 @@ export const WORKSPACE_STORAGE_KEY = {
showEmbeddingOptions: 'showEmbeddingOptions',
showProviders: 'showProviders',
reviewOwner: 'reviewOwner',
reviewRepo: 'reviewRepo',
reviewRepo: 'reviewRepo'
}

export const EXTENSION_SETTING_KEY = {
Expand Down Expand Up @@ -220,7 +220,22 @@ export const STOP_DEEPSEEK = [
'<|end▁of▁sentence|>'
]

export const STOP_STARCODER_QWEN = ['<file_sep>', '<fim_prefix>', '<|file_sep|>', '<|fim_prefix|>', '<|endoftext|>']
export const STOP_STARCODER = [
'<|endoftext|>',
'<file_sep>',
'<file_sep>',
'<fim_prefix>',
'<repo_name>',
]

export const STOP_QWEN = [
'<|endoftext|>',
'<|file_sep|>',
'<|fim_prefix|>',
'<|im_end|>',
'<|im_start|>',
'<|repo_name|>',
]

export const STOP_CODEGEMMA = ['<|file_separator|>', '<|end_of_turn|>', '<eos>']

Expand Down
26 changes: 15 additions & 11 deletions src/extension/fim-templates.ts
Original file line number Diff line number Diff line change
@@ -1,10 +1,11 @@
import {
FIM_TEMPLATE_FORMAT,
STOP_DEEPSEEK,
STOP_LLAMA,
STOP_STARCODER_QWEN,
STOP_CODEGEMMA,
STOP_CODESTRAL,
STOP_DEEPSEEK,
STOP_LLAMA,
STOP_QWEN,
STOP_STARCODER,
} from '../common/constants'
import { supportedLanguages } from '../common/languages'
import { FimPromptTemplate } from '../common/types'
Expand Down Expand Up @@ -80,7 +81,7 @@ export const getFimPromptTemplateCodestral = ({
}

export const getFimPromptTemplateQwen = ({
prefixSuffix,
prefixSuffix
}: FimPromptTemplate) => {
const { prefix, suffix } = prefixSuffix
return `<|fim_prefix|>${prefix}<|fim_suffix|>${suffix}<|fim_middle|>`
Expand Down Expand Up @@ -187,10 +188,13 @@ export const getStopWordsAuto = (fimModel: string) => {

if (
fimModel.includes(FIM_TEMPLATE_FORMAT.stableCode) ||
fimModel.includes(FIM_TEMPLATE_FORMAT.starcoder) ||
fimModel.includes(FIM_TEMPLATE_FORMAT.codeqwen)
fimModel.includes(FIM_TEMPLATE_FORMAT.starcoder)
) {
return STOP_STARCODER_QWEN
return STOP_STARCODER
}

if (fimModel.includes(FIM_TEMPLATE_FORMAT.codeqwen)) {
return STOP_QWEN
}

if (fimModel.includes(FIM_TEMPLATE_FORMAT.codegemma)) {
Expand All @@ -207,14 +211,14 @@ export const getStopWordsAuto = (fimModel: string) => {
export const getStopWordsChosen = (format: string) => {
if (format === FIM_TEMPLATE_FORMAT.codellama) return STOP_LLAMA
if (format === FIM_TEMPLATE_FORMAT.deepseek) return STOP_DEEPSEEK
if (format === FIM_TEMPLATE_FORMAT.codeqwen) return STOP_STARCODER_QWEN
if (format === FIM_TEMPLATE_FORMAT.codeqwen) return STOP_QWEN
if (format === FIM_TEMPLATE_FORMAT.codegemma) return STOP_CODEGEMMA
if (format === FIM_TEMPLATE_FORMAT.codestral) return STOP_CODESTRAL
if (
format === FIM_TEMPLATE_FORMAT.stableCode ||
format === FIM_TEMPLATE_FORMAT.starcoder
)
return STOP_STARCODER_QWEN
if (format === FIM_TEMPLATE_FORMAT.codegemma) return STOP_CODEGEMMA
if (format === FIM_TEMPLATE_FORMAT.codestral) return STOP_CODESTRAL
return STOP_STARCODER
return STOP_LLAMA
}

Expand Down

0 comments on commit 1064518

Please sign in to comment.