add fix to HF endpoint
This commit is contained in:
parent
6a7e378baf
commit
3cda348e95
|
|
@ -78,9 +78,7 @@ export class HuggingFaceInference extends LLM implements HFInput {
|
|||
async _call(prompt: string, options: this['ParsedCallOptions']): Promise<string> {
|
||||
const { HfInference } = await HuggingFaceInference.imports()
|
||||
const hf = new HfInference(this.apiKey)
|
||||
if (this.endpoint) hf.endpoint(this.endpoint)
|
||||
const res = await this.caller.callWithOptions({ signal: options.signal }, hf.textGeneration.bind(hf), {
|
||||
model: this.model,
|
||||
const obj: any = {
|
||||
parameters: {
|
||||
// make it behave similar to openai, returning only the generated text
|
||||
return_full_text: false,
|
||||
|
|
@ -91,7 +89,13 @@ export class HuggingFaceInference extends LLM implements HFInput {
|
|||
repetition_penalty: this.frequencyPenalty
|
||||
},
|
||||
inputs: prompt
|
||||
})
|
||||
}
|
||||
if (this.endpoint) {
|
||||
hf.endpoint(this.endpoint)
|
||||
} else {
|
||||
obj.model = this.model
|
||||
}
|
||||
const res = await this.caller.callWithOptions({ signal: options.signal }, hf.textGeneration.bind(hf), obj)
|
||||
return res.generated_text
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -30,12 +30,11 @@ export class HuggingFaceInferenceEmbeddings extends Embeddings implements Huggin
|
|||
async _embed(texts: string[]): Promise<number[][]> {
|
||||
// replace newlines, which can negatively affect performance.
|
||||
const clean = texts.map((text) => text.replace(/\n/g, ' '))
|
||||
return this.caller.call(() =>
|
||||
this.client.featureExtraction({
|
||||
model: this.model,
|
||||
inputs: clean
|
||||
})
|
||||
) as Promise<number[][]>
|
||||
const obj: any = {
|
||||
inputs: clean
|
||||
}
|
||||
if (!this.endpoint) obj.model = this.model
|
||||
return this.caller.call(() => this.client.featureExtraction(obj)) as Promise<number[][]>
|
||||
}
|
||||
|
||||
embedQuery(document: string): Promise<number[]> {
|
||||
|
|
|
|||
|
|
@ -78,9 +78,7 @@ export class HuggingFaceInference extends LLM implements HFInput {
|
|||
async _call(prompt: string, options: this['ParsedCallOptions']): Promise<string> {
|
||||
const { HfInference } = await HuggingFaceInference.imports()
|
||||
const hf = new HfInference(this.apiKey)
|
||||
if (this.endpoint) hf.endpoint(this.endpoint)
|
||||
const res = await this.caller.callWithOptions({ signal: options.signal }, hf.textGeneration.bind(hf), {
|
||||
model: this.model,
|
||||
const obj: any = {
|
||||
parameters: {
|
||||
// make it behave similar to openai, returning only the generated text
|
||||
return_full_text: false,
|
||||
|
|
@ -91,7 +89,13 @@ export class HuggingFaceInference extends LLM implements HFInput {
|
|||
repetition_penalty: this.frequencyPenalty
|
||||
},
|
||||
inputs: prompt
|
||||
})
|
||||
}
|
||||
if (this.endpoint) {
|
||||
hf.endpoint(this.endpoint)
|
||||
} else {
|
||||
obj.model = this.model
|
||||
}
|
||||
const res = await this.caller.callWithOptions({ signal: options.signal }, hf.textGeneration.bind(hf), obj)
|
||||
return res.generated_text
|
||||
}
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue