add fix to HF endpoint

packages/components/nodes/chatmodels/ChatHuggingFace/core.ts

@@ -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 obj: any = {
-        const res = await this.caller.callWithOptions({ signal: options.signal }, hf.textGeneration.bind(hf), {
-            model: this.model,
             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
     }
 

packages/components/nodes/embeddings/HuggingFaceInferenceEmbedding/core.ts

@@ -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(() =>
+        const obj: any = {
-            this.client.featureExtraction({
+            inputs: clean
-                model: this.model,
+        }
-                inputs: clean
+        if (!this.endpoint) obj.model = this.model
-            })
+        return this.caller.call(() => this.client.featureExtraction(obj)) as Promise<number[][]>
-        ) as Promise<number[][]>
     }
 
     embedQuery(document: string): Promise<number[]> {

packages/components/nodes/llms/HuggingFaceInference/core.ts

@@ -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 obj: any = {
-        const res = await this.caller.callWithOptions({ signal: options.signal }, hf.textGeneration.bind(hf), {
-            model: this.model,
             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
     }