fix

src/app/queries/predictNextPanels.ts

@@ -61,19 +61,21 @@ export const predictNextPanels = async ({
   try {
     // console.log(`calling predict(${query}, ${nbTotalPanels})`)
     result = `${await predict(query, nbPanelsToGenerate) || ""}`.trim()
+    console.log("LLM result (1st trial):", result)
     if (!result.length) {
-      throw new Error("empty result!")
+      throw new Error("empty result on 1st trial!")
     }
   } catch (err) {
     // console.log(`prediction of the story failed, trying again..`)
     try {
-      result = `${await predict(query+".", nbPanelsToGenerate) || ""}`.trim()
+      result = `${await predict(query + " \n ", nbPanelsToGenerate) || ""}`.trim()
+      console.log("LLM result (2nd trial):", result)
       if (!result.length) {
-        throw new Error("empty result!")
+        throw new Error("empty result on 2nd trial!")
       }
     } catch (err) {
-      console.error(`prediction of the story failed again 💩`)
-      throw new Error(`failed to generate the story ${err}`)
+      console.error(`prediction of the story failed twice 💩`)
+      throw new Error(`failed to generate the story twice 💩 ${err}`)
     }
   }
 

src/app/queries/predictWithHuggingFace.ts

@@ -14,7 +14,7 @@ export async function predict(inputs: string, nbPanels: number): Promise<string>
 
   // we don't require a lot of token for our task
   // but to be safe, let's count ~110 tokens per panel
-  const nbMaxNewTokens = nbPanels * 110
+  const nbMaxNewTokens = nbPanels * 130 // 110 isn't enough anymore for long dialogues
 
   switch (llmEngine) {
     case "INFERENCE_ENDPOINT":