Rm jsonMode

src/lib/components/InferencePlayground/InferencePlayground.svelte

@@ -21,7 +21,7 @@
 		{
 			id: String(Math.random()),
 			model: '01-ai/Yi-1.5-34B-Chat',
-			config: { temperature: 0.5, maxTokens: 2048, streaming: true, jsonMode: false },
+			config: { temperature: 0.5, maxTokens: 2048, streaming: true },
 			messages: startMessages
 		}
 	];
@@ -128,7 +128,6 @@
 				requestMessages,
 				conversation.config.temperature,
 				conversation.config.maxTokens,
-				conversation.config.jsonMode,
 				(content) => {
 					if (streamingMessage) {
 						streamingMessage.content = content;
@@ -146,7 +145,6 @@
 				requestMessages,
 				conversation.config.temperature,
 				conversation.config.maxTokens,
-				conversation.config.jsonMode
 			);
 			// check if the user did not abort the request
 			if (waitForNonStreaming) {
@@ -386,7 +384,7 @@
 								{
 									id: String(Math.random()),
 									model: e.target.value,
-									config: { temperature: 0.5, maxTokens: 2048, streaming: true, jsonMode: false },
+									config: { temperature: 0.5, maxTokens: 2048, streaming: true },
 									messages: [...conversations[0].messages]
 								}
 							];

src/lib/components/InferencePlayground/InferencePlaygroundGenerationConfig.svelte

@@ -1,9 +1,4 @@
 <script lang="ts">
-	// export let temperature = 0.5;
-	// export let maxTokens = 2048;
-	// export let streaming = true;
-	// export let jsonMode = true;
-
 	export let config;
 	export let classNames = '';
 </script>

src/lib/components/InferencePlayground/inferencePlaygroundUtils.ts

@@ -18,7 +18,6 @@ export async function handleStreamingResponse(
 	messages: ChatCompletionInputMessage[],
 	temperature: number,
 	maxTokens: number,
-	jsonMode: boolean,
 	onChunk: (content: string) => void,
 	abortController: AbortController
 ): Promise<void> {
@@ -30,7 +29,6 @@ export async function handleStreamingResponse(
 				messages: messages,
 				temperature: temperature,
 				max_tokens: maxTokens,
-				json_mode: jsonMode
 			},
 			{ signal: abortController.signal }
 		)) {
@@ -54,14 +52,12 @@ export async function handleNonStreamingResponse(
 	messages: ChatCompletionInputMessage[],
 	temperature: number,
 	maxTokens: number,
-	jsonMode: boolean
 ): Promise<ChatCompletionInputMessage> {
 	const response = await hf.chatCompletion({
 		model: model,
 		messages: messages,
 		temperature: temperature,
 		max_tokens: maxTokens,
-		json_mode: jsonMode
 	});
 
 	if (response.choices && response.choices.length > 0) {

src/lib/types/index.d.ts

@@ -7,7 +7,6 @@ type GenerationConfig = {
 	temperature: number;
 	maxTokens: number;
 	streaming: boolean;
-	jsonMode: boolean;
 };
 
 type Conversation = {