@huggingface/tasks 0.13.4 → 0.13.5

package/dist/esm/local-apps.spec.js

@@ -0,0 +1,112 @@
+import { describe, expect, it } from "vitest";
+import { LOCAL_APPS } from "./local-apps.js";
+describe("local-apps", () => {
+    it("llama.cpp conversational", async () => {
+        const { snippet: snippetFunc } = LOCAL_APPS["llama.cpp"];
+        const model = {
+            id: "bartowski/Llama-3.2-3B-Instruct-GGUF",
+            tags: ["conversational"],
+            inference: "",
+        };
+        const snippet = snippetFunc(model);
+        expect(snippet[0].content).toEqual(`# Load and run the model:
+llama-cli \\
+  --hf-repo "bartowski/Llama-3.2-3B-Instruct-GGUF" \\
+  --hf-file {{GGUF_FILE}} \\
+  -p "You are a helpful assistant" \\
+  --conversation`);
+    });
+    it("llama.cpp non-conversational", async () => {
+        const { snippet: snippetFunc } = LOCAL_APPS["llama.cpp"];
+        const model = {
+            id: "mlabonne/gemma-2b-GGUF",
+            tags: [],
+            inference: "",
+        };
+        const snippet = snippetFunc(model);
+        expect(snippet[0].content).toEqual(`# Load and run the model:
+llama-cli \\
+  --hf-repo "mlabonne/gemma-2b-GGUF" \\
+  --hf-file {{GGUF_FILE}} \\
+  -p "Once upon a time,"`);
+    });
+    it("vLLM conversational llm", async () => {
+        const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
+        const model = {
+            id: "meta-llama/Llama-3.2-3B-Instruct",
+            pipeline_tag: "text-generation",
+            tags: ["conversational"],
+            inference: "",
+        };
+        const snippet = snippetFunc(model);
+        expect(snippet[0].content.join("\n")).toEqual(`# Load and run the model:
+vllm serve "meta-llama/Llama-3.2-3B-Instruct"
+# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/chat/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "meta-llama/Llama-3.2-3B-Instruct",
+		"messages": [
+			{
+				"role": "user",
+				"content": "What is the capital of France?"
+			}
+		]
+	}'`);
+    });
+    it("vLLM non-conversational llm", async () => {
+        const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
+        const model = {
+            id: "meta-llama/Llama-3.2-3B",
+            tags: [""],
+            inference: "",
+        };
+        const snippet = snippetFunc(model);
+        expect(snippet[0].content.join("\n")).toEqual(`# Load and run the model:
+vllm serve "meta-llama/Llama-3.2-3B"
+# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "meta-llama/Llama-3.2-3B",
+		"prompt": "Once upon a time,",
+		"max_tokens": 512,
+		"temperature": 0.5
+	}'`);
+    });
+    it("vLLM conversational vlm", async () => {
+        const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
+        const model = {
+            id: "meta-llama/Llama-3.2-11B-Vision-Instruct",
+            pipeline_tag: "image-text-to-text",
+            tags: ["conversational"],
+            inference: "",
+        };
+        const snippet = snippetFunc(model);
+        expect(snippet[0].content.join("\n")).toEqual(`# Load and run the model:
+vllm serve "meta-llama/Llama-3.2-11B-Vision-Instruct"
+# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/chat/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "meta-llama/Llama-3.2-11B-Vision-Instruct",
+		"messages": [
+			{
+				"role": "user",
+				"content": [
+					{
+						"type": "text",
+						"text": "Describe this image in one sentence."
+					},
+					{
+						"type": "image_url",
+						"image_url": {
+							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
+						}
+					}
+				]
+			}
+		]
+	}'`);
+    });
+});

package/dist/esm/model-libraries-snippets.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"model-libraries-snippets.d.ts","sourceRoot":"","sources":["../../src/model-libraries-snippets.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAejD,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAKjD,CAAC;AAkBF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAKjD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAIjD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAkBlD,CAAC;AAaF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAIjD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAEF,eAAO,MAAM,iBAAiB,UAAW,SAAS,KAAG,MAAM,EA6C1D,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAwBlD,CAAC;AAuCF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAUlD,CAAC;AAEF,eAAO,MAAM,YAAY,UAAW,SAAS,KAAG,MAAM,EAwCrD,CAAC;AAEF,eAAO,MAAM,gBAAgB,UAAW,SAAS,KAAG,MAAM,EAgBzD,CAAC;AAEF,eAAO,MAAM,YAAY,UAAW,SAAS,KAAG,MAAM,EAmBrD,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAgB/C,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAMlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EASlD,CAAC;AAIF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAO/C,CAAC;AAEF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAMhD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AAEF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAehD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAS9C,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAUlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAalD,CAAC;AAEF,eAAO,MAAM,gBAAgB,UAAW,SAAS,KAAG,MAAM,EAgBzD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAOjD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAIlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAKlD,CAAC;AAEF,eAAO,MAAM,aAAa,QAAO,MAAM,EAQtC,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAKlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAsBlD,CAAC;AAEF,eAAO,MAAM,uBAAuB,UAAW,SAAS,KAAG,MAAM,EAehE,CAAC;AAiBF,eAAO,MAAM,cAAc,UAAW,SAAS,KAAG,MAAM,EAKvD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAyBF,eAAO,MAAM,aAAa,UAAW,SAAS,KAAG,MAAM,EAOtD,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAI7C,CAAC;AAEF,eAAO,MAAM,OAAO,QAA6B,MAAM,EAQtD,CAAC;AAEF,eAAO,MAAM,UAAU,QAAO,MAAM,EAanC,CAAC;AAsCF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAehD,CAAC;AAEF,eAAO,MAAM,kBAAkB,UAAW,SAAS,KAAG,MAAM,EAmC3D,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EA2B7C,CAAC;AAEF,eAAO,MAAM,aAAa,UAAW,SAAS,KAAG,MAAM,EAEtD,CAAC;AASF,eAAO,MAAM,oBAAoB,UAAW,SAAS,KAAG,MAAM,EAoB7D,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAU9C,CAAC;AAEF,eAAO,MAAM,WAAW,UAAW,SAAS,KAAG,MAAM,EAIpD,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAK/C,CAAC;AAkBF,eAAO,MAAM,WAAW,UAAW,SAAS,KAAG,MAAM,EAkBpD,CAAC;AAEF,eAAO,MAAM,YAAY,UAAW,SAAS,KAAG,MAAM,EA4CrD,CAAC;AAEF,eAAO,MAAM,cAAc,UAAW,SAAS,KAAG,MAAM,EAcvD,CAAC;AAiBF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAkB7C,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAKjD,CAAC;AAEF,eAAO,MAAM,gBAAgB,UAAW,SAAS,KAAG,MAAM,EAMzD,CAAC;AAgBF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAEjD,CAAC;AAEF,eAAO,MAAM,MAAM,QAA6B,MAAM,EAMrD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAIjD,CAAC;AAEF,eAAO,MAAM,UAAU,UAAW,SAAS,KAAG,MAAM,EAInD,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,MAAM,EAYhC,CAAC;AAEF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAOhD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAYjD,CAAC;AAEF,eAAO,MAAM,GAAG,UAAW,SAAS,KAAG,MAAM,EAK5C,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAIlD,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAQ7C,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAI7C,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AA6BF,eAAO,MAAM,UAAU,UAAW,SAAS,KAAG,MAAM,EAUnD,CAAC;AAEF,eAAO,MAAM,UAAU,QAAO,MAAM,EAYnC,CAAC;AAEF,eAAO,MAAM,cAAc,UAAW,SAAS,KAAG,MAAM,EAKvD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC"}
+{"version":3,"file":"model-libraries-snippets.d.ts","sourceRoot":"","sources":["../../src/model-libraries-snippets.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAkBjD,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAKjD,CAAC;AAkBF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAKjD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAIjD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAkBlD,CAAC;AAaF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAIjD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAEF,eAAO,MAAM,iBAAiB,UAAW,SAAS,KAAG,MAAM,EA6C1D,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAwBlD,CAAC;AAuCF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAUlD,CAAC;AAEF,eAAO,MAAM,YAAY,UAAW,SAAS,KAAG,MAAM,EAwCrD,CAAC;AAEF,eAAO,MAAM,gBAAgB,UAAW,SAAS,KAAG,MAAM,EAgBzD,CAAC;AAEF,eAAO,MAAM,YAAY,UAAW,SAAS,KAAG,MAAM,EAmBrD,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAgB/C,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAMlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EASlD,CAAC;AAIF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAO/C,CAAC;AAEF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAMhD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AAEF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAehD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAS9C,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAUlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAalD,CAAC;AAEF,eAAO,MAAM,gBAAgB,UAAW,SAAS,KAAG,MAAM,EA0BzD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAOjD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAIlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAKlD,CAAC;AAEF,eAAO,MAAM,aAAa,QAAO,MAAM,EAQtC,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAKlD,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAsBlD,CAAC;AAEF,eAAO,MAAM,uBAAuB,UAAW,SAAS,KAAG,MAAM,EAehE,CAAC;AAiBF,eAAO,MAAM,cAAc,UAAW,SAAS,KAAG,MAAM,EAKvD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAyBF,eAAO,MAAM,aAAa,UAAW,SAAS,KAAG,MAAM,EAOtD,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAI7C,CAAC;AAEF,eAAO,MAAM,OAAO,QAA6B,MAAM,EAQtD,CAAC;AAEF,eAAO,MAAM,UAAU,QAAO,MAAM,EAanC,CAAC;AAsCF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAehD,CAAC;AAEF,eAAO,MAAM,kBAAkB,UAAW,SAAS,KAAG,MAAM,EAmC3D,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EA2B7C,CAAC;AAEF,eAAO,MAAM,aAAa,UAAW,SAAS,KAAG,MAAM,EAEtD,CAAC;AASF,eAAO,MAAM,oBAAoB,UAAW,SAAS,KAAG,MAAM,EAoB7D,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAU9C,CAAC;AAEF,eAAO,MAAM,WAAW,UAAW,SAAS,KAAG,MAAM,EAIpD,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAK/C,CAAC;AAkBF,eAAO,MAAM,WAAW,UAAW,SAAS,KAAG,MAAM,EAkBpD,CAAC;AAEF,eAAO,MAAM,YAAY,UAAW,SAAS,KAAG,MAAM,EA4CrD,CAAC;AAEF,eAAO,MAAM,cAAc,UAAW,SAAS,KAAG,MAAM,EAcvD,CAAC;AAiBF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAkB7C,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAKjD,CAAC;AAEF,eAAO,MAAM,gBAAgB,UAAW,SAAS,KAAG,MAAM,EAMzD,CAAC;AAgBF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAEjD,CAAC;AAEF,eAAO,MAAM,MAAM,QAA6B,MAAM,EAMrD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAIjD,CAAC;AAEF,eAAO,MAAM,UAAU,UAAW,SAAS,KAAG,MAAM,EAInD,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,MAAM,EAYhC,CAAC;AAEF,eAAO,MAAM,OAAO,UAAW,SAAS,KAAG,MAAM,EAOhD,CAAC;AAEF,eAAO,MAAM,QAAQ,UAAW,SAAS,KAAG,MAAM,EAYjD,CAAC;AAEF,eAAO,MAAM,GAAG,UAAW,SAAS,KAAG,MAAM,EAK5C,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC;AAEF,eAAO,MAAM,SAAS,UAAW,SAAS,KAAG,MAAM,EAIlD,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAQ7C,CAAC;AAEF,eAAO,MAAM,IAAI,UAAW,SAAS,KAAG,MAAM,EAI7C,CAAC;AAEF,eAAO,MAAM,MAAM,UAAW,SAAS,KAAG,MAAM,EAI/C,CAAC;AA6BF,eAAO,MAAM,UAAU,UAAW,SAAS,KAAG,MAAM,EAUnD,CAAC;AAEF,eAAO,MAAM,UAAU,QAAO,MAAM,EAYnC,CAAC;AAEF,eAAO,MAAM,cAAc,UAAW,SAAS,KAAG,MAAM,EAKvD,CAAC;AAEF,eAAO,MAAM,KAAK,UAAW,SAAS,KAAG,MAAM,EAI9C,CAAC"}

package/dist/esm/model-libraries-snippets.js

@@ -1,4 +1,6 @@
 import { LIBRARY_TASK_MAPPING } from "./library-to-tasks.js";
+import { getModelInputSnippet } from "./snippets/inputs.js";
+import { stringifyMessages } from "./snippets/common.js";
 const TAG_CUSTOM_CODE = "custom_code";
 function nameWithoutNamespace(modelId) {
     const splitted = modelId.split("/");
@@ -374,23 +376,32 @@ model = keras_hub.models.CausalLM.from_preset("hf://${model.id}", dtype="bfloat1
 # full list here: https://keras.io/api/keras_hub/models/#api-documentation
 `,
 ];
-export const llama_cpp_python = (model) => [
-    `from llama_cpp import Llama
+export const llama_cpp_python = (model) => {
+    const snippets = [
+        `from llama_cpp import Llama
 
 llm = Llama.from_pretrained(
 	repo_id="${model.id}",
 	filename="{{GGUF_FILE}}",
 )
-
-llm.create_chat_completion(
-	messages = [
-		{
-			"role": "user",
-			"content": "What is the capital of France?"
-		}
-	]
-)`,
-];
+`,
+    ];
+    if (model.tags.includes("conversational")) {
+        const messages = getModelInputSnippet(model);
+        snippets.push(`llm.create_chat_completion(
+	messages = ${stringifyMessages(messages, { attributeKeyQuotes: true, indent: "\t" })}
+)`);
+    }
+    else {
+        snippets.push(`output = llm(
+	"Once upon a time,",
+	max_tokens=512,
+	echo=True
+)
+print(output)`);
+    }
+    return snippets;
+};
 export const tf_keras = (model) => [
     `# Note: 'keras<3.x' or 'tf_keras' must be installed (legacy)
 # See https://github.com/keras-team/tf-keras for more details.

package/dist/esm/model-libraries-snippets.spec.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=model-libraries-snippets.spec.d.ts.map

package/dist/esm/model-libraries-snippets.spec.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"model-libraries-snippets.spec.d.ts","sourceRoot":"","sources":["../../src/model-libraries-snippets.spec.ts"],"names":[],"mappings":""}

package/dist/esm/model-libraries-snippets.spec.js

@@ -0,0 +1,49 @@
+import { describe, expect, it } from "vitest";
+import { llama_cpp_python } from "./model-libraries-snippets.js";
+describe("model-libraries-snippets", () => {
+    it("llama_cpp_python conversational", async () => {
+        const model = {
+            id: "bartowski/Llama-3.2-3B-Instruct-GGUF",
+            pipeline_tag: "text-generation",
+            tags: ["conversational"],
+            inference: "",
+        };
+        const snippet = llama_cpp_python(model);
+        expect(snippet.join("\n")).toEqual(`from llama_cpp import Llama
+
+llm = Llama.from_pretrained(
+	repo_id="bartowski/Llama-3.2-3B-Instruct-GGUF",
+	filename="{{GGUF_FILE}}",
+)
+
+llm.create_chat_completion(
+	messages = [
+		{
+			"role": "user",
+			"content": "What is the capital of France?"
+		}
+	]
+)`);
+    });
+    it("llama_cpp_python non-conversational", async () => {
+        const model = {
+            id: "mlabonne/gemma-2b-GGUF",
+            tags: [""],
+            inference: "",
+        };
+        const snippet = llama_cpp_python(model);
+        expect(snippet.join("\n")).toEqual(`from llama_cpp import Llama
+
+llm = Llama.from_pretrained(
+	repo_id="mlabonne/gemma-2b-GGUF",
+	filename="{{GGUF_FILE}}",
+)
+
+output = llm(
+	"Once upon a time,",
+	max_tokens=512,
+	echo=True
+)
+print(output)`);
+    });
+});

package/dist/esm/tasks/audio-classification/inference.d.ts

@@ -24,6 +24,9 @@ export interface AudioClassificationInput {
  * Additional inference parameters for Audio Classification
  */
 export interface AudioClassificationParameters {
+    /**
+     * The function to apply to the model outputs in order to retrieve the scores.
+     */
     function_to_apply?: ClassificationOutputTransform;
     /**
      * When specified, limits the output to the top K most probable classes.

package/dist/esm/tasks/audio-classification/inference.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/audio-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,wBAAwB;IACxC;;;OAGG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;OAEG;IACH,UAAU,CAAC,EAAE,6BAA6B,CAAC;IAC3C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,6BAA6B;IAC7C,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,yBAAyB,GAAG,gCAAgC,EAAE,CAAC;AAC3E;;GAEG;AACH,MAAM,WAAW,gCAAgC;IAChD;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}
+{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/audio-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,wBAAwB;IACxC;;;OAGG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;OAEG;IACH,UAAU,CAAC,EAAE,6BAA6B,CAAC;IAC3C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,6BAA6B;IAC7C;;OAEG;IACH,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,yBAAyB,GAAG,gCAAgC,EAAE,CAAC;AAC3E;;GAEG;AACH,MAAM,WAAW,gCAAgC;IAChD;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}

package/dist/esm/tasks/image-classification/inference.d.ts

@@ -24,6 +24,9 @@ export interface ImageClassificationInput {
  * Additional inference parameters for Image Classification
  */
 export interface ImageClassificationParameters {
+    /**
+     * The function to apply to the model outputs in order to retrieve the scores.
+     */
     function_to_apply?: ClassificationOutputTransform;
     /**
      * When specified, limits the output to the top K most probable classes.

package/dist/esm/tasks/image-classification/inference.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/image-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,wBAAwB;IACxC;;;OAGG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;OAEG;IACH,UAAU,CAAC,EAAE,6BAA6B,CAAC;IAC3C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,6BAA6B;IAC7C,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,yBAAyB,GAAG,gCAAgC,EAAE,CAAC;AAC3E;;GAEG;AACH,MAAM,WAAW,gCAAgC;IAChD;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}
+{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/image-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,wBAAwB;IACxC;;;OAGG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;OAEG;IACH,UAAU,CAAC,EAAE,6BAA6B,CAAC;IAC3C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,6BAA6B;IAC7C;;OAEG;IACH,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,yBAAyB,GAAG,gCAAgC,EAAE,CAAC;AAC3E;;GAEG;AACH,MAAM,WAAW,gCAAgC;IAChD;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}

package/dist/esm/tasks/text-classification/inference.d.ts

@@ -23,6 +23,9 @@ export interface TextClassificationInput {
  * Additional inference parameters for Text Classification
  */
 export interface TextClassificationParameters {
+    /**
+     * The function to apply to the model outputs in order to retrieve the scores.
+     */
     function_to_apply?: ClassificationOutputTransform;
     /**
      * When specified, limits the output to the top K most probable classes.

package/dist/esm/tasks/text-classification/inference.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/text-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,uBAAuB;IACvC;;OAEG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;OAEG;IACH,UAAU,CAAC,EAAE,4BAA4B,CAAC;IAC1C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,4BAA4B;IAC5C,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,wBAAwB,GAAG,+BAA+B,EAAE,CAAC;AACzE;;GAEG;AACH,MAAM,WAAW,+BAA+B;IAC/C;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}
+{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/text-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,uBAAuB;IACvC;;OAEG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;OAEG;IACH,UAAU,CAAC,EAAE,4BAA4B,CAAC;IAC1C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,4BAA4B;IAC5C;;OAEG;IACH,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,wBAAwB,GAAG,+BAA+B,EAAE,CAAC;AACzE;;GAEG;AACH,MAAM,WAAW,+BAA+B;IAC/C;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}

package/dist/esm/tasks/video-classification/inference.d.ts

@@ -27,6 +27,9 @@ export interface VideoClassificationParameters {
      * The sampling rate used to select frames from the video.
      */
     frame_sampling_rate?: number;
+    /**
+     * The function to apply to the model outputs in order to retrieve the scores.
+     */
     function_to_apply?: ClassificationOutputTransform;
     /**
      * The number of sampled frames to consider for classification.

package/dist/esm/tasks/video-classification/inference.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/video-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,wBAAwB;IACxC;;OAEG;IACH,MAAM,EAAE,OAAO,CAAC;IAChB;;OAEG;IACH,UAAU,CAAC,EAAE,6BAA6B,CAAC;IAC3C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,6BAA6B;IAC7C;;OAEG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,yBAAyB,GAAG,gCAAgC,EAAE,CAAC;AAC3E;;GAEG;AACH,MAAM,WAAW,gCAAgC;IAChD;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}
+{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/video-classification/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,wBAAwB;IACxC;;OAEG;IACH,MAAM,EAAE,OAAO,CAAC;IAChB;;OAEG;IACH,UAAU,CAAC,EAAE,6BAA6B,CAAC;IAC3C,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;;;GAIG;AACH,MAAM,WAAW,6BAA6B;IAC7C;;OAEG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B;;OAEG;IACH,iBAAiB,CAAC,EAAE,6BAA6B,CAAC;IAClD;;OAEG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,MAAM,6BAA6B,GAAG,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAC3E,MAAM,MAAM,yBAAyB,GAAG,gCAAgC,EAAE,CAAC;AAC3E;;GAEG;AACH,MAAM,WAAW,gCAAgC;IAChD;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@huggingface/tasks",
   "packageManager": "pnpm@8.10.5",
-  "version": "0.13.4",
+  "version": "0.13.5",
   "description": "List of ML tasks for huggingface.co/tasks",
   "repository": "https://github.com/huggingface/huggingface.js.git",
   "publishConfig": {

package/src/local-apps.spec.ts

@@ -0,0 +1,123 @@
+import { describe, expect, it } from "vitest";
+import { LOCAL_APPS } from "./local-apps.js";
+import type { ModelData } from "./model-data.js";
+
+describe("local-apps", () => {
+	it("llama.cpp conversational", async () => {
+		const { snippet: snippetFunc } = LOCAL_APPS["llama.cpp"];
+		const model: ModelData = {
+			id: "bartowski/Llama-3.2-3B-Instruct-GGUF",
+			tags: ["conversational"],
+			inference: "",
+		};
+		const snippet = snippetFunc(model);
+
+		expect(snippet[0].content).toEqual(`# Load and run the model:
+llama-cli \\
+  --hf-repo "bartowski/Llama-3.2-3B-Instruct-GGUF" \\
+  --hf-file {{GGUF_FILE}} \\
+  -p "You are a helpful assistant" \\
+  --conversation`);
+	});
+
+	it("llama.cpp non-conversational", async () => {
+		const { snippet: snippetFunc } = LOCAL_APPS["llama.cpp"];
+		const model: ModelData = {
+			id: "mlabonne/gemma-2b-GGUF",
+			tags: [],
+			inference: "",
+		};
+		const snippet = snippetFunc(model);
+
+		expect(snippet[0].content).toEqual(`# Load and run the model:
+llama-cli \\
+  --hf-repo "mlabonne/gemma-2b-GGUF" \\
+  --hf-file {{GGUF_FILE}} \\
+  -p "Once upon a time,"`);
+	});
+
+	it("vLLM conversational llm", async () => {
+		const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
+		const model: ModelData = {
+			id: "meta-llama/Llama-3.2-3B-Instruct",
+			pipeline_tag: "text-generation",
+			tags: ["conversational"],
+			inference: "",
+		};
+		const snippet = snippetFunc(model);
+
+		expect((snippet[0].content as string[]).join("\n")).toEqual(`# Load and run the model:
+vllm serve "meta-llama/Llama-3.2-3B-Instruct"
+# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/chat/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "meta-llama/Llama-3.2-3B-Instruct",
+		"messages": [
+			{
+				"role": "user",
+				"content": "What is the capital of France?"
+			}
+		]
+	}'`);
+	});
+
+	it("vLLM non-conversational llm", async () => {
+		const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
+		const model: ModelData = {
+			id: "meta-llama/Llama-3.2-3B",
+			tags: [""],
+			inference: "",
+		};
+		const snippet = snippetFunc(model);
+
+		expect((snippet[0].content as string[]).join("\n")).toEqual(`# Load and run the model:
+vllm serve "meta-llama/Llama-3.2-3B"
+# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "meta-llama/Llama-3.2-3B",
+		"prompt": "Once upon a time,",
+		"max_tokens": 512,
+		"temperature": 0.5
+	}'`);
+	});
+
+	it("vLLM conversational vlm", async () => {
+		const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
+		const model: ModelData = {
+			id: "meta-llama/Llama-3.2-11B-Vision-Instruct",
+			pipeline_tag: "image-text-to-text",
+			tags: ["conversational"],
+			inference: "",
+		};
+		const snippet = snippetFunc(model);
+
+		expect((snippet[0].content as string[]).join("\n")).toEqual(`# Load and run the model:
+vllm serve "meta-llama/Llama-3.2-11B-Vision-Instruct"
+# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/chat/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "meta-llama/Llama-3.2-11B-Vision-Instruct",
+		"messages": [
+			{
+				"role": "user",
+				"content": [
+					{
+						"type": "text",
+						"text": "Describe this image in one sentence."
+					},
+					{
+						"type": "image_url",
+						"image_url": {
+							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
+						}
+					}
+				]
+			}
+		]
+	}'`);
+	});
+});

package/src/local-apps.ts

@@ -1,6 +1,9 @@
 import { parseGGUFQuantLabel } from "./gguf.js";
 import type { ModelData } from "./model-data.js";
 import type { PipelineType } from "./pipelines.js";
+import { stringifyMessages } from "./snippets/common.js";
+import { getModelInputSnippet } from "./snippets/inputs.js";
+import type { ChatCompletionInputMessage } from "./tasks/index.js";
 
 export interface LocalAppSnippet {
 	/**
@@ -92,15 +95,20 @@ function isMlxModel(model: ModelData) {
 }
 
 const snippetLlamacpp = (model: ModelData, filepath?: string): LocalAppSnippet[] => {
-	const command = (binary: string) =>
-		[
+	const command = (binary: string) => {
+		const snippet = [
 			"# Load and run the model:",
 			`${binary} \\`,
 			`  --hf-repo "${model.id}" \\`,
 			`  --hf-file ${filepath ?? "{{GGUF_FILE}}"} \\`,
-			'  -p "You are a helpful assistant" \\',
-			"  --conversation",
-		].join("\n");
+			`  -p "${model.tags.includes("conversational") ? "You are a helpful assistant" : "Once upon a time,"}"`,
+		];
+		if (model.tags.includes("conversational")) {
+			snippet[snippet.length - 1] += " \\";
+			snippet.push("  --conversation");
+		}
+		return snippet.join("\n");
+	};
 	return [
 		{
 			title: "Install from brew",
@@ -178,22 +186,33 @@ const snippetLocalAI = (model: ModelData, filepath?: string): LocalAppSnippet[]
 };
 
 const snippetVllm = (model: ModelData): LocalAppSnippet[] => {
-	const runCommand = [
-		"# Call the server using curl:",
-		`curl -X POST "http://localhost:8000/v1/chat/completions" \\`,
-		`	-H "Content-Type: application/json" \\`,
-		`	--data '{`,
-		`		"model": "${model.id}",`,
-		`		"messages": [`,
-		`			{"role": "user", "content": "Hello!"}`,
-		`		]`,
-		`	}'`,
-	];
+	const messages = getModelInputSnippet(model) as ChatCompletionInputMessage[];
+	const runCommandInstruct = `# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/chat/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "${model.id}",
+		"messages": ${stringifyMessages(messages, {
+			indent: "\t\t",
+			attributeKeyQuotes: true,
+			customContentEscaper: (str) => str.replace(/'/g, "'\\''"),
+		})}
+	}'`;
+	const runCommandNonInstruct = `# Call the server using curl:
+curl -X POST "http://localhost:8000/v1/completions" \\
+	-H "Content-Type: application/json" \\
+	--data '{
+		"model": "${model.id}",
+		"prompt": "Once upon a time,",
+		"max_tokens": 512,
+		"temperature": 0.5
+	}'`;
+	const runCommand = model.tags.includes("conversational") ? runCommandInstruct : runCommandNonInstruct;
 	return [
 		{
 			title: "Install from pip",
 			setup: ["# Install vLLM from pip:", "pip install vllm"].join("\n"),
-			content: [`# Load and run the model:\nvllm serve "${model.id}"`, runCommand.join("\n")],
+			content: [`# Load and run the model:\nvllm serve "${model.id}"`, runCommand],
 		},
 		{
 			title: "Use Docker images",
@@ -210,7 +229,7 @@ const snippetVllm = (model: ModelData): LocalAppSnippet[] => {
 			].join("\n"),
 			content: [
 				`# Load and run the model:\ndocker exec -it my_vllm_container bash -c "vllm serve ${model.id}"`,
-				runCommand.join("\n"),
+				runCommand,
 			],
 		},
 	];

package/src/model-libraries-snippets.spec.ts

@@ -0,0 +1,54 @@
+import { describe, expect, it } from "vitest";
+import type { ModelData } from "./model-data.js";
+import { llama_cpp_python } from "./model-libraries-snippets.js";
+
+describe("model-libraries-snippets", () => {
+	it("llama_cpp_python conversational", async () => {
+		const model: ModelData = {
+			id: "bartowski/Llama-3.2-3B-Instruct-GGUF",
+			pipeline_tag: "text-generation",
+			tags: ["conversational"],
+			inference: "",
+		};
+		const snippet = llama_cpp_python(model);
+
+		expect(snippet.join("\n")).toEqual(`from llama_cpp import Llama
+
+llm = Llama.from_pretrained(
+	repo_id="bartowski/Llama-3.2-3B-Instruct-GGUF",
+	filename="{{GGUF_FILE}}",
+)
+
+llm.create_chat_completion(
+	messages = [
+		{
+			"role": "user",
+			"content": "What is the capital of France?"
+		}
+	]
+)`);
+	});
+
+	it("llama_cpp_python non-conversational", async () => {
+		const model: ModelData = {
+			id: "mlabonne/gemma-2b-GGUF",
+			tags: [""],
+			inference: "",
+		};
+		const snippet = llama_cpp_python(model);
+
+		expect(snippet.join("\n")).toEqual(`from llama_cpp import Llama
+
+llm = Llama.from_pretrained(
+	repo_id="mlabonne/gemma-2b-GGUF",
+	filename="{{GGUF_FILE}}",
+)
+
+output = llm(
+	"Once upon a time,",
+	max_tokens=512,
+	echo=True
+)
+print(output)`);
+	});
+});

package/src/model-libraries-snippets.ts

@@ -1,6 +1,9 @@
 import type { ModelData } from "./model-data.js";
 import type { WidgetExampleTextInput, WidgetExampleSentenceSimilarityInput } from "./widget-example.js";
 import { LIBRARY_TASK_MAPPING } from "./library-to-tasks.js";
+import { getModelInputSnippet } from "./snippets/inputs.js";
+import type { ChatCompletionInputMessage } from "./tasks/index.js";
+import { stringifyMessages } from "./snippets/common.js";
 
 const TAG_CUSTOM_CODE = "custom_code";
 
@@ -418,23 +421,33 @@ model = keras_hub.models.CausalLM.from_preset("hf://${model.id}", dtype="bfloat1
 `,
 ];
 
-export const llama_cpp_python = (model: ModelData): string[] => [
-	`from llama_cpp import Llama
+export const llama_cpp_python = (model: ModelData): string[] => {
+	const snippets = [
+		`from llama_cpp import Llama
 
 llm = Llama.from_pretrained(
 	repo_id="${model.id}",
 	filename="{{GGUF_FILE}}",
 )
+`,
+	];
 
-llm.create_chat_completion(
-	messages = [
-		{
-			"role": "user",
-			"content": "What is the capital of France?"
-		}
-	]
-)`,
-];
+	if (model.tags.includes("conversational")) {
+		const messages = getModelInputSnippet(model) as ChatCompletionInputMessage[];
+		snippets.push(`llm.create_chat_completion(
+	messages = ${stringifyMessages(messages, { attributeKeyQuotes: true, indent: "\t" })}
+)`);
+	} else {
+		snippets.push(`output = llm(
+	"Once upon a time,",
+	max_tokens=512,
+	echo=True
+)
+print(output)`);
+	}
+
+	return snippets;
+};
 
 export const tf_keras = (model: ModelData): string[] => [
 	`# Note: 'keras<3.x' or 'tf_keras' must be installed (legacy)

package/src/tasks/audio-classification/inference.ts

@@ -24,6 +24,9 @@ export interface AudioClassificationInput {
  * Additional inference parameters for Audio Classification
  */
 export interface AudioClassificationParameters {
+	/**
+	 * The function to apply to the model outputs in order to retrieve the scores.
+	 */
 	function_to_apply?: ClassificationOutputTransform;
 	/**
 	 * When specified, limits the output to the top K most probable classes.