@huggingface/tasks 0.19.3 → 0.19.5

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

@@ -226,6 +226,13 @@ export declare const MODEL_LIBRARIES_UI_ELEMENTS: {
         filter: false;
         countDownloads: string;
     };
+    "describe-anything": {
+        prettyLabel: string;
+        repoName: string;
+        repoUrl: string;
+        snippets: (model: ModelData) => string[];
+        filter: false;
+    };
     "dia-tts": {
         prettyLabel: string;
         repoName: string;
@@ -653,6 +660,14 @@ export declare const MODEL_LIBRARIES_UI_ELEMENTS: {
         snippets: (model: ModelData) => string[];
         countDownloads: string;
     };
+    "phantom-wan": {
+        prettyLabel: string;
+        repoName: string;
+        repoUrl: string;
+        snippets: (model: ModelData) => string[];
+        filter: false;
+        countDownloads: string;
+    };
     pxia: {
         prettyLabel: string;
         repoName: string;
@@ -999,6 +1014,14 @@ export declare const MODEL_LIBRARIES_UI_ELEMENTS: {
         countDownloads: string;
         snippets: (model: ModelData) => string[];
     };
+    zonos: {
+        prettyLabel: string;
+        repoName: string;
+        repoUrl: string;
+        docsUrl: string;
+        snippets: (model: ModelData) => string[];
+        filter: false;
+    };
     "3dtopia-xl": {
         prettyLabel: string;
         repoName: string;
@@ -1010,5 +1033,5 @@ export declare const MODEL_LIBRARIES_UI_ELEMENTS: {
 };
 export type ModelLibraryKey = keyof typeof MODEL_LIBRARIES_UI_ELEMENTS;
 export declare const ALL_MODEL_LIBRARY_KEYS: ModelLibraryKey[];
-export declare const ALL_DISPLAY_MODEL_LIBRARY_KEYS: ("adapter-transformers" | "allennlp" | "anemoi" | "araclip" | "asteroid" | "audiocraft" | "audioseal" | "ben2" | "bertopic" | "big_vision" | "birder" | "birefnet" | "bm25s" | "champ" | "chat_tts" | "colpali" | "comet" | "cosmos" | "cxr-foundation" | "deepforest" | "depth-anything-v2" | "depth-pro" | "derm-foundation" | "dia-tts" | "diffree" | "diffusers" | "diffusionkit" | "doctr" | "cartesia_pytorch" | "cartesia_mlx" | "clipscope" | "cosyvoice" | "cotracker" | "edsnlp" | "elm" | "espnet" | "fairseq" | "fastai" | "fasttext" | "flair" | "gemma.cpp" | "geometry-crafter" | "gliner" | "glyph-byt5" | "grok" | "hallo" | "hezar" | "htrflow" | "hunyuan-dit" | "hunyuan3d-2" | "imstoucan" | "index-tts" | "infinite-you" | "keras" | "tf-keras" | "keras-hub" | "k2" | "lightning-ir" | "liveportrait" | "llama-cpp-python" | "mini-omni2" | "mindspore" | "mamba-ssm" | "mars5-tts" | "matanyone" | "mesh-anything" | "merlin" | "medvae" | "mitie" | "ml-agents" | "mlx" | "mlx-image" | "mlc-llm" | "model2vec" | "moshi" | "nemo" | "open-oasis" | "open_clip" | "open-sora" | "outetts" | "paddlenlp" | "peft" | "perception-encoder" | "pxia" | "pyannote-audio" | "py-feat" | "pythae" | "recurrentgemma" | "relik" | "refiners" | "reverb" | "saelens" | "sam2" | "sample-factory" | "sapiens" | "sentence-transformers" | "setfit" | "sklearn" | "spacy" | "span-marker" | "speechbrain" | "ssr-speech" | "stable-audio-tools" | "diffusion-single-file" | "seed-story" | "soloaudio" | "stable-baselines3" | "stanza" | "swarmformer" | "f5-tts" | "genmo" | "tensorflowtts" | "tabpfn" | "terratorch" | "tic-clip" | "timesfm" | "timm" | "transformers" | "transformers.js" | "trellis" | "ultralytics" | "uni-3dar" | "unity-sentis" | "sana" | "vfi-mamba" | "voicecraft" | "wham" | "whisperkit" | "yolov10" | "3dtopia-xl")[];
+export declare const ALL_DISPLAY_MODEL_LIBRARY_KEYS: ("adapter-transformers" | "allennlp" | "anemoi" | "araclip" | "asteroid" | "audiocraft" | "audioseal" | "ben2" | "bertopic" | "big_vision" | "birder" | "birefnet" | "bm25s" | "champ" | "chat_tts" | "colpali" | "comet" | "cosmos" | "cxr-foundation" | "deepforest" | "depth-anything-v2" | "depth-pro" | "derm-foundation" | "describe-anything" | "dia-tts" | "diffree" | "diffusers" | "diffusionkit" | "doctr" | "cartesia_pytorch" | "cartesia_mlx" | "clipscope" | "cosyvoice" | "cotracker" | "edsnlp" | "elm" | "espnet" | "fairseq" | "fastai" | "fasttext" | "flair" | "gemma.cpp" | "geometry-crafter" | "gliner" | "glyph-byt5" | "grok" | "hallo" | "hezar" | "htrflow" | "hunyuan-dit" | "hunyuan3d-2" | "imstoucan" | "index-tts" | "infinite-you" | "keras" | "tf-keras" | "keras-hub" | "k2" | "lightning-ir" | "liveportrait" | "llama-cpp-python" | "mini-omni2" | "mindspore" | "mamba-ssm" | "mars5-tts" | "matanyone" | "mesh-anything" | "merlin" | "medvae" | "mitie" | "ml-agents" | "mlx" | "mlx-image" | "mlc-llm" | "model2vec" | "moshi" | "nemo" | "open-oasis" | "open_clip" | "open-sora" | "outetts" | "paddlenlp" | "peft" | "perception-encoder" | "phantom-wan" | "pxia" | "pyannote-audio" | "py-feat" | "pythae" | "recurrentgemma" | "relik" | "refiners" | "reverb" | "saelens" | "sam2" | "sample-factory" | "sapiens" | "sentence-transformers" | "setfit" | "sklearn" | "spacy" | "span-marker" | "speechbrain" | "ssr-speech" | "stable-audio-tools" | "diffusion-single-file" | "seed-story" | "soloaudio" | "stable-baselines3" | "stanza" | "swarmformer" | "f5-tts" | "genmo" | "tensorflowtts" | "tabpfn" | "terratorch" | "tic-clip" | "timesfm" | "timm" | "transformers" | "transformers.js" | "trellis" | "ultralytics" | "uni-3dar" | "unity-sentis" | "sana" | "vfi-mamba" | "voicecraft" | "wham" | "whisperkit" | "yolov10" | "zonos" | "3dtopia-xl")[];
 //# sourceMappingURL=model-libraries.d.ts.map

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

@@ -1 +1 @@
-{"version":3,"file":"model-libraries.d.ts","sourceRoot":"","sources":["../../src/model-libraries.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,gCAAgC,CAAC;AAEzE;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC;;;;OAIG;IACH,WAAW,EAAE,MAAM,CAAC;IACpB;;OAEG;IACH,QAAQ,EAAE,MAAM,CAAC;IACjB;;OAEG;IACH,OAAO,EAAE,MAAM,CAAC;IAChB;;OAEG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB;;OAEG;IACH,QAAQ,CAAC,EAAE,CAAC,KAAK,EAAE,SAAS,KAAK,MAAM,EAAE,CAAC;IAC1C;;;;;OAKG;IACH,cAAc,CAAC,EAAE,kBAAkB,CAAC;IACpC;;;OAGG;IACH,MAAM,CAAC,EAAE,OAAO,CAAC;CACjB;AAED;;;;;;;;;;;;;GAaG;AAEH,eAAO,MAAM,2BAA2B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAg8BI,CAAC;AAE7C,MAAM,MAAM,eAAe,GAAG,MAAM,OAAO,2BAA2B,CAAC;AAEvE,eAAO,MAAM,sBAAsB,EAA+C,eAAe,EAAE,CAAC;AAEpG,eAAO,MAAM,8BAA8B,qwDAQ1B,CAAC"}
+{"version":3,"file":"model-libraries.d.ts","sourceRoot":"","sources":["../../src/model-libraries.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,gCAAgC,CAAC;AAEzE;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC;;;;OAIG;IACH,WAAW,EAAE,MAAM,CAAC;IACpB;;OAEG;IACH,QAAQ,EAAE,MAAM,CAAC;IACjB;;OAEG;IACH,OAAO,EAAE,MAAM,CAAC;IAChB;;OAEG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB;;OAEG;IACH,QAAQ,CAAC,EAAE,CAAC,KAAK,EAAE,SAAS,KAAK,MAAM,EAAE,CAAC;IAC1C;;;;;OAKG;IACH,cAAc,CAAC,EAAE,kBAAkB,CAAC;IACpC;;;OAGG;IACH,MAAM,CAAC,EAAE,OAAO,CAAC;CACjB;AAED;;;;;;;;;;;;;GAaG;AAEH,eAAO,MAAM,2BAA2B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAu9BI,CAAC;AAE7C,MAAM,MAAM,eAAe,GAAG,MAAM,OAAO,2BAA2B,CAAC;AAEvE,eAAO,MAAM,sBAAsB,EAA+C,eAAe,EAAE,CAAC;AAEpG,eAAO,MAAM,8BAA8B,qzDAQ1B,CAAC"}

package/dist/esm/model-libraries.js

@@ -186,6 +186,13 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
         filter: false,
         countDownloads: `path:"scin_dataset_precomputed_embeddings.npz" OR path:"saved_model.pb"`,
     },
+    "describe-anything": {
+        prettyLabel: "Describe Anything",
+        repoName: "Describe Anything",
+        repoUrl: "https://github.com/NVlabs/describe-anything",
+        snippets: snippets.describe_anything,
+        filter: false,
+    },
     "dia-tts": {
         prettyLabel: "Dia",
         repoName: "Dia",
@@ -618,6 +625,14 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
         snippets: snippets.perception_encoder,
         countDownloads: `path_extension:"pt"`,
     },
+    "phantom-wan": {
+        prettyLabel: "Phantom",
+        repoName: "Phantom",
+        repoUrl: "https://github.com/Phantom-video/Phantom",
+        snippets: snippets.phantom_wan,
+        filter: false,
+        countDownloads: `path_extension:"pth"`,
+    },
     pxia: {
         prettyLabel: "pxia",
         repoName: "pxia",
@@ -965,6 +980,14 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
         countDownloads: `path_extension:"pt" OR path_extension:"safetensors"`,
         snippets: snippets.ultralytics,
     },
+    zonos: {
+        prettyLabel: "Zonos",
+        repoName: "Zonos",
+        repoUrl: "https://github.com/Zyphra/Zonos",
+        docsUrl: "https://github.com/Zyphra/Zonos",
+        snippets: snippets.zonos,
+        filter: false,
+    },
     "3dtopia-xl": {
         prettyLabel: "3DTopia-XL",
         repoName: "3DTopia-XL",

package/dist/esm/tasks/index.js

@@ -113,7 +113,7 @@ export const TASKS_MODEL_LIBRARIES = {
 /**
  * Return the whole TaskData object for a certain task.
  * If the partialTaskData argument is left undefined,
- * the default placholder data will be used.
+ * the default placeholder data will be used.
  */
 function getData(type, partialTaskData = placeholder) {
     return {

package/package.json

@@ -1,7 +1,6 @@
 {
   "name": "@huggingface/tasks",
-  "packageManager": "pnpm@8.10.5",
-  "version": "0.19.3",
+  "version": "0.19.5",
   "description": "List of ML tasks for huggingface.co/tasks",
   "repository": "https://github.com/huggingface/huggingface.js.git",
   "publishConfig": {

package/src/gguf.ts

@@ -1,3 +1,170 @@
+// This list is copied from gguf/types.ts, but will all types available (for backward compatibility)
+// NOT to be confused with GGMLQuantizationType, a FileQuantization can contain multiple GGMLQuantizationType
+// For example, Q4_K_M model can contains Q4_K and Q6_K tensors
+export enum GGMLFileQuantizationType {
+	F32 = 0,
+	F16 = 1,
+	Q4_0 = 2,
+	Q4_1 = 3,
+	Q4_1_SOME_F16 = 4,
+	Q4_2 = 5,
+	Q4_3 = 6,
+	Q8_0 = 7,
+	Q5_0 = 8,
+	Q5_1 = 9,
+	Q2_K = 10,
+	Q3_K_S = 11,
+	Q3_K_M = 12,
+	Q3_K_L = 13,
+	Q4_K_S = 14,
+	Q4_K_M = 15,
+	Q5_K_S = 16,
+	Q5_K_M = 17,
+	Q6_K = 18,
+	IQ2_XXS = 19,
+	IQ2_XS = 20,
+	Q2_K_S = 21,
+	IQ3_XS = 22,
+	IQ3_XXS = 23,
+	IQ1_S = 24,
+	IQ4_NL = 25,
+	IQ3_S = 26,
+	IQ3_M = 27,
+	IQ2_S = 28,
+	IQ2_M = 29,
+	IQ4_XS = 30,
+	IQ1_M = 31,
+	BF16 = 32,
+	Q4_0_4_4 = 33,
+	Q4_0_4_8 = 34,
+	Q4_0_8_8 = 35,
+	TQ1_0 = 36,
+	TQ2_0 = 37,
+
+	// custom quants used by unsloth
+	// they are not officially a scheme enum value in GGUF, but only here for naming
+	Q2_K_XL = 1000,
+	Q3_K_XL = 1001,
+	Q4_K_XL = 1002,
+	Q5_K_XL = 1003,
+	Q6_K_XL = 1004,
+	Q8_K_XL = 1005,
+}
+
+const ggufQuants = Object.values(GGMLFileQuantizationType).filter((v): v is string => typeof v === "string");
+export const GGUF_QUANT_RE = new RegExp(`(?<quant>${ggufQuants.join("|")})` + "(_(?<sizeVariation>[A-Z]+))?");
+export const GGUF_QUANT_RE_GLOBAL = new RegExp(GGUF_QUANT_RE, "g");
+
+export function parseGGUFQuantLabel(fname: string): string | undefined {
+	const quantLabel = fname.toUpperCase().match(GGUF_QUANT_RE_GLOBAL)?.at(-1); // if there is multiple quant substrings in a name, we prefer the last one
+	return quantLabel;
+}
+
+// order of quantization, from biggest to smallest
+// this list must be in sync with the order in GGMLFileQuantizationType
+// the gguf.spec.ts tests are using verify if the order is correct
+export const GGUF_QUANT_ORDER: GGMLFileQuantizationType[] = [
+	GGMLFileQuantizationType.F32,
+	GGMLFileQuantizationType.BF16,
+	GGMLFileQuantizationType.F16,
+	GGMLFileQuantizationType.Q8_K_XL,
+	GGMLFileQuantizationType.Q8_0,
+
+	// 6-bit quantizations
+	GGMLFileQuantizationType.Q6_K_XL,
+	GGMLFileQuantizationType.Q6_K,
+
+	// 5-bit quantizations
+	GGMLFileQuantizationType.Q5_K_XL,
+	GGMLFileQuantizationType.Q5_K_M,
+	GGMLFileQuantizationType.Q5_K_S,
+	GGMLFileQuantizationType.Q5_0,
+	GGMLFileQuantizationType.Q5_1,
+
+	// 4-bit quantizations
+	GGMLFileQuantizationType.Q4_K_XL,
+	GGMLFileQuantizationType.Q4_K_M,
+	GGMLFileQuantizationType.Q4_K_S,
+	GGMLFileQuantizationType.IQ4_NL,
+	GGMLFileQuantizationType.IQ4_XS,
+	GGMLFileQuantizationType.Q4_0_4_4,
+	GGMLFileQuantizationType.Q4_0_4_8,
+	GGMLFileQuantizationType.Q4_0_8_8,
+	GGMLFileQuantizationType.Q4_1_SOME_F16,
+	GGMLFileQuantizationType.Q4_0,
+	GGMLFileQuantizationType.Q4_1,
+	GGMLFileQuantizationType.Q4_2,
+	GGMLFileQuantizationType.Q4_3,
+
+	// 3-bit quantizations
+	GGMLFileQuantizationType.Q3_K_XL,
+	GGMLFileQuantizationType.Q3_K_L,
+	GGMLFileQuantizationType.Q3_K_M,
+	GGMLFileQuantizationType.Q3_K_S,
+	GGMLFileQuantizationType.IQ3_M,
+	GGMLFileQuantizationType.IQ3_S,
+	GGMLFileQuantizationType.IQ3_XS,
+	GGMLFileQuantizationType.IQ3_XXS,
+
+	// 2-bit quantizations
+	GGMLFileQuantizationType.Q2_K_XL,
+	GGMLFileQuantizationType.Q2_K,
+	GGMLFileQuantizationType.Q2_K_S,
+	GGMLFileQuantizationType.IQ2_M,
+	GGMLFileQuantizationType.IQ2_S,
+	GGMLFileQuantizationType.IQ2_XS,
+	GGMLFileQuantizationType.IQ2_XXS,
+
+	// 1-bit quantizations
+	GGMLFileQuantizationType.IQ1_S,
+	GGMLFileQuantizationType.IQ1_M,
+	GGMLFileQuantizationType.TQ1_0,
+	GGMLFileQuantizationType.TQ2_0,
+];
+
+// This function finds the nearest quantization type that is less than or equal to the given quantization type.
+// It returns undefined if no such quantization type is found.
+export function findNearestQuantType(
+	quant: GGMLFileQuantizationType,
+	availableQuants: GGMLFileQuantizationType[]
+): GGMLFileQuantizationType | undefined {
+	// Create a map for quick index lookup from the defined order
+	const orderMap = new Map<GGMLFileQuantizationType, number>();
+	GGUF_QUANT_ORDER.forEach((q, index) => {
+		orderMap.set(q, index);
+	});
+
+	const targetIndex = orderMap.get(quant) ?? 0; // the 0 case should never happen
+
+	// Filter the available quantizations to include only those defined in the order map,
+	// then sort them according to the GGUF_QUANT_ORDER (from largest/index 0 to smallest/highest index).
+	const sortedAvailable = availableQuants
+		.filter((q) => orderMap.has(q))
+		.sort((a, b) => (orderMap.get(a) ?? Infinity) - (orderMap.get(b) ?? Infinity));
+
+	// If no valid quantizations are available after filtering
+	if (sortedAvailable.length === 0) {
+		return undefined;
+	}
+
+	// Iterate through the sorted available quantizations (largest to smallest).
+	// Find the first one whose order index is >= the target index.
+	// This means finding the largest quantization that is smaller than or equal to the target.
+	for (const availableQuant of sortedAvailable) {
+		// We know the key exists due to the filter above.
+		const availableIndex = orderMap.get(availableQuant) ?? 0;
+		if (availableIndex >= targetIndex) {
+			return availableQuant;
+		}
+	}
+
+	// If the loop completes, it means all available quantizations are larger (have a smaller index)
+	// than the target quantization. In this case, return the "smallest" available quantization,
+	// which is the last element in the sorted list (highest index among available).
+	return sortedAvailable[sortedAvailable.length - 1];
+}
+
+// This list is only used to calculate the size of the model, NOT to be confused with the quantization FILE type
 export enum GGMLQuantizationType {
 	F32 = 0,
 	F16 = 1,
@@ -28,13 +195,6 @@ export enum GGMLQuantizationType {
 	F64 = 28,
 	IQ1_M = 29,
 	BF16 = 30,
-}
-
-const ggufQuants = Object.values(GGMLQuantizationType).filter((v): v is string => typeof v === "string");
-export const GGUF_QUANT_RE = new RegExp(`(?<quant>${ggufQuants.join("|")})` + "(_(?<sizeVariation>[A-Z]+))?");
-export const GGUF_QUANT_RE_GLOBAL = new RegExp(GGUF_QUANT_RE, "g");
-
-export function parseGGUFQuantLabel(fname: string): string | undefined {
-	const quantLabel = fname.toUpperCase().match(GGUF_QUANT_RE_GLOBAL)?.at(-1); // if there is multiple quant substrings in a name, we prefer the last one
-	return quantLabel;
+	TQ1_0 = 34,
+	TQ2_0 = 35,
 }

package/src/local-apps.ts

@@ -262,6 +262,41 @@ const snippetTgi = (model: ModelData): LocalAppSnippet[] => {
 	];
 };
 
+const snippetMlxLm = (model: ModelData): LocalAppSnippet[] => {
+	const openaiCurl = [
+		"# Calling the OpenAI-compatible server with curl",
+		`curl -X POST "http://localhost:8000/v1/chat/completions" \\`,
+		`   -H "Content-Type: application/json" \\`,
+		`   --data '{`,
+		`     "model": "${model.id}",`,
+		`     "messages": [`,
+		`       {"role": "user", "content": "Hello"}`,
+		`     ]`,
+		`   }'`,
+	];
+
+	return [
+		{
+			title: "Generate or start a chat session",
+			setup: ["# Install MLX LM", "uv tool install mlx-lm"].join("\n"),
+			content: [
+				...(model.tags.includes("conversational")
+					? ["# Interactive chat REPL", `mlx_lm.chat --model "${model.id}"`]
+					: ["# Generate some text", `mlx_lm.generate --model "${model.id}" --prompt "Once upon a time"`]),
+			].join("\n"),
+		},
+		...(model.tags.includes("conversational")
+			? [
+					{
+						title: "Run an OpenAI-compatible server",
+						setup: ["# Install MLX LM", "uv tool install mlx-lm"].join("\n"),
+						content: ["# Start the server", `mlx_lm.server --model "${model.id}"`, ...openaiCurl].join("\n"),
+					},
+			  ]
+			: []),
+	];
+};
+
 /**
  * Add your new local app here.
  *
@@ -302,6 +337,13 @@ export const LOCAL_APPS = {
 			(model.pipeline_tag === "text-generation" || model.pipeline_tag === "image-text-to-text"),
 		snippet: snippetVllm,
 	},
+	"mlx-lm": {
+		prettyLabel: "MLX LM",
+		docsUrl: "https://github.com/ml-explore/mlx-lm",
+		mainTask: "text-generation",
+		displayOnModelPage: (model) => model.pipeline_tag === "text-generation" && isMlxModel(model),
+		snippet: snippetMlxLm,
+	},
 	tgi: {
 		prettyLabel: "TGI",
 		docsUrl: "https://huggingface.co/docs/text-generation-inference/",

package/src/model-libraries-snippets.ts

@@ -237,6 +237,20 @@ output = model.generate(text)
 sf.write("simple.mp3", output, 44100)`,
 ];
 
+export const describe_anything = (model: ModelData): string[] => [
+	`# pip install git+https://github.com/NVlabs/describe-anything
+from huggingface_hub import snapshot_download
+from dam import DescribeAnythingModel
+
+snapshot_download(${model.id}, local_dir="checkpoints")
+
+dam = DescribeAnythingModel(
+	model_path="checkpoints",
+	conv_mode="v1",
+	prompt_mode="focal_prompt",
+)`,
+];
+
 const diffusersDefaultPrompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k";
 
 const diffusers_default = (model: ModelData) => [
@@ -727,6 +741,17 @@ model = pe.VisionTransformer.from_config("${model.id}", pretrained=True)`;
 		return [vision_encoder];
 	}
 };
+export const phantom_wan = (model: ModelData): string[] => [
+	`from huggingface_hub import snapshot_download
+from phantom_wan import WANI2V, configs
+
+checkpoint_dir = snapshot_download("${model.id}")
+wan_i2v = WanI2V(
+            config=configs.WAN_CONFIGS['i2v-14B'],
+            checkpoint_dir=checkpoint_dir,
+        )
+ video = wan_i2v.generate(text_prompt, image_prompt)`,
+];
 
 export const pyannote_audio_pipeline = (model: ModelData): string[] => [
 	`from pyannote.audio import Pipeline
@@ -1117,13 +1142,31 @@ export const transformers = (model: ModelData): string[] => {
 	}
 
 	if (model.pipeline_tag && LIBRARY_TASK_MAPPING.transformers?.includes(model.pipeline_tag)) {
-		const pipelineSnippet = ["# Use a pipeline as a high-level helper", "from transformers import pipeline", ""];
+		const pipelineSnippet = [
+			"# Use a pipeline as a high-level helper",
+			"from transformers import pipeline",
+			"",
+			`pipe = pipeline("${model.pipeline_tag}", model="${model.id}"` + remote_code_snippet + ")",
+		];
 
-		if (model.tags.includes("conversational") && model.config?.tokenizer_config?.chat_template) {
-			pipelineSnippet.push("messages = [", '    {"role": "user", "content": "Who are you?"},', "]");
-		}
-		pipelineSnippet.push(`pipe = pipeline("${model.pipeline_tag}", model="${model.id}"` + remote_code_snippet + ")");
-		if (model.tags.includes("conversational") && model.config?.tokenizer_config?.chat_template) {
+		if (model.tags.includes("conversational")) {
+			if (model.tags.includes("image-text-to-text")) {
+				pipelineSnippet.push(
+					"messages = [",
+					[
+						"    {",
+						'        "role": "user",',
+						'        "content": [',
+						'            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},',
+						'            {"type": "text", "text": "What animal is on the candy?"}',
+						"        ]",
+						"    },",
+					].join("\n"),
+					"]"
+				);
+			} else {
+				pipelineSnippet.push("messages = [", '    {"role": "user", "content": "Who are you?"},', "]");
+			}
 			pipelineSnippet.push("pipe(messages)");
 		}
 
@@ -1310,30 +1353,66 @@ model = SwarmFormerModel.from_pretrained("${model.id}")
 ];
 
 const mlx_unknown = (model: ModelData): string[] => [
-	`pip install huggingface_hub hf_transfer
+	`# Download the model from the Hub
+pip install huggingface_hub hf_transfer
 
 export HF_HUB_ENABLE_HF_TRANSFER=1
 huggingface-cli download --local-dir ${nameWithoutNamespace(model.id)} ${model.id}`,
 ];
 
 const mlxlm = (model: ModelData): string[] => [
-	`pip install --upgrade mlx-lm
+	`# Make sure mlx-lm is installed
+pip install --upgrade mlx-lm
+
+# Generate text with mlx-lm
+from mlx_lm import load, generate
+
+model, tokenizer = load("${model.id}")
 
-mlx_lm.generate --model ${model.id} --prompt "Hello"`,
+prompt = "Once upon a time in"
+text = generate(model, tokenizer, prompt=prompt, verbose=True)`,
 ];
 
 const mlxchat = (model: ModelData): string[] => [
-	`pip install --upgrade mlx-lm
+	`# Make sure mlx-lm is installed
+pip install --upgrade mlx-lm
 
-mlx_lm.chat --model ${model.id}`,
+# Generate text with mlx-lm
+from mlx_lm import load, generate
+
+model, tokenizer = load("${model.id}")
+
+prompt = "Write a story about Einstein"
+messages = [{"role": "user", "content": prompt}]
+prompt = tokenizer.apply_chat_template(
+    messages, add_generation_prompt=True
+)
+
+text = generate(model, tokenizer, prompt=prompt, verbose=True)`,
 ];
 
 const mlxvlm = (model: ModelData): string[] => [
-	`pip install --upgrade mlx-vlm
+	`Make sure mlx-vlm is installed
+from mlx_vlm import load, generate
+from mlx_vlm.prompt_utils import apply_chat_template
+from mlx_vlm.utils import load_config
+
+# Load the model
+model, processor = load("${model.id}")
+config = load_config("${model.id}")
+
+# Prepare input
+image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
+prompt = "Describe this image."
+
+# Apply chat template
+formatted_prompt = apply_chat_template(
+    processor, config, prompt, num_images=1
+)
 
-mlx_vlm.generate --model ${model.id} \\
-	--prompt "Describe this image." \\
-	--image "https://huggingface.co/datasets/huggingface/documentation-images/resolve/0052a70beed5bf71b92610a43a52df6d286cd5f3/diffusers/rabbit.jpg"`,
+# Generate output
+output = generate(model, processor, formatted_prompt, image)
+print(output)`,
 ];
 
 export const mlxim = (model: ModelData): string[] => [
@@ -1343,11 +1422,11 @@ model = create_model(${model.id})`,
 ];
 
 export const mlx = (model: ModelData): string[] => {
-	if (model.tags.includes("image-text-to-text")) {
+	if (model.pipeline_tag === "image-text-to-text") {
 		return mlxvlm(model);
 	}
-	if (model.tags.includes("conversational")) {
-		if (model.config?.tokenizer_config?.chat_template) {
+	if (model.pipeline_tag === "text-generation") {
+		if (model.tags.includes("conversational")) {
 			return mlxchat(model);
 		} else {
 			return mlxlm(model);
@@ -1482,4 +1561,23 @@ export const hezar = (model: ModelData): string[] => [
 
 model = Model.load("${model.id}")`,
 ];
+
+export const zonos = (model: ModelData): string[] => [
+	`# pip install git+https://github.com/Zyphra/Zonos.git
+import torchaudio
+from zonos.model import Zonos
+from zonos.conditioning import make_cond_dict
+
+model = Zonos.from_pretrained("${model.id}", device="cuda")
+
+wav, sr = torchaudio.load("speaker.wav")           # 5-10s reference clip
+speaker = model.make_speaker_embedding(wav, sr)
+
+cond  = make_cond_dict(text="Hello, world!", speaker=speaker, language="en-us")
+codes = model.generate(model.prepare_conditioning(cond))
+
+audio = model.autoencoder.decode(codes)[0].cpu()
+torchaudio.save("sample.wav", audio, model.autoencoder.sampling_rate)
+`,
+];
 //#endregion

package/src/model-libraries.ts

@@ -230,6 +230,13 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
 		filter: false,
 		countDownloads: `path:"scin_dataset_precomputed_embeddings.npz" OR path:"saved_model.pb"`,
 	},
+	"describe-anything": {
+		prettyLabel: "Describe Anything",
+		repoName: "Describe Anything",
+		repoUrl: "https://github.com/NVlabs/describe-anything",
+		snippets: snippets.describe_anything,
+		filter: false,
+	},
 	"dia-tts": {
 		prettyLabel: "Dia",
 		repoName: "Dia",
@@ -662,6 +669,14 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
 		snippets: snippets.perception_encoder,
 		countDownloads: `path_extension:"pt"`,
 	},
+	"phantom-wan": {
+		prettyLabel: "Phantom",
+		repoName: "Phantom",
+		repoUrl: "https://github.com/Phantom-video/Phantom",
+		snippets: snippets.phantom_wan,
+		filter: false,
+		countDownloads: `path_extension:"pth"`,
+	},
 	pxia: {
 		prettyLabel: "pxia",
 		repoName: "pxia",
@@ -1009,6 +1024,14 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
 		countDownloads: `path_extension:"pt" OR path_extension:"safetensors"`,
 		snippets: snippets.ultralytics,
 	},
+	zonos: {
+		prettyLabel: "Zonos",
+		repoName: "Zonos",
+		repoUrl: "https://github.com/Zyphra/Zonos",
+		docsUrl: "https://github.com/Zyphra/Zonos",
+		snippets: snippets.zonos,
+		filter: false,
+	},
 	"3dtopia-xl": {
 		prettyLabel: "3DTopia-XL",
 		repoName: "3DTopia-XL",

package/src/tasks/index.ts

@@ -184,7 +184,7 @@ export const TASKS_MODEL_LIBRARIES: Record<PipelineType, ModelLibraryKey[]> = {
 /**
  * Return the whole TaskData object for a certain task.
  * If the partialTaskData argument is left undefined,
- * the default placholder data will be used.
+ * the default placeholder data will be used.
  */
 function getData(type: PipelineType, partialTaskData: TaskDataCustom = placeholder): TaskData {
 	return {

package/src/tasks/reinforcement-learning/about.md

@@ -111,7 +111,7 @@ env = gym.make("LunarLander-v2")
 # Loading the saved model
 model = PPO.load("PPO-LunarLander-v2",env=env)
 
-# Initializating the evaluation environment
+# Initializing the evaluation environment
 eval_env = gym.make("LunarLander-v2")
 
 # Running the trained agent on eval_env for 10 time steps and getting the mean reward