@huggingface/tasks 0.19.8 → 0.19.9
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/commonjs/hardware.d.ts +7 -0
- package/dist/commonjs/hardware.d.ts.map +1 -1
- package/dist/commonjs/hardware.js +7 -0
- package/dist/commonjs/local-apps.js +10 -10
- package/dist/commonjs/local-apps.spec.js +2 -3
- package/dist/commonjs/model-libraries.d.ts +8 -1
- package/dist/commonjs/model-libraries.d.ts.map +1 -1
- package/dist/commonjs/model-libraries.js +8 -1
- package/dist/commonjs/snippets/inputs.d.ts.map +1 -1
- package/dist/commonjs/snippets/inputs.js +5 -0
- package/dist/commonjs/tasks/image-to-video/data.d.ts +4 -0
- package/dist/commonjs/tasks/image-to-video/data.d.ts.map +1 -0
- package/dist/commonjs/tasks/image-to-video/data.js +119 -0
- package/dist/commonjs/tasks/image-to-video/inference.d.ts +75 -0
- package/dist/commonjs/tasks/image-to-video/inference.d.ts.map +1 -0
- package/dist/commonjs/tasks/image-to-video/inference.js +2 -0
- package/dist/commonjs/tasks/index.d.ts +1 -0
- package/dist/commonjs/tasks/index.d.ts.map +1 -1
- package/dist/commonjs/tasks/index.js +63 -62
- package/dist/esm/hardware.d.ts +7 -0
- package/dist/esm/hardware.d.ts.map +1 -1
- package/dist/esm/hardware.js +7 -0
- package/dist/esm/local-apps.js +10 -10
- package/dist/esm/local-apps.spec.js +2 -3
- package/dist/esm/model-libraries.d.ts +8 -1
- package/dist/esm/model-libraries.d.ts.map +1 -1
- package/dist/esm/model-libraries.js +8 -1
- package/dist/esm/snippets/inputs.d.ts.map +1 -1
- package/dist/esm/snippets/inputs.js +5 -0
- package/dist/esm/tasks/image-to-video/data.d.ts +4 -0
- package/dist/esm/tasks/image-to-video/data.d.ts.map +1 -0
- package/dist/esm/tasks/image-to-video/data.js +117 -0
- package/dist/esm/tasks/image-to-video/inference.d.ts +75 -0
- package/dist/esm/tasks/image-to-video/inference.d.ts.map +1 -0
- package/dist/esm/tasks/image-to-video/inference.js +1 -0
- package/dist/esm/tasks/index.d.ts +1 -0
- package/dist/esm/tasks/index.d.ts.map +1 -1
- package/dist/esm/tasks/index.js +2 -1
- package/package.json +1 -1
- package/src/hardware.ts +7 -0
- package/src/local-apps.spec.ts +2 -3
- package/src/local-apps.ts +10 -10
- package/src/model-libraries.ts +8 -1
- package/src/snippets/inputs.ts +6 -0
- package/src/tasks/image-text-to-text/about.md +2 -1
- package/src/tasks/image-to-video/about.md +51 -0
- package/src/tasks/image-to-video/data.ts +126 -0
- package/src/tasks/image-to-video/inference.ts +74 -0
- package/src/tasks/image-to-video/spec/input.json +64 -0
- package/src/tasks/image-to-video/spec/output.json +13 -0
- package/src/tasks/index.ts +3 -1
package/dist/esm/hardware.d.ts
CHANGED
|
@@ -96,6 +96,10 @@ export declare const SKUS: {
|
|
|
96
96
|
tflops: number;
|
|
97
97
|
memory: number[];
|
|
98
98
|
};
|
|
99
|
+
"RTX A5000": {
|
|
100
|
+
tflops: number;
|
|
101
|
+
memory: number[];
|
|
102
|
+
};
|
|
99
103
|
"RTX A4000": {
|
|
100
104
|
tflops: number;
|
|
101
105
|
memory: number[];
|
|
@@ -529,6 +533,9 @@ export declare const SKUS: {
|
|
|
529
533
|
"Intel Core Ultra 7 265KF": {
|
|
530
534
|
tflops: number;
|
|
531
535
|
};
|
|
536
|
+
"Intel Core 14th Generation (i7)": {
|
|
537
|
+
tflops: number;
|
|
538
|
+
};
|
|
532
539
|
"Intel Core 13th Generation (i9)": {
|
|
533
540
|
tflops: number;
|
|
534
541
|
};
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"hardware.d.ts","sourceRoot":"","sources":["../../src/hardware.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,eAAO,MAAM,iDAAiD,QAAW,CAAC;AAC1E,eAAO,MAAM,yDAAyD,QAAW,CAAC;AAClF,eAAO,MAAM,oCAAoC,QAAU,CAAC;AAE5D;;;GAGG;AACH,eAAO,MAAM,+CAA+C,QAAW,CAAC;AAExE,MAAM,WAAW,YAAY;IAC5B;;;;;;;;;OASG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;;OAGG;IACH,MAAM,CAAC,EAAE,MAAM,EAAE,CAAC;CAClB;AAED,eAAO,MAAM,sBAAsB,UAAqD,CAAC;AAEzF,eAAO,MAAM,IAAI
|
|
1
|
+
{"version":3,"file":"hardware.d.ts","sourceRoot":"","sources":["../../src/hardware.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,eAAO,MAAM,iDAAiD,QAAW,CAAC;AAC1E,eAAO,MAAM,yDAAyD,QAAW,CAAC;AAClF,eAAO,MAAM,oCAAoC,QAAU,CAAC;AAE5D;;;GAGG;AACH,eAAO,MAAM,+CAA+C,QAAW,CAAC;AAExE,MAAM,WAAW,YAAY;IAC5B;;;;;;;;;OASG;IACH,MAAM,EAAE,MAAM,CAAC;IACf;;;OAGG;IACH,MAAM,CAAC,EAAE,MAAM,EAAE,CAAC;CAClB;AAED,eAAO,MAAM,sBAAsB,UAAqD,CAAC;AAEzF,eAAO,MAAM,IAAI;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAwqBuD,CAAC;AAEzE,MAAM,MAAM,OAAO,GAAG,MAAM,OAAO,IAAI,CAAC"}
|
package/dist/esm/hardware.js
CHANGED
|
@@ -78,6 +78,10 @@ export const SKUS = {
|
|
|
78
78
|
tflops: 38.7,
|
|
79
79
|
memory: [48],
|
|
80
80
|
},
|
|
81
|
+
"RTX A5000": {
|
|
82
|
+
tflops: 34.1,
|
|
83
|
+
memory: [24],
|
|
84
|
+
},
|
|
81
85
|
"RTX A4000": {
|
|
82
86
|
tflops: 19.2,
|
|
83
87
|
memory: [16],
|
|
@@ -511,6 +515,9 @@ export const SKUS = {
|
|
|
511
515
|
"Intel Core Ultra 7 265KF": {
|
|
512
516
|
tflops: 1.53,
|
|
513
517
|
},
|
|
518
|
+
"Intel Core 14th Generation (i7)": {
|
|
519
|
+
tflops: 0.8,
|
|
520
|
+
},
|
|
514
521
|
"Intel Core 13th Generation (i9)": {
|
|
515
522
|
tflops: 0.85,
|
|
516
523
|
},
|
package/dist/esm/local-apps.js
CHANGED
|
@@ -36,18 +36,18 @@ function getQuantTag(filepath) {
|
|
|
36
36
|
const snippetLlamacpp = (model, filepath) => {
|
|
37
37
|
const command = (binary) => {
|
|
38
38
|
const snippet = ["# Load and run the model:", `${binary} -hf ${model.id}${getQuantTag(filepath)}`];
|
|
39
|
-
if (!model.tags.includes("conversational")) {
|
|
40
|
-
// for non-conversational models, add a prompt
|
|
41
|
-
snippet[snippet.length - 1] += " \\";
|
|
42
|
-
snippet.push(' -p "Once upon a time,"');
|
|
43
|
-
}
|
|
44
39
|
return snippet.join("\n");
|
|
45
40
|
};
|
|
46
41
|
return [
|
|
47
42
|
{
|
|
48
43
|
title: "Install from brew",
|
|
49
44
|
setup: "brew install llama.cpp",
|
|
50
|
-
content: command("llama-
|
|
45
|
+
content: command("llama-server"),
|
|
46
|
+
},
|
|
47
|
+
{
|
|
48
|
+
title: "Install from WinGet (Windows)",
|
|
49
|
+
setup: "winget install llama.cpp",
|
|
50
|
+
content: command("llama-server"),
|
|
51
51
|
},
|
|
52
52
|
{
|
|
53
53
|
title: "Use pre-built binary",
|
|
@@ -56,17 +56,17 @@ const snippetLlamacpp = (model, filepath) => {
|
|
|
56
56
|
"# Download pre-built binary from:",
|
|
57
57
|
"# https://github.com/ggerganov/llama.cpp/releases",
|
|
58
58
|
].join("\n"),
|
|
59
|
-
content: command("./llama-
|
|
59
|
+
content: command("./llama-server"),
|
|
60
60
|
},
|
|
61
61
|
{
|
|
62
62
|
title: "Build from source code",
|
|
63
63
|
setup: [
|
|
64
64
|
"git clone https://github.com/ggerganov/llama.cpp.git",
|
|
65
65
|
"cd llama.cpp",
|
|
66
|
-
"cmake -B build
|
|
67
|
-
"cmake --build build -j --target llama-
|
|
66
|
+
"cmake -B build",
|
|
67
|
+
"cmake --build build -j --target llama-server",
|
|
68
68
|
].join("\n"),
|
|
69
|
-
content: command("./build/bin/llama-
|
|
69
|
+
content: command("./build/bin/llama-server"),
|
|
70
70
|
},
|
|
71
71
|
];
|
|
72
72
|
};
|
|
@@ -10,7 +10,7 @@ describe("local-apps", () => {
|
|
|
10
10
|
};
|
|
11
11
|
const snippet = snippetFunc(model);
|
|
12
12
|
expect(snippet[0].content).toEqual(`# Load and run the model:
|
|
13
|
-
llama-
|
|
13
|
+
llama-server -hf bartowski/Llama-3.2-3B-Instruct-GGUF:{{QUANT_TAG}}`);
|
|
14
14
|
});
|
|
15
15
|
it("llama.cpp non-conversational", async () => {
|
|
16
16
|
const { snippet: snippetFunc } = LOCAL_APPS["llama.cpp"];
|
|
@@ -21,8 +21,7 @@ llama-cli -hf bartowski/Llama-3.2-3B-Instruct-GGUF:{{QUANT_TAG}}`);
|
|
|
21
21
|
};
|
|
22
22
|
const snippet = snippetFunc(model);
|
|
23
23
|
expect(snippet[0].content).toEqual(`# Load and run the model:
|
|
24
|
-
llama-
|
|
25
|
-
-p "Once upon a time,"`);
|
|
24
|
+
llama-server -hf mlabonne/gemma-2b-GGUF:{{QUANT_TAG}}`);
|
|
26
25
|
});
|
|
27
26
|
it("vLLM conversational llm", async () => {
|
|
28
27
|
const { snippet: snippetFunc } = LOCAL_APPS["vllm"];
|
|
@@ -112,6 +112,13 @@ export declare const MODEL_LIBRARIES_UI_ELEMENTS: {
|
|
|
112
112
|
countDownloads: string;
|
|
113
113
|
snippets: (model: ModelData) => string[];
|
|
114
114
|
};
|
|
115
|
+
"bagel-mot": {
|
|
116
|
+
prettyLabel: string;
|
|
117
|
+
repoName: string;
|
|
118
|
+
repoUrl: string;
|
|
119
|
+
filter: false;
|
|
120
|
+
countDownloads: string;
|
|
121
|
+
};
|
|
115
122
|
ben2: {
|
|
116
123
|
prettyLabel: string;
|
|
117
124
|
repoName: string;
|
|
@@ -1048,5 +1055,5 @@ export declare const MODEL_LIBRARIES_UI_ELEMENTS: {
|
|
|
1048
1055
|
};
|
|
1049
1056
|
export type ModelLibraryKey = keyof typeof MODEL_LIBRARIES_UI_ELEMENTS;
|
|
1050
1057
|
export declare const ALL_MODEL_LIBRARY_KEYS: ModelLibraryKey[];
|
|
1051
|
-
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" | "kimi-audio" | "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" | "torchgeo" | "transformers" | "transformers.js" | "trellis" | "ultralytics" | "uni-3dar" | "unity-sentis" | "sana" | "vfi-mamba" | "voicecraft" | "wham" | "whisperkit" | "yolov10" | "zonos" | "3dtopia-xl")[];
|
|
1058
|
+
export declare const ALL_DISPLAY_MODEL_LIBRARY_KEYS: ("adapter-transformers" | "allennlp" | "anemoi" | "araclip" | "asteroid" | "audiocraft" | "audioseal" | "bagel-mot" | "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" | "kimi-audio" | "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" | "torchgeo" | "transformers" | "transformers.js" | "trellis" | "ultralytics" | "uni-3dar" | "unity-sentis" | "sana" | "vfi-mamba" | "voicecraft" | "wham" | "whisperkit" | "yolov10" | "zonos" | "3dtopia-xl")[];
|
|
1052
1059
|
//# sourceMappingURL=model-libraries.d.ts.map
|
|
@@ -1 +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
|
|
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;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA6+BI,CAAC;AAE7C,MAAM,MAAM,eAAe,GAAG,MAAM,OAAO,2BAA2B,CAAC;AAEvE,eAAO,MAAM,sBAAsB,EAA+C,eAAe,EAAE,CAAC;AAEpG,eAAO,MAAM,8BAA8B,+1DAQ1B,CAAC"}
|
|
@@ -72,6 +72,13 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
|
|
|
72
72
|
countDownloads: `path_extension:"pth"`,
|
|
73
73
|
snippets: snippets.audioseal,
|
|
74
74
|
},
|
|
75
|
+
"bagel-mot": {
|
|
76
|
+
prettyLabel: "Bagel",
|
|
77
|
+
repoName: "Bagel",
|
|
78
|
+
repoUrl: "https://github.com/ByteDance-Seed/Bagel/",
|
|
79
|
+
filter: false,
|
|
80
|
+
countDownloads: `path:"llm_config.json"`,
|
|
81
|
+
},
|
|
75
82
|
ben2: {
|
|
76
83
|
prettyLabel: "BEN2",
|
|
77
84
|
repoName: "BEN2",
|
|
@@ -885,7 +892,7 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
|
|
|
885
892
|
repoName: "timesfm",
|
|
886
893
|
repoUrl: "https://github.com/google-research/timesfm",
|
|
887
894
|
filter: false,
|
|
888
|
-
countDownloads: `path:"checkpoints/checkpoint_1100000/state/checkpoint"`,
|
|
895
|
+
countDownloads: `path:"checkpoints/checkpoint_1100000/state/checkpoint" OR path:"checkpoints/checkpoint_2150000/state/checkpoint" OR path_extension:"ckpt"`,
|
|
889
896
|
},
|
|
890
897
|
timm: {
|
|
891
898
|
prettyLabel: "timm",
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"inputs.d.ts","sourceRoot":"","sources":["../../../src/snippets/inputs.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,0BAA0B,EAAE,MAAM,mBAAmB,CAAC;AACpE,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;
|
|
1
|
+
{"version":3,"file":"inputs.d.ts","sourceRoot":"","sources":["../../../src/snippets/inputs.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,0BAA0B,EAAE,MAAM,mBAAmB,CAAC;AACpE,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AA4JnD,wBAAgB,oBAAoB,CACnC,KAAK,EAAE,gBAAgB,EACvB,MAAM,UAAQ,EACd,QAAQ,UAAQ,GACd,MAAM,GAAG,0BAA0B,EAAE,CAmBvC"}
|
|
@@ -65,6 +65,10 @@ const inputsImageToImage = () => `{
|
|
|
65
65
|
"image": "cat.png",
|
|
66
66
|
"prompt": "Turn the cat into a tiger."
|
|
67
67
|
}`;
|
|
68
|
+
const inputsImageToVideo = () => `{
|
|
69
|
+
"image": "cat.png",
|
|
70
|
+
"prompt": "The cat starts to dance"
|
|
71
|
+
}`;
|
|
68
72
|
const inputsImageSegmentation = () => `"cats.jpg"`;
|
|
69
73
|
const inputsObjectDetection = () => `"cats.jpg"`;
|
|
70
74
|
const inputsAudioToAudio = () => `"sample1.flac"`;
|
|
@@ -86,6 +90,7 @@ const modelInputSnippets = {
|
|
|
86
90
|
"image-classification": inputsImageClassification,
|
|
87
91
|
"image-to-text": inputsImageToText,
|
|
88
92
|
"image-to-image": inputsImageToImage,
|
|
93
|
+
"image-to-video": inputsImageToVideo,
|
|
89
94
|
"image-segmentation": inputsImageSegmentation,
|
|
90
95
|
"object-detection": inputsObjectDetection,
|
|
91
96
|
"question-answering": inputsQuestionAnswering,
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"data.d.ts","sourceRoot":"","sources":["../../../../src/tasks/image-to-video/data.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAElD,QAAA,MAAM,QAAQ,EAAE,cAyHf,CAAC;AAEF,eAAe,QAAQ,CAAC"}
|
|
@@ -0,0 +1,117 @@
|
|
|
1
|
+
const taskData = {
|
|
2
|
+
datasets: [
|
|
3
|
+
{
|
|
4
|
+
description: "A benchmark dataset for reference image controlled video generation.",
|
|
5
|
+
id: "ali-vilab/VACE-Benchmark",
|
|
6
|
+
},
|
|
7
|
+
{
|
|
8
|
+
description: "A dataset of video generation style preferences.",
|
|
9
|
+
id: "Rapidata/sora-video-generation-style-likert-scoring",
|
|
10
|
+
},
|
|
11
|
+
{
|
|
12
|
+
description: "A dataset with videos and captions throughout the videos.",
|
|
13
|
+
id: "BestWishYsh/ChronoMagic",
|
|
14
|
+
},
|
|
15
|
+
],
|
|
16
|
+
demo: {
|
|
17
|
+
inputs: [
|
|
18
|
+
{
|
|
19
|
+
filename: "image-to-video-input.jpg",
|
|
20
|
+
type: "img",
|
|
21
|
+
},
|
|
22
|
+
{
|
|
23
|
+
label: "Optional Text Prompt",
|
|
24
|
+
content: "This penguin is dancing",
|
|
25
|
+
type: "text",
|
|
26
|
+
},
|
|
27
|
+
],
|
|
28
|
+
outputs: [
|
|
29
|
+
{
|
|
30
|
+
filename: "image-to-video-output.gif",
|
|
31
|
+
type: "img",
|
|
32
|
+
},
|
|
33
|
+
],
|
|
34
|
+
},
|
|
35
|
+
metrics: [
|
|
36
|
+
{
|
|
37
|
+
description: "Fréchet Video Distance (FVD) measures the perceptual similarity between the distributions of generated videos and a set of real videos, assessing overall visual quality and temporal coherence of the video generated from an input image.",
|
|
38
|
+
id: "fvd",
|
|
39
|
+
},
|
|
40
|
+
{
|
|
41
|
+
description: "CLIP Score measures the semantic similarity between a textual prompt (if provided alongside the input image) and the generated video frames. It evaluates how well the video's generated content and motion align with the textual description, conditioned on the initial image.",
|
|
42
|
+
id: "clip_score",
|
|
43
|
+
},
|
|
44
|
+
{
|
|
45
|
+
description: "First Frame Fidelity, often measured using LPIPS (Learned Perceptual Image Patch Similarity), PSNR, or SSIM, quantifies how closely the first frame of the generated video matches the input conditioning image.",
|
|
46
|
+
id: "lpips",
|
|
47
|
+
},
|
|
48
|
+
{
|
|
49
|
+
description: "Identity Preservation Score measures the consistency of identity (e.g., a person's face or a specific object's characteristics) between the input image and throughout the generated video frames, often calculated using features from specialized models like face recognition (e.g., ArcFace) or re-identification models.",
|
|
50
|
+
id: "identity_preservation",
|
|
51
|
+
},
|
|
52
|
+
{
|
|
53
|
+
description: "Motion Score evaluates the quality, realism, and temporal consistency of motion in the video generated from a static image. This can be based on optical flow analysis (e.g., smoothness, magnitude), consistency of object trajectories, or specific motion plausibility assessments.",
|
|
54
|
+
id: "motion_score",
|
|
55
|
+
},
|
|
56
|
+
],
|
|
57
|
+
models: [
|
|
58
|
+
{
|
|
59
|
+
description: "LTX-Video, a 13B parameter model for high quality video generation",
|
|
60
|
+
id: "Lightricks/LTX-Video-0.9.7-dev",
|
|
61
|
+
},
|
|
62
|
+
{
|
|
63
|
+
description: "A 14B parameter model for reference image controlled video generation",
|
|
64
|
+
id: "Wan-AI/Wan2.1-VACE-14B",
|
|
65
|
+
},
|
|
66
|
+
{
|
|
67
|
+
description: "An image-to-video generation model using FramePack F1 methodology with Hunyuan-DiT architecture",
|
|
68
|
+
id: "lllyasviel/FramePack_F1_I2V_HY_20250503",
|
|
69
|
+
},
|
|
70
|
+
{
|
|
71
|
+
description: "A distilled version of the LTX-Video-0.9.7-dev model for faster inference",
|
|
72
|
+
id: "Lightricks/LTX-Video-0.9.7-distilled",
|
|
73
|
+
},
|
|
74
|
+
{
|
|
75
|
+
description: "An image-to-video generation model by Skywork AI, 14B parameters, producing 720p videos.",
|
|
76
|
+
id: "Skywork/SkyReels-V2-I2V-14B-720P",
|
|
77
|
+
},
|
|
78
|
+
{
|
|
79
|
+
description: "Image-to-video variant of Tencent's HunyuanVideo.",
|
|
80
|
+
id: "tencent/HunyuanVideo-I2V",
|
|
81
|
+
},
|
|
82
|
+
{
|
|
83
|
+
description: "A 14B parameter model for 720p image-to-video generation by Wan-AI.",
|
|
84
|
+
id: "Wan-AI/Wan2.1-I2V-14B-720P",
|
|
85
|
+
},
|
|
86
|
+
{
|
|
87
|
+
description: "A Diffusers version of the Wan2.1-I2V-14B-720P model for 720p image-to-video generation.",
|
|
88
|
+
id: "Wan-AI/Wan2.1-I2V-14B-720P-Diffusers",
|
|
89
|
+
},
|
|
90
|
+
],
|
|
91
|
+
spaces: [
|
|
92
|
+
{
|
|
93
|
+
description: "An application to generate videos fast.",
|
|
94
|
+
id: "Lightricks/ltx-video-distilled",
|
|
95
|
+
},
|
|
96
|
+
{
|
|
97
|
+
description: "Generate videos with the FramePack-F1",
|
|
98
|
+
id: "linoyts/FramePack-F1",
|
|
99
|
+
},
|
|
100
|
+
{
|
|
101
|
+
description: "Generate videos with the FramePack",
|
|
102
|
+
id: "lisonallen/framepack-i2v",
|
|
103
|
+
},
|
|
104
|
+
{
|
|
105
|
+
description: "Wan2.1 with CausVid LoRA",
|
|
106
|
+
id: "multimodalart/wan2-1-fast",
|
|
107
|
+
},
|
|
108
|
+
{
|
|
109
|
+
description: "A demo for Stable Video Diffusion",
|
|
110
|
+
id: "multimodalart/stable-video-diffusion",
|
|
111
|
+
},
|
|
112
|
+
],
|
|
113
|
+
summary: "Image-to-video models take a still image as input and generate a video. These models can be guided by text prompts to influence the content and style of the output video.",
|
|
114
|
+
widgetModels: [],
|
|
115
|
+
youtubeId: undefined,
|
|
116
|
+
};
|
|
117
|
+
export default taskData;
|
|
@@ -0,0 +1,75 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Inference code generated from the JSON schema spec in ./spec
|
|
3
|
+
*
|
|
4
|
+
* Using src/scripts/inference-codegen
|
|
5
|
+
*/
|
|
6
|
+
/**
|
|
7
|
+
* Inputs for Image To Video inference
|
|
8
|
+
*/
|
|
9
|
+
export interface ImageToVideoInput {
|
|
10
|
+
/**
|
|
11
|
+
* The input image data as a base64-encoded string. If no `parameters` are provided, you can
|
|
12
|
+
* also provide the image data as a raw bytes payload.
|
|
13
|
+
*/
|
|
14
|
+
inputs: Blob;
|
|
15
|
+
/**
|
|
16
|
+
* Additional inference parameters for Image To Video
|
|
17
|
+
*/
|
|
18
|
+
parameters?: ImageToVideoParameters;
|
|
19
|
+
[property: string]: unknown;
|
|
20
|
+
}
|
|
21
|
+
/**
|
|
22
|
+
* Additional inference parameters for Image To Video
|
|
23
|
+
*/
|
|
24
|
+
export interface ImageToVideoParameters {
|
|
25
|
+
/**
|
|
26
|
+
* For diffusion models. A higher guidance scale value encourages the model to generate
|
|
27
|
+
* videos closely linked to the text prompt at the expense of lower image quality.
|
|
28
|
+
*/
|
|
29
|
+
guidance_scale?: number;
|
|
30
|
+
/**
|
|
31
|
+
* One prompt to guide what NOT to include in video generation.
|
|
32
|
+
*/
|
|
33
|
+
negative_prompt?: string;
|
|
34
|
+
/**
|
|
35
|
+
* The num_frames parameter determines how many video frames are generated.
|
|
36
|
+
*/
|
|
37
|
+
num_frames?: number;
|
|
38
|
+
/**
|
|
39
|
+
* The number of denoising steps. More denoising steps usually lead to a higher quality
|
|
40
|
+
* video at the expense of slower inference.
|
|
41
|
+
*/
|
|
42
|
+
num_inference_steps?: number;
|
|
43
|
+
/**
|
|
44
|
+
* The text prompt to guide the video generation.
|
|
45
|
+
*/
|
|
46
|
+
prompt?: string;
|
|
47
|
+
/**
|
|
48
|
+
* Seed for the random number generator.
|
|
49
|
+
*/
|
|
50
|
+
seed?: number;
|
|
51
|
+
/**
|
|
52
|
+
* The size in pixel of the output video frames.
|
|
53
|
+
*/
|
|
54
|
+
target_size?: TargetSize;
|
|
55
|
+
[property: string]: unknown;
|
|
56
|
+
}
|
|
57
|
+
/**
|
|
58
|
+
* The size in pixel of the output video frames.
|
|
59
|
+
*/
|
|
60
|
+
export interface TargetSize {
|
|
61
|
+
height: number;
|
|
62
|
+
width: number;
|
|
63
|
+
[property: string]: unknown;
|
|
64
|
+
}
|
|
65
|
+
/**
|
|
66
|
+
* Outputs of inference for the Image To Video task
|
|
67
|
+
*/
|
|
68
|
+
export interface ImageToVideoOutput {
|
|
69
|
+
/**
|
|
70
|
+
* The generated video returned as raw bytes in the payload.
|
|
71
|
+
*/
|
|
72
|
+
video: unknown;
|
|
73
|
+
[property: string]: unknown;
|
|
74
|
+
}
|
|
75
|
+
//# sourceMappingURL=inference.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"inference.d.ts","sourceRoot":"","sources":["../../../../src/tasks/image-to-video/inference.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH;;GAEG;AACH,MAAM,WAAW,iBAAiB;IACjC;;;OAGG;IACH,MAAM,EAAE,IAAI,CAAC;IACb;;OAEG;IACH,UAAU,CAAC,EAAE,sBAAsB,CAAC;IACpC,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,WAAW,sBAAsB;IACtC;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB;;OAEG;IACH,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB;;OAEG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB;;;OAGG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B;;OAEG;IACH,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB;;OAEG;IACH,IAAI,CAAC,EAAE,MAAM,CAAC;IACd;;OAEG;IACH,WAAW,CAAC,EAAE,UAAU,CAAC;IACzB,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B;AACD;;GAEG;AACH,MAAM,WAAW,kBAAkB;IAClC;;OAEG;IACH,KAAK,EAAE,OAAO,CAAC;IACf,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;CAC5B"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
export {};
|
|
@@ -9,6 +9,7 @@ export type { ImageClassificationInput, ImageClassificationOutput, ImageClassifi
|
|
|
9
9
|
export type * from "./image-to-image/inference.js";
|
|
10
10
|
export type { ImageToTextInput, ImageToTextOutput, ImageToTextParameters } from "./image-to-text/inference.js";
|
|
11
11
|
export type * from "./image-segmentation/inference.js";
|
|
12
|
+
export type { ImageToVideoInput, ImageToVideoOutput, ImageToVideoParameters } from "./image-to-video/inference.js";
|
|
12
13
|
export type * from "./object-detection/inference.js";
|
|
13
14
|
export type * from "./depth-estimation/inference.js";
|
|
14
15
|
export type * from "./question-answering/inference.js";
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/tasks/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/tasks/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAgDpD,mBAAmB,qCAAqC,CAAC;AACzD,mBAAmB,6CAA6C,CAAC;AACjE,YAAY,EACX,mBAAmB,EACnB,0BAA0B,EAC1B,oBAAoB,EACpB,4BAA4B,EAC5B,2BAA2B,EAC3B,0BAA0B,EAC1B,gCAAgC,EAChC,+BAA+B,GAC/B,MAAM,gCAAgC,CAAC;AACxC,mBAAmB,4CAA4C,CAAC;AAChE,mBAAmB,mCAAmC,CAAC;AACvD,mBAAmB,0BAA0B,CAAC;AAC9C,YAAY,EACX,wBAAwB,EACxB,yBAAyB,EACzB,gCAAgC,EAChC,6BAA6B,GAC7B,MAAM,qCAAqC,CAAC;AAC7C,mBAAmB,+BAA+B,CAAC;AACnD,YAAY,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,qBAAqB,EAAE,MAAM,8BAA8B,CAAC;AAC/G,mBAAmB,mCAAmC,CAAC;AACvD,YAAY,EAAE,iBAAiB,EAAE,kBAAkB,EAAE,sBAAsB,EAAE,MAAM,+BAA+B,CAAC;AACnH,mBAAmB,iCAAiC,CAAC;AACrD,mBAAmB,iCAAiC,CAAC;AACrD,mBAAmB,mCAAmC,CAAC;AACvD,mBAAmB,oCAAoC,CAAC;AACxD,mBAAmB,8BAA8B,CAAC;AAClD,mBAAmB,yCAAyC,CAAC;AAC7D,YAAY,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,qBAAqB,EAAE,MAAM,8BAA8B,CAAC;AAC/G,YAAY,EAAE,qBAAqB,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,8BAA8B,CAAC;AAC/G,YAAY,EAAE,sBAAsB,EAAE,iBAAiB,EAAE,kBAAkB,EAAE,MAAM,+BAA+B,CAAC;AACnH,mBAAmB,qCAAqC,CAAC;AACzD,YAAY,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,MAAM,4BAA4B,CAAC;AACtF,YAAY,EACX,6BAA6B,EAC7B,uBAAuB,EACvB,wBAAwB,EACxB,+BAA+B,EAC/B,4BAA4B,GAC5B,MAAM,oCAAoC,CAAC;AAC5C,YAAY,EACX,gCAAgC,EAChC,gCAAgC,EAChC,mBAAmB,EACnB,oBAAoB,EACpB,2BAA2B,EAC3B,qCAAqC,EACrC,kCAAkC,EAClC,yBAAyB,EACzB,uCAAuC,EACvC,0BAA0B,GAC1B,MAAM,gCAAgC,CAAC;AACxC,mBAAmB,qCAAqC,CAAC;AACzD,mBAAmB,0CAA0C,CAAC;AAC9D,mBAAmB,yCAAyC,CAAC;AAC7D,mBAAmB,+CAA+C,CAAC;AACnE,YAAY,EACX,WAAW,EACX,4BAA4B,EAC5B,6BAA6B,EAC7B,oCAAoC,GACpC,MAAM,2CAA2C,CAAC;AAEnD,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAE7D;;GAEG;AACH,eAAO,MAAM,qBAAqB,EAAE,MAAM,CAAC,YAAY,EAAE,eAAe,EAAE,CAgEzE,CAAC;AAoBF,eAAO,MAAM,UAAU,EAAE,MAAM,CAAC,YAAY,EAAE,QAAQ,GAAG,SAAS,CAwDxD,CAAC;AAEX,MAAM,WAAW,WAAW;IAC3B,WAAW,EAAE,MAAM,CAAC;IACpB,EAAE,EAAE,MAAM,CAAC;CACX;AAED,MAAM,MAAM,aAAa,GACtB;IACA,QAAQ,EAAE,MAAM,CAAC;IACjB,IAAI,EAAE,OAAO,CAAC;CACb,GACD;IACA,IAAI,EAAE,KAAK,CAAC;QACX,KAAK,EAAE,MAAM,CAAC;QACd,KAAK,EAAE,MAAM,CAAC;KACd,CAAC,CAAC;IACH,IAAI,EAAE,OAAO,CAAC;CACb,GACD;IACA,QAAQ,EAAE,MAAM,CAAC;IACjB,IAAI,EAAE,KAAK,CAAC;CACX,GACD;IACA,KAAK,EAAE,MAAM,EAAE,EAAE,CAAC;IAClB,IAAI,EAAE,SAAS,CAAC;CACf,GACD;IACA,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;CACZ,GACD;IACA,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,KAAK,CAAC;QACb,GAAG,EAAE,MAAM,CAAC;QACZ,KAAK,EAAE,MAAM,CAAC;QACd,IAAI,EAAE,MAAM,CAAC;KACb,CAAC,CAAC;IACH,IAAI,EAAE,kBAAkB,CAAC;CACxB,CAAC;AAEL,MAAM,WAAW,QAAQ;IACxB,MAAM,EAAE,aAAa,EAAE,CAAC;IACxB,OAAO,EAAE,aAAa,EAAE,CAAC;CACzB;AAED,MAAM,WAAW,QAAQ;IACxB,QAAQ,EAAE,WAAW,EAAE,CAAC;IACxB,IAAI,EAAE,QAAQ,CAAC;IACf,EAAE,EAAE,YAAY,CAAC;IACjB,WAAW,CAAC,EAAE,YAAY,CAAC;IAC3B,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,KAAK,EAAE,MAAM,CAAC;IACd,SAAS,EAAE,eAAe,EAAE,CAAC;IAC7B,OAAO,EAAE,WAAW,EAAE,CAAC;IACvB,MAAM,EAAE,WAAW,EAAE,CAAC;IACtB,MAAM,EAAE,WAAW,EAAE,CAAC;IACtB,OAAO,EAAE,MAAM,CAAC;IAChB,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB,SAAS,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,MAAM,cAAc,GAAG,IAAI,CAAC,QAAQ,EAAE,IAAI,GAAG,OAAO,GAAG,WAAW,CAAC,CAAC"}
|
package/dist/esm/tasks/index.js
CHANGED
|
@@ -12,6 +12,7 @@ import imageToImage from "./image-to-image/data.js";
|
|
|
12
12
|
import imageToText from "./image-to-text/data.js";
|
|
13
13
|
import imageTextToText from "./image-text-to-text/data.js";
|
|
14
14
|
import imageSegmentation from "./image-segmentation/data.js";
|
|
15
|
+
import imageToVideo from "./image-to-video/data.js";
|
|
15
16
|
import maskGeneration from "./mask-generation/data.js";
|
|
16
17
|
import objectDetection from "./object-detection/data.js";
|
|
17
18
|
import depthEstimation from "./depth-estimation/data.js";
|
|
@@ -145,7 +146,7 @@ export const TASKS_DATA = {
|
|
|
145
146
|
"image-to-image": getData("image-to-image", imageToImage),
|
|
146
147
|
"image-text-to-text": getData("image-text-to-text", imageTextToText),
|
|
147
148
|
"image-to-text": getData("image-to-text", imageToText),
|
|
148
|
-
"image-to-video":
|
|
149
|
+
"image-to-video": getData("image-to-video", imageToVideo),
|
|
149
150
|
"keypoint-detection": getData("keypoint-detection", keypointDetection),
|
|
150
151
|
"mask-generation": getData("mask-generation", maskGeneration),
|
|
151
152
|
"multiple-choice": undefined,
|
package/package.json
CHANGED
package/src/hardware.ts
CHANGED
|
@@ -100,6 +100,10 @@ export const SKUS = {
|
|
|
100
100
|
tflops: 38.7,
|
|
101
101
|
memory: [48],
|
|
102
102
|
},
|
|
103
|
+
"RTX A5000": {
|
|
104
|
+
tflops: 34.1,
|
|
105
|
+
memory: [24],
|
|
106
|
+
},
|
|
103
107
|
"RTX A4000": {
|
|
104
108
|
tflops: 19.2,
|
|
105
109
|
memory: [16],
|
|
@@ -533,6 +537,9 @@ export const SKUS = {
|
|
|
533
537
|
"Intel Core Ultra 7 265KF": {
|
|
534
538
|
tflops: 1.53,
|
|
535
539
|
},
|
|
540
|
+
"Intel Core 14th Generation (i7)": {
|
|
541
|
+
tflops: 0.8,
|
|
542
|
+
},
|
|
536
543
|
"Intel Core 13th Generation (i9)": {
|
|
537
544
|
tflops: 0.85,
|
|
538
545
|
},
|
package/src/local-apps.spec.ts
CHANGED
|
@@ -13,7 +13,7 @@ describe("local-apps", () => {
|
|
|
13
13
|
const snippet = snippetFunc(model);
|
|
14
14
|
|
|
15
15
|
expect(snippet[0].content).toEqual(`# Load and run the model:
|
|
16
|
-
llama-
|
|
16
|
+
llama-server -hf bartowski/Llama-3.2-3B-Instruct-GGUF:{{QUANT_TAG}}`);
|
|
17
17
|
});
|
|
18
18
|
|
|
19
19
|
it("llama.cpp non-conversational", async () => {
|
|
@@ -26,8 +26,7 @@ llama-cli -hf bartowski/Llama-3.2-3B-Instruct-GGUF:{{QUANT_TAG}}`);
|
|
|
26
26
|
const snippet = snippetFunc(model);
|
|
27
27
|
|
|
28
28
|
expect(snippet[0].content).toEqual(`# Load and run the model:
|
|
29
|
-
llama-
|
|
30
|
-
-p "Once upon a time,"`);
|
|
29
|
+
llama-server -hf mlabonne/gemma-2b-GGUF:{{QUANT_TAG}}`);
|
|
31
30
|
});
|
|
32
31
|
|
|
33
32
|
it("vLLM conversational llm", async () => {
|
package/src/local-apps.ts
CHANGED
|
@@ -108,18 +108,18 @@ function getQuantTag(filepath?: string): string {
|
|
|
108
108
|
const snippetLlamacpp = (model: ModelData, filepath?: string): LocalAppSnippet[] => {
|
|
109
109
|
const command = (binary: string) => {
|
|
110
110
|
const snippet = ["# Load and run the model:", `${binary} -hf ${model.id}${getQuantTag(filepath)}`];
|
|
111
|
-
if (!model.tags.includes("conversational")) {
|
|
112
|
-
// for non-conversational models, add a prompt
|
|
113
|
-
snippet[snippet.length - 1] += " \\";
|
|
114
|
-
snippet.push(' -p "Once upon a time,"');
|
|
115
|
-
}
|
|
116
111
|
return snippet.join("\n");
|
|
117
112
|
};
|
|
118
113
|
return [
|
|
119
114
|
{
|
|
120
115
|
title: "Install from brew",
|
|
121
116
|
setup: "brew install llama.cpp",
|
|
122
|
-
content: command("llama-
|
|
117
|
+
content: command("llama-server"),
|
|
118
|
+
},
|
|
119
|
+
{
|
|
120
|
+
title: "Install from WinGet (Windows)",
|
|
121
|
+
setup: "winget install llama.cpp",
|
|
122
|
+
content: command("llama-server"),
|
|
123
123
|
},
|
|
124
124
|
{
|
|
125
125
|
title: "Use pre-built binary",
|
|
@@ -128,17 +128,17 @@ const snippetLlamacpp = (model: ModelData, filepath?: string): LocalAppSnippet[]
|
|
|
128
128
|
"# Download pre-built binary from:",
|
|
129
129
|
"# https://github.com/ggerganov/llama.cpp/releases",
|
|
130
130
|
].join("\n"),
|
|
131
|
-
content: command("./llama-
|
|
131
|
+
content: command("./llama-server"),
|
|
132
132
|
},
|
|
133
133
|
{
|
|
134
134
|
title: "Build from source code",
|
|
135
135
|
setup: [
|
|
136
136
|
"git clone https://github.com/ggerganov/llama.cpp.git",
|
|
137
137
|
"cd llama.cpp",
|
|
138
|
-
"cmake -B build
|
|
139
|
-
"cmake --build build -j --target llama-
|
|
138
|
+
"cmake -B build",
|
|
139
|
+
"cmake --build build -j --target llama-server",
|
|
140
140
|
].join("\n"),
|
|
141
|
-
content: command("./build/bin/llama-
|
|
141
|
+
content: command("./build/bin/llama-server"),
|
|
142
142
|
},
|
|
143
143
|
];
|
|
144
144
|
};
|
package/src/model-libraries.ts
CHANGED
|
@@ -116,6 +116,13 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
|
|
|
116
116
|
countDownloads: `path_extension:"pth"`,
|
|
117
117
|
snippets: snippets.audioseal,
|
|
118
118
|
},
|
|
119
|
+
"bagel-mot": {
|
|
120
|
+
prettyLabel: "Bagel",
|
|
121
|
+
repoName: "Bagel",
|
|
122
|
+
repoUrl: "https://github.com/ByteDance-Seed/Bagel/",
|
|
123
|
+
filter: false,
|
|
124
|
+
countDownloads: `path:"llm_config.json"`,
|
|
125
|
+
},
|
|
119
126
|
ben2: {
|
|
120
127
|
prettyLabel: "BEN2",
|
|
121
128
|
repoName: "BEN2",
|
|
@@ -929,7 +936,7 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
|
|
|
929
936
|
repoName: "timesfm",
|
|
930
937
|
repoUrl: "https://github.com/google-research/timesfm",
|
|
931
938
|
filter: false,
|
|
932
|
-
countDownloads: `path:"checkpoints/checkpoint_1100000/state/checkpoint"`,
|
|
939
|
+
countDownloads: `path:"checkpoints/checkpoint_1100000/state/checkpoint" OR path:"checkpoints/checkpoint_2150000/state/checkpoint" OR path_extension:"ckpt"`,
|
|
933
940
|
},
|
|
934
941
|
timm: {
|
|
935
942
|
prettyLabel: "timm",
|
package/src/snippets/inputs.ts
CHANGED
|
@@ -91,6 +91,11 @@ const inputsImageToImage = () => `{
|
|
|
91
91
|
"prompt": "Turn the cat into a tiger."
|
|
92
92
|
}`;
|
|
93
93
|
|
|
94
|
+
const inputsImageToVideo = () => `{
|
|
95
|
+
"image": "cat.png",
|
|
96
|
+
"prompt": "The cat starts to dance"
|
|
97
|
+
}`;
|
|
98
|
+
|
|
94
99
|
const inputsImageSegmentation = () => `"cats.jpg"`;
|
|
95
100
|
|
|
96
101
|
const inputsObjectDetection = () => `"cats.jpg"`;
|
|
@@ -126,6 +131,7 @@ const modelInputSnippets: {
|
|
|
126
131
|
"image-classification": inputsImageClassification,
|
|
127
132
|
"image-to-text": inputsImageToText,
|
|
128
133
|
"image-to-image": inputsImageToImage,
|
|
134
|
+
"image-to-video": inputsImageToVideo,
|
|
129
135
|
"image-segmentation": inputsImageSegmentation,
|
|
130
136
|
"object-detection": inputsObjectDetection,
|
|
131
137
|
"question-answering": inputsQuestionAnswering,
|
|
@@ -85,9 +85,10 @@ curl https://router.huggingface.co/hf-inference/models/meta-llama/Llama-3.2-11B-
|
|
|
85
85
|
|
|
86
86
|
## Useful Resources
|
|
87
87
|
|
|
88
|
+
- [Vision Language Models (Better, Faster, Stronger)](https://huggingface.co/blog/vlms-2025)
|
|
88
89
|
- [Vision Language Models Explained](https://huggingface.co/blog/vlms)
|
|
89
90
|
- [Welcome PaliGemma 2 – New vision language models by Google](https://huggingface.co/blog/paligemma2)
|
|
90
91
|
- [SmolVLM - small yet mighty Vision Language Model](https://huggingface.co/blog/smolvlm)
|
|
91
92
|
- [Multimodal RAG using ColPali and Qwen2-VL](https://github.com/merveenoyan/smol-vision/blob/main/ColPali_%2B_Qwen2_VL.ipynb)
|
|
92
|
-
- [Image-text-to-text task guide](https://huggingface.co/tasks/image-text-to-text)
|
|
93
93
|
- [Preference Optimization for Vision Language Models with TRL](https://huggingface.co/blog/dpo_vlm)
|
|
94
|
+
- [Image-text-to-text task guide](https://huggingface.co/docs/transformers/tasks/image_text_to_text)
|