@huggingface/tasks 0.21.4 → 0.21.5

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.
Files changed (8) hide show
  1. package/dist/commonjs/eval.d.ts +10 -0
  2. package/dist/commonjs/eval.d.ts.map +1 -1
  3. package/dist/commonjs/eval.js +10 -0
  4. package/dist/esm/eval.d.ts +10 -0
  5. package/dist/esm/eval.d.ts.map +1 -1
  6. package/dist/esm/eval.js +10 -0
  7. package/package.json +1 -1
  8. package/src/eval.ts +12 -0
package/dist/commonjs/eval.d.ts CHANGED
@@ -32,6 +32,11 @@ export declare const EVALUATION_FRAMEWORKS: {
32
32
  readonly description: "Harbor is a framework for evaluating and optimizing agents and language models.";
33
33
  readonly url: "https://github.com/laude-institute/harbor";
34
34
  };
35
+ readonly pier: {
36
+ readonly name: "pier";
37
+ readonly description: "Pier is a Harbor fork built for DeepSWE, with stronger support for CLI agents in no-internet tasks and more faithful, consistent agent trajectories.";
38
+ readonly url: "https://github.com/datacurve-ai/pier";
39
+ };
35
40
  readonly archipelago: {
36
41
  readonly name: "archipelago";
37
42
  readonly description: "Archipelago is a system for running and evaluating AI agents against MCP applications.";
@@ -112,5 +117,10 @@ export declare const EVALUATION_FRAMEWORKS: {
112
117
  readonly description: "WBench is a comprehensive multi-turn benchmark for interactive video world model evaluation, assessing models across 5 dimensions (video quality, setting adherence, interaction adherence, consistency, physics compliance) and 22 metrics over 289 multi-turn interaction cases.";
113
118
  readonly url: "https://github.com/meituan-longcat/WBench";
114
119
  };
120
+ readonly nanofold: {
121
+ readonly name: "nanofold";
122
+ readonly description: "nanoFold is a data-efficiency benchmark for protein structure prediction. Its goal is to evaluate models on scenarios with scarce data.";
123
+ readonly url: "https://github.com/ChrisHayduk/nanoFold-Competition";
124
+ };
115
125
  };
116
126
  //# sourceMappingURL=eval.d.ts.map
package/dist/commonjs/eval.d.ts.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"eval.d.ts","sourceRoot":"","sources":["../../src/eval.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,eAAO,MAAM,qBAAqB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA4HxB,CAAC"}
1
+ {"version":3,"file":"eval.d.ts","sourceRoot":"","sources":["../../src/eval.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,eAAO,MAAM,qBAAqB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAwIxB,CAAC"}
package/dist/commonjs/eval.js CHANGED
@@ -35,6 +35,11 @@ exports.EVALUATION_FRAMEWORKS = {
35
35
  description: "Harbor is a framework for evaluating and optimizing agents and language models.",
36
36
  url: "https://github.com/laude-institute/harbor",
37
37
  },
38
+ pier: {
39
+ name: "pier",
40
+ description: "Pier is a Harbor fork built for DeepSWE, with stronger support for CLI agents in no-internet tasks and more faithful, consistent agent trajectories.",
41
+ url: "https://github.com/datacurve-ai/pier",
42
+ },
38
43
  archipelago: {
39
44
  name: "archipelago",
40
45
  description: "Archipelago is a system for running and evaluating AI agents against MCP applications.",
@@ -115,4 +120,9 @@ exports.EVALUATION_FRAMEWORKS = {
115
120
  description: "WBench is a comprehensive multi-turn benchmark for interactive video world model evaluation, assessing models across 5 dimensions (video quality, setting adherence, interaction adherence, consistency, physics compliance) and 22 metrics over 289 multi-turn interaction cases.",
116
121
  url: "https://github.com/meituan-longcat/WBench",
117
122
  },
123
+ nanofold: {
124
+ name: "nanofold",
125
+ description: "nanoFold is a data-efficiency benchmark for protein structure prediction. Its goal is to evaluate models on scenarios with scarce data.",
126
+ url: "https://github.com/ChrisHayduk/nanoFold-Competition",
127
+ },
118
128
  };
package/dist/esm/eval.d.ts CHANGED
@@ -32,6 +32,11 @@ export declare const EVALUATION_FRAMEWORKS: {
32
32
  readonly description: "Harbor is a framework for evaluating and optimizing agents and language models.";
33
33
  readonly url: "https://github.com/laude-institute/harbor";
34
34
  };
35
+ readonly pier: {
36
+ readonly name: "pier";
37
+ readonly description: "Pier is a Harbor fork built for DeepSWE, with stronger support for CLI agents in no-internet tasks and more faithful, consistent agent trajectories.";
38
+ readonly url: "https://github.com/datacurve-ai/pier";
39
+ };
35
40
  readonly archipelago: {
36
41
  readonly name: "archipelago";
37
42
  readonly description: "Archipelago is a system for running and evaluating AI agents against MCP applications.";
@@ -112,5 +117,10 @@ export declare const EVALUATION_FRAMEWORKS: {
112
117
  readonly description: "WBench is a comprehensive multi-turn benchmark for interactive video world model evaluation, assessing models across 5 dimensions (video quality, setting adherence, interaction adherence, consistency, physics compliance) and 22 metrics over 289 multi-turn interaction cases.";
113
118
  readonly url: "https://github.com/meituan-longcat/WBench";
114
119
  };
120
+ readonly nanofold: {
121
+ readonly name: "nanofold";
122
+ readonly description: "nanoFold is a data-efficiency benchmark for protein structure prediction. Its goal is to evaluate models on scenarios with scarce data.";
123
+ readonly url: "https://github.com/ChrisHayduk/nanoFold-Competition";
124
+ };
115
125
  };
116
126
  //# sourceMappingURL=eval.d.ts.map
package/dist/esm/eval.d.ts.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"eval.d.ts","sourceRoot":"","sources":["../../src/eval.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,eAAO,MAAM,qBAAqB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA4HxB,CAAC"}
1
+ {"version":3,"file":"eval.d.ts","sourceRoot":"","sources":["../../src/eval.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,eAAO,MAAM,qBAAqB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAwIxB,CAAC"}
package/dist/esm/eval.js CHANGED
@@ -32,6 +32,11 @@ export const EVALUATION_FRAMEWORKS = {
32
32
  description: "Harbor is a framework for evaluating and optimizing agents and language models.",
33
33
  url: "https://github.com/laude-institute/harbor",
34
34
  },
35
+ pier: {
36
+ name: "pier",
37
+ description: "Pier is a Harbor fork built for DeepSWE, with stronger support for CLI agents in no-internet tasks and more faithful, consistent agent trajectories.",
38
+ url: "https://github.com/datacurve-ai/pier",
39
+ },
35
40
  archipelago: {
36
41
  name: "archipelago",
37
42
  description: "Archipelago is a system for running and evaluating AI agents against MCP applications.",
@@ -112,4 +117,9 @@ export const EVALUATION_FRAMEWORKS = {
112
117
  description: "WBench is a comprehensive multi-turn benchmark for interactive video world model evaluation, assessing models across 5 dimensions (video quality, setting adherence, interaction adherence, consistency, physics compliance) and 22 metrics over 289 multi-turn interaction cases.",
113
118
  url: "https://github.com/meituan-longcat/WBench",
114
119
  },
120
+ nanofold: {
121
+ name: "nanofold",
122
+ description: "nanoFold is a data-efficiency benchmark for protein structure prediction. Its goal is to evaluate models on scenarios with scarce data.",
123
+ url: "https://github.com/ChrisHayduk/nanoFold-Competition",
124
+ },
115
125
  };
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@huggingface/tasks",
3
- "version": "0.21.4",
3
+ "version": "0.21.5",
4
4
  "description": "List of ML tasks for huggingface.co/tasks",
5
5
  "keywords": [
6
6
  "hub",
package/src/eval.ts CHANGED
@@ -33,6 +33,12 @@ export const EVALUATION_FRAMEWORKS = {
33
33
  description: "Harbor is a framework for evaluating and optimizing agents and language models.",
34
34
  url: "https://github.com/laude-institute/harbor",
35
35
  },
36
+ pier: {
37
+ name: "pier",
38
+ description:
39
+ "Pier is a Harbor fork built for DeepSWE, with stronger support for CLI agents in no-internet tasks and more faithful, consistent agent trajectories.",
40
+ url: "https://github.com/datacurve-ai/pier",
41
+ },
36
42
  archipelago: {
37
43
  name: "archipelago",
38
44
  description: "Archipelago is a system for running and evaluating AI agents against MCP applications.",
@@ -125,4 +131,10 @@ export const EVALUATION_FRAMEWORKS = {
125
131
  "WBench is a comprehensive multi-turn benchmark for interactive video world model evaluation, assessing models across 5 dimensions (video quality, setting adherence, interaction adherence, consistency, physics compliance) and 22 metrics over 289 multi-turn interaction cases.",
126
132
  url: "https://github.com/meituan-longcat/WBench",
127
133
  },
134
+ nanofold: {
135
+ name: "nanofold",
136
+ description:
137
+ "nanoFold is a data-efficiency benchmark for protein structure prediction. Its goal is to evaluate models on scenarios with scarce data.",
138
+ url: "https://github.com/ChrisHayduk/nanoFold-Competition",
139
+ },
128
140
  } as const;