Merge branch 'main' into add-video-to-video

Author: merveenoyan

packages/hub/src/index.ts

@@ -18,6 +18,7 @@ export type {
 	SpaceStage,
 } from "./types/public";
 export { HubApiError, InvalidApiResponseFormatError } from "./error";
+export { HUB_URL } from "./consts";
 /**
  * Only exported for E2Es convenience
  */

packages/inference/src/providers/featherless-ai.ts

@@ -7,6 +7,7 @@ import type {
 import { InferenceOutputError } from "../lib/InferenceOutputError.js";
 import type { BodyParams } from "../types.js";
 import { BaseConversationalTask, BaseTextGenerationTask } from "./providerHelper.js";
+import { omit } from "../utils/omit.js";
 
 interface FeatherlessAITextCompletionOutput extends Omit<ChatCompletionOutput, "choices"> {
 	choices: Array<{
@@ -33,9 +34,14 @@ export class FeatherlessAITextGenerationTask extends BaseTextGenerationTask {
 
 	override preparePayload(params: BodyParams<TextGenerationInput>): Record<string, unknown> {
 		return {
-			...params.args,
-			...params.args.parameters,
 			model: params.model,
+			...omit(params.args, ["inputs", "parameters"]),
+			...(params.args.parameters
+				? {
+						max_tokens: params.args.parameters.max_new_tokens,
+						...omit(params.args.parameters, "max_new_tokens"),
+				  }
+				: undefined),
 			prompt: params.args.inputs,
 		};
 	}

packages/ollama-utils/src/chat-template-automap.ts

@@ -5,19 +5,9 @@ import type { OllamaChatTemplateMapEntry } from "./types";
 
 /**
  * Skipped these models due to error:
+ * - library/llama4:latest
  * - library/llama3.1:latest
- * - library/gemma3:latest
- * - library/llama3.3:70b
- * - library/command-r-plus:latest
- * - library/codeqwen:latest
- * - library/bakllava:latest
- * - library/command-r-plus:latest
- * - library/qwen2-math:1.5b
- * - library/reflection:latest
- * - library/deepscaler:1.5b
- * - library/nemotron:latest
- * - library/nemotron:70b
- * - library/tulu3:latest
+ * - library/granite3.2:latest
  */
 
 export const OLLAMA_CHAT_TEMPLATE_MAPPING: OllamaChatTemplateMapEntry[] = [
@@ -337,6 +327,15 @@ export const OLLAMA_CHAT_TEMPLATE_MAPPING: OllamaChatTemplateMapEntry[] = [
 			},
 		},
 	},
+	{
+		model: "library/devstral:latest",
+		gguf: "{%- set today = strftime_now(\"%Y-%m-%d\") %}\n{%- set default_system_message = \"You are Devstral, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\\nYour knowledge base was last updated on 2023-10-01. The current date is \" + today + \".\\n\\nWhen you're not sure about some information, you say that you don't have the information and don't make up anything.\\nIf the user's question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \\\"What are some good restaurants around me?\\\" => \\\"Where are you?\\\" or \\\"When is the next flight to Tokyo\\\" => \\\"Where do you travel from?\\\")\" %}\n\n{{- bos_token }}\n\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content'] %}\n    {%- set loop_messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = default_system_message %}\n    {%- set loop_messages = messages %}\n{%- endif %}\n{{- '[SYSTEM_PROMPT]' + system_message + '[/SYSTEM_PROMPT]' }}\n\n{%- for message in loop_messages %}\n    {%- if message['role'] == 'user' %}\n        {{- '[INST]' + message['content'] + '[/INST]' }}\n    {%- elif message['role'] == 'system' %}\n        {{- '[SYSTEM_PROMPT]' + message['content'] + '[/SYSTEM_PROMPT]' }}\n    {%- elif message['role'] == 'assistant' %}\n        {{- message['content'] + eos_token }}\n    {%- else %}\n        {{- raise_exception('Only user, system and assistant roles are supported!') }}\n    {%- endif %}\n{%- endfor %}",
+		ollama: {
+			template:
+				'{{- range $index, $_ := .Messages }}\n{{- if eq .Role "system" }}[SYSTEM_PROMPT]{{ .Content }}[/SYSTEM_PROMPT]\n{{- else if eq .Role "user" }}\n{{- if and (le (len (slice $.Messages $index)) 2) $.Tools }}[AVAILABLE_TOOLS]{{ $.Tools }}[/AVAILABLE_TOOLS]\n{{- end }}[INST]{{ .Content }}[/INST]\n{{- else if eq .Role "assistant" }}\n{{- if .Content }}{{ .Content }}\n{{- if not (eq (len (slice $.Messages $index)) 1) }}</s>\n{{- end }}\n{{- else if .ToolCalls }}[TOOL_CALLS][\n{{- range .ToolCalls }}{"name": "{{ .Function.Name }}", "arguments": {{ .Function.Arguments }}}\n{{- end }}]</s>\n{{- end }}\n{{- else if eq .Role "tool" }}[TOOL_RESULTS]{"content": {{ .Content }}}[/TOOL_RESULTS]\n{{- end }}\n{{- end }}',
+			tokens: ["[INST]"],
+		},
+	},
 	{
 		model: "library/dolphin-llama3:8b",
 		gguf: "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}",

packages/tasks-gen/scripts/generate-snippets-fixtures.ts

@@ -19,6 +19,7 @@ import { existsSync as pathExists } from "node:fs";
 import * as fs from "node:fs/promises";
 import * as path from "node:path/posix";
 
+import type { InferenceProviderOrPolicy } from "@huggingface/inference";
 import { snippets } from "@huggingface/inference";
 import type { InferenceSnippet, ModelDataMinimal, SnippetInferenceProvider, WidgetType } from "@huggingface/tasks";
 import { inferenceSnippetLanguages } from "@huggingface/tasks";
@@ -31,7 +32,7 @@ const TEST_CASES: {
 	testName: string;
 	task: WidgetType;
 	model: ModelDataMinimal;
-	providers: SnippetInferenceProvider[];
+	providers: InferenceProviderOrPolicy[];
 	lora?: boolean;
 	opts?: snippets.InferenceSnippetOptions;
 }[] = [
@@ -306,7 +307,7 @@ function getFixtureFolder(testName: string): string {
 function generateInferenceSnippet(
 	model: ModelDataMinimal,
 	language: Language,
-	provider: SnippetInferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: WidgetType,
 	lora: boolean = false,
 	opts?: Record<string, unknown>

packages/tasks/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "@huggingface/tasks",
 	"packageManager": "[email protected]",
-	"version": "0.19.8",
+	"version": "0.19.9",
 	"description": "List of ML tasks for huggingface.co/tasks",
 	"repository": "https://github.com/huggingface/huggingface.js.git",
 	"publishConfig": {

packages/tasks/src/hardware.ts

@@ -100,6 +100,10 @@ export const SKUS = {
 				tflops: 38.7,
 				memory: [48],
 			},
+			"RTX A5000": {
+				tflops: 34.1,
+				memory: [24],
+			},
 			"RTX A4000": {
 				tflops: 19.2,
 				memory: [16],
@@ -533,6 +537,9 @@ export const SKUS = {
 			"Intel Core Ultra 7 265KF": {
 				tflops: 1.53,
 			},
+			"Intel Core 14th Generation (i7)": {
+				tflops: 0.8,
+			},
 			"Intel Core 13th Generation (i9)": {
 				tflops: 0.85,
 			},

packages/tasks/src/model-libraries.ts

@@ -936,7 +936,7 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
 		repoName: "timesfm",
 		repoUrl: "https://github.com/google-research/timesfm",
 		filter: false,
-		countDownloads: `path:"checkpoints/checkpoint_1100000/state/checkpoint"`,
+		countDownloads: `path:"checkpoints/checkpoint_1100000/state/checkpoint" OR path:"checkpoints/checkpoint_2150000/state/checkpoint" OR path_extension:"ckpt"`,
 	},
 	timm: {
 		prettyLabel: "timm",