/**
 * WebInferContext manages WebGPU device and provides core functionality
 */
import type { DType } from './types.ts';
import { Tensor } from './tensor.ts';
/**
 * Device vendor detection
 */
export type Vendor = 'apple' | 'nvidia' | 'amd' | 'intel' | 'unknown';
/**
 * Device features detected at runtime
 */
export interface DeviceFeatures {
    readonly vendor: Vendor;
    readonly hasSubgroups: boolean;
    readonly hasF16: boolean;
    readonly maxWorkgroupSize: number;
    readonly maxStorageBufferSize: number;
    readonly maxBufferSize: number;
    readonly maxWorkgroupsPerDimension: number;
}
/**
 * WebInferContext - main context for all WebInfer operations
 */
export declare class WebInferContext {
    readonly device: GPUDevice;
    readonly features: DeviceFeatures;
    private pipelineCache;
    private disposed;
    private constructor();
    /**
     * Create a WebInferContext from an existing GPUDevice.
     * Use this when integrating with TVM or other WebGPU frameworks
     * that already have a GPUDevice to enable zero-copy buffer sharing.
     *
     * @param device - The existing GPUDevice to use
     * @param featureOverrides - Optional overrides for detected features
     * @returns A new WebInferContext using the provided device
     */
    static fromDevice(device: GPUDevice, featureOverrides?: Partial<DeviceFeatures>): WebInferContext;
    /**
     * Create a new WebInferContext with its own GPUDevice
     */
    static create(): Promise<WebInferContext>;
    /**
     * Create a GPU buffer
     */
    createBuffer(size: number, usage?: GPUBufferUsageFlags): GPUBuffer;
    /**
     * Create a tensor with uninitialized data
     */
    createTensor(shape: number[], dtype: DType): Tensor;
    /**
     * Create a tensor from TypedArray data
     */
    createTensorFromData(data: Float32Array | Uint16Array, shape: number[], dtype: DType): Tensor;
    /**
     * Get or create a cached compute pipeline
     */
    getOrCreatePipeline(key: string, createFn: () => Promise<GPUComputePipeline> | GPUComputePipeline): Promise<GPUComputePipeline>;
    /**
     * Clear pipeline cache
     */
    clearPipelineCache(): void;
    /**
     * Dispose the context and cleanup resources
     */
    dispose(): void;
    private checkDisposed;
}