MarkBase Admin
ac75faa0cc
CI / build-and-test (push) Has been cancelled
- E4B-MarkBase model (42 layers, 4.4GB) loaded successfully - All Phase 1-6 tests passed (model loading, forward pass, vision/audio towers, token generation, performance) - All stress tests passed (5/5 in 127.6s) - Concurrent inference - Memory stress (67.5 tok/s, 0 NaN) - Continuous generation - Batch processing - Long-running stability - Swift Metal inference engine with multimodal support
110 lines
3.2 KiB
Swift
110 lines
3.2 KiB
Swift
import Foundation
|
|
import MarkBase
|
|
|
|
// ─────────────────────────────────────────────────────────────
|
|
// Model API Models
|
|
// ─────────────────────────────────────────────────────────────
|
|
|
|
/// Chat completion request (for testing)
|
|
public struct ChatCompletionRequest: Codable {
|
|
public let model: String
|
|
public let messages: [ChatMessage]
|
|
public let max_tokens: Int?
|
|
public let temperature: Float?
|
|
public let stream: Bool?
|
|
|
|
public func toGenerationConfig() -> GenerationConfig {
|
|
GenerationConfig(
|
|
maxTokens: max_tokens ?? 100,
|
|
temperature: temperature ?? 1.0
|
|
)
|
|
}
|
|
}
|
|
|
|
/// Multimodal chat completion request (for testing)
|
|
public struct MultimodalChatCompletionRequest: Codable {
|
|
public let model: String
|
|
public let messages: [MultimodalMessage]
|
|
public let max_tokens: Int?
|
|
public let stream: Bool?
|
|
|
|
public func toGenerationConfig() -> GenerationConfig {
|
|
GenerationConfig(maxTokens: max_tokens ?? 100)
|
|
}
|
|
}
|
|
|
|
/// Model capabilities
|
|
public struct ModelCapabilities: Codable, Sendable {
|
|
public let text: Bool
|
|
public let vision: Bool
|
|
public let audio: Bool
|
|
public let embeddings: Bool
|
|
public let streaming: Bool
|
|
|
|
public init(
|
|
text: Bool = true,
|
|
vision: Bool = true,
|
|
audio: Bool = true,
|
|
embeddings: Bool = true,
|
|
streaming: Bool = true
|
|
) {
|
|
self.text = text
|
|
self.vision = vision
|
|
self.audio = audio
|
|
self.embeddings = embeddings
|
|
self.streaming = streaming
|
|
}
|
|
}
|
|
|
|
/// Model parameters
|
|
public struct ModelParameters: Codable, Sendable {
|
|
public let context_length: Int
|
|
public let num_hidden_layers: Int
|
|
public let hidden_size: Int
|
|
public let vocab_size: Int
|
|
public let num_attention_heads: Int
|
|
public let num_kv_heads: Int
|
|
|
|
public init(
|
|
context_length: Int,
|
|
num_hidden_layers: Int,
|
|
hidden_size: Int,
|
|
vocab_size: Int,
|
|
num_attention_heads: Int,
|
|
num_kv_heads: Int
|
|
) {
|
|
self.context_length = context_length
|
|
self.num_hidden_layers = num_hidden_layers
|
|
self.hidden_size = hidden_size
|
|
self.vocab_size = vocab_size
|
|
self.num_attention_heads = num_attention_heads
|
|
self.num_kv_heads = num_kv_heads
|
|
}
|
|
}
|
|
|
|
/// Model details response
|
|
public struct ModelDetails: Codable, Sendable {
|
|
public let id: String
|
|
public let object: String
|
|
public let created: Int
|
|
public let owned_by: String
|
|
public let capabilities: ModelCapabilities
|
|
public let parameters: ModelParameters
|
|
|
|
public init(
|
|
id: String,
|
|
object: String = "model",
|
|
created: Int = Int(Date().timeIntervalSince1970),
|
|
owned_by: String = "markbase",
|
|
capabilities: ModelCapabilities,
|
|
parameters: ModelParameters
|
|
) {
|
|
self.id = id
|
|
self.object = object
|
|
self.created = created
|
|
self.owned_by = owned_by
|
|
self.capabilities = capabilities
|
|
self.parameters = parameters
|
|
}
|
|
}
|