LLMFarm_core swift library to work with large language models (LLM). It allows you to load different LLMs with certain parameters.
Based on ggml and llama.cpp by Georgi Gerganov.
Also used sources from:
- rwkv.cpp by saharNooby.
- Mia by byroneverson.
- MacOS (13+)
- iOS (16+)
- Various inferences
- Various sampling methods
- Metal (dont work on intel Mac)
- Model setting templates
- LoRA adapters support (read more)
- LoRA train support
- LoRA export as model support
- Restore context state (now only chat history)
- LLaMA
- GPTNeoX
- Replit
- GPT2 + Cerebras
- Starcoder(Santacoder)
- RWKV (20B tokenizer)
- Falcon
- MPT
- Bloom
- StableLM-3b-4e1t
- Qwen
- Temperature (temp, tok-k, top-p)
- Tail Free Sampling (TFS)
- Locally Typical Sampling
- Mirostat
- Greedy
- Grammar (dont work for GPTNeoX, GPT-2, RWKV)
- Classifier-Free Guidance
git clone https://github.com/guinmoon/llmfarm_core.swift
Add llmfarm_core to your project using Xcode (File > Add Packages...) or by adding it to your project's Package.swift file:
dependencies: [
.package(url: "https://github.com/guinmoon/llmfarm_core.swift")
]To Debug llmfarm_core package, do not forget to comment .unsafeFlags(["-Ofast"]) in Package.swift.
Don't forget that the debug version is slower than the release version.
To build with QKK_64 support uncomment .unsafeFlags(["-DGGML_QKK_64"]) in Package.swift.
import Foundation
import llmfarm_core
let maxOutputLength = 256
var total_output = 0
func mainCallback(_ str: String, _ time: Double) -> Bool {
print("\(str)",terminator: "")
total_output += str.count
if(total_output>maxOutputLength){
return true
}
return false
}
var input_text = "State the meaning of life."
let ai = AI(_modelPath: "llama-2-7b.q4_K_M.gguf",_chatName: "chat")
var params:ModelContextParams = .default
params.use_metal = true
try? ai.loadModel(ModelInference.LLama_gguf,contextParams: params)
ai.model.promptFormat = .LLaMa
let output = try? ai.model.predict(input_text, mainCallback)
App to run LLaMA and other large language models locally on iOS and MacOS.