Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/ptsochantaris/emeltal

Local ML voice chat using high-end models.
https://github.com/ptsochantaris/emeltal

ai llama-cpp machine-learning macos ml natural-language-processing speech-recognition swift swiftui user-interface whisper-cpp

Last synced: 4 days ago
JSON representation

Local ML voice chat using high-end models.

Host: GitHub
URL: https://github.com/ptsochantaris/emeltal
Owner: ptsochantaris
License: mit
Created: 2023-12-17T14:18:07.000Z (11 months ago)
Default Branch: main
Last Pushed: 2024-11-03T13:39:45.000Z (8 days ago)
Last Synced: 2024-11-03T14:27:31.369Z (8 days ago)
Topics: ai, llama-cpp, machine-learning, macos, ml, natural-language-processing, speech-recognition, swift, swiftui, user-interface, whisper-cpp
Language: C++
Homepage:
Size: 47.8 MB
Stars: 142
Watchers: 3
Forks: 8
Open Issues: 1
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

awesome-local-llms - emeltal - end models. | 142 | 8 | 1 | 1 | 0 | MIT License | 5 days, 5 hrs, 5 mins | (Open-Source Local LLM Projects)

README

        

Emeltal

====

_The wise cheese_

Local ML voice chat using high-end models, aiming for a self contained, user-friendly out-of-the-box experience as much as possible.

This is a work in progress with frequent updates; [TestFlight builds are available here](https://testflight.apple.com/join/NTIomxyk) for macOS, iOS and visionOS.

|Selection|Full|Mini|

|---------|----|----|

||||



## Emellink

A light helper app which can run on an iPhone or device with not enough processing power, which automatically detects and connects to Emeltal on the network and provides the same voice interface. [Testflight link for this app is here](https://testflight.apple.com/join/s0EYVO5P)

## Currently supported models

Emeltal offers a curated list of proven open-source high-performance models, aiming to provide the best model for each category/size combination. This list often changes as new models become available, or others are superceeded by much better performing ones. Most models (with the exception of certain extremely large variants, which are capped at 16384 tokens) run at their maximum context size.

#### Qwen Series

- [Qwen 2.5 72b] (https://huggingface.co/bartowski/Qwen2.5-72B-Instruct-GGUF)

- [Qwen 2.5 32b] (https://huggingface.co/bartowski/Qwen2.5-32B-Instruct-GGUF)

- [Qwen 2.5 14b] (https://huggingface.co/Qwen/Qwen2.5-14B-Instruct-GGUF)

- [Qwen 2.5 7b] (https://huggingface.co/bartowski/Qwen2.5-7B-Instruct-GGUF)

#### Dolphin Series

- [Dolphin 2.9.2 on Qwen 2.5](https://huggingface.co/mradermacher/dolphin-2.9.2-qwen2-72b-i1-GGUF)

- [Dolphin 2.7 on Mixtral](https://huggingface.co/cognitivecomputations/dolphin-2.7-mixtral-8x7b)

- [Dolphin 2.9.3 on Mistral Nemo](https://huggingface.co/cognitivecomputations/dolphin-2.9.3-mistral-nemo-12b-gguf)

- [Dolphin 2.8.1 on TinyLlama](https://huggingface.co/cognitivecomputations/TinyDolphin-2.8-1.1b)

#### Samantha Series

- [Samantha 1.11 70b](https://huggingface.co/cognitivecomputations/Samantha-1.11-70b)

- [Samantha 1.1 7b](https://huggingface.co/cognitivecomputations/samantha-1.1-westlake-7b)

#### Llama Series

- [Llama 3.1 70b](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct)

- [Llama 3.1 8b](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct)

- [Llama 3.2 3b](https://huggingface.co/meta-llama/Meta-Llama-3.2-8B-Instruct)

- [Llama 3.2 1b](https://huggingface.co/meta-llama/Meta-Llama-3.2-8B-Instruct)

#### Coding

- [Dolphin Coder](https://huggingface.co/cognitivecomputations/dolphincoder-starcoder2-15b)

- [Deepseek Coder 33b](https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct)

- [Deepseek Coder 7b](https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct-v1.5)

- [Everyone Coder](https://huggingface.co/rombodawg/Everyone-Coder-33b-v2-Base)

- [CodeLlama 70b](https://huggingface.co/codellama/CodeLlama-70b-Instruct-hf)

- [Codestral](https://huggingface.co/mistralai/Codestral-22B-v0.1)

#### Creative

- [MythoMax 13b](https://huggingface.co/Gryphe/MythoMax-L2-13b)

- [Neural Story](https://huggingface.co/NeuralNovel/Mistral-7B-Instruct-v0.2-Neural-Story)

#### Other

- [Supernova Medius](https://huggingface.co/arcee-ai/SuperNova-Medius)

- [Shuttle 3](https://huggingface.co/shuttleai/shuttle-3)

- [SmolLM 2](https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct)

#### Voice Recognition

- [Whisper](https://huggingface.co/ggerganov/whisper.cpp)

## Packages

- Emeltal heavily relies on the [llama.cpp](https://github.com/ggerganov/llama.cpp) for LLM processing, and [whisper.cpp](https://github.com/ggerganov/whisper.cpp) for voice recognition.

- Text rendering uses [Ink](https://github.com/JohnSundell/Ink) to convert between Markdown and HTML.

- Uses my [PopTimer](https://github.com/ptsochantaris/pop-timer) for debouncing things.

## License

Released under the terms of the MIT license, see the [LICENSE](LICENSE.txt) file for license rights and limitations (MIT).

All model data which is downloaded locally by the app comes from HuggingFace, and use of the models and data is subject to the respective license of each specific model.

## Copyright

Copyright (c) 2023-2024 Paul Tsochantaris