Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: RekaAI/reka-flash-3
|
| 3 |
+
pipeline_tag: text-generation
|
| 4 |
+
inference: true
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
license: apache-2.0
|
| 8 |
+
model_creator: RekaAI
|
| 9 |
+
model_name: reka-flash-3
|
| 10 |
+
model_type: llama
|
| 11 |
+
quantized_by: brittlewis12
|
| 12 |
+
tags:
|
| 13 |
+
- reasoning
|
| 14 |
+
- llama
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
# Reka Flash 3 (21B) GGUF
|
| 18 |
+
|
| 19 |
+
**Original model**: [Reka Flash 3](https://huggingface.co/RekaAI/reka-flash-3)
|
| 20 |
+
|
| 21 |
+
**Model creator**: [RekaAI](https://huggingface.co/RekaAI)
|
| 22 |
+
|
| 23 |
+
> Reka Flash 3 is a 21B general-purpose reasoning model that was trained from scratch. It was trained in synthetic and public datasets for supervised finetuning, followed by RLOO with model-based and rule-based rewards. It performs competitively with proprietary models such as OpenAI o1-mini, making it a good foundation to build applications that require low latency or on-device deployment.
|
| 24 |
+
|
| 25 |
+
This repo contains GGUF format model files for RekaAI’s Reka Flash 3.
|
| 26 |
+
|
| 27 |
+
### What is GGUF?
|
| 28 |
+
|
| 29 |
+
GGUF is a file format for representing AI models. It is the third version of the format,
|
| 30 |
+
introduced by the llama.cpp team on August 21st 2023.
|
| 31 |
+
|
| 32 |
+
Converted with llama.cpp build b4872 (revision [ba76543](https://github.com/ggml-org/llama.cpp/commits/ba7654380a3c7c1b5ae154bea19134a3a9417a1e)),
|
| 33 |
+
using [autogguf-rs](https://github.com/brittlewis12/autogguf-rs).
|
| 34 |
+
|
| 35 |
+
### Prompt template: [Reka 3 Chat](https://huggingface.co/RekaAI/reka-flash-3/blob/5ebbf19d65a3d4e41c6dcb87cf61bcfbf0d08e0f/tokenizer_config.json#L197)
|
| 36 |
+
|
| 37 |
+
```
|
| 38 |
+
human: {{system_message}} {{prompt}} <sep> assistant:
|
| 39 |
+
```
|
| 40 |
+
|
| 41 |
+
---
|
| 42 |
+
|
| 43 |
+
## Download & run with [cnvrs](https://twitter.com/cnvrsai) on iPhone, iPad, and Mac!
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
[cnvrs](https://testflight.apple.com/join/sFWReS7K) is the best app for private, local AI on your device:
|
| 48 |
+
- create & save **Characters** with custom system prompts & temperature settings
|
| 49 |
+
- download and experiment with any **GGUF model** you can [find on HuggingFace](https://huggingface.co/models?library=gguf)!
|
| 50 |
+
* or, use an API key with the chat completions-compatible model provider of your choice -- ChatGPT, Claude, Gemini, DeepSeek, & more!
|
| 51 |
+
- make it your own with custom **Theme colors**
|
| 52 |
+
- powered by Metal ⚡️ & [Llama.cpp](https://github.com/ggml-org/llama.cpp), with **haptics** during response streaming!
|
| 53 |
+
- **try it out** yourself today, on [Testflight](https://testflight.apple.com/join/sFWReS7K)!
|
| 54 |
+
* if you **already have the app**, download Reka Flash 3 now!
|
| 55 |
+
* <cnvrsai:///models/search/hf?id=brittlewis12/reka-flash-3-GGUF>
|
| 56 |
+
- follow [cnvrs on twitter](https://twitter.com/cnvrsai) to stay up to date
|
| 57 |
+
|
| 58 |
+
### Reka Flash 3 in cnvrs on macOS
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
---
|
| 63 |
+
|
| 64 |
+
## Original Model Evaluation
|
| 65 |
+
|
| 66 |
+
|