⚠️ ARCHIVED / LEGACY MODEL NOTICE
This repository is part of a legacy collection quantized around 2023. To manage storage quotas and maintain active community projects, some rarely used quantization formats (e.g., Q2_K, Q3_K, Q4_1, Q5_1) have been permanently removed.
Only the most popular and stable formats (Q4_0, Q4_K_M, Q5_K_M, Q6_K, and Q8_0) remain available.
💡 Looking for something modern?
If you are starting a new project, we highly recommend using newer architectures (like Llama 3, Mistral, or Qwen) provided by official maintainers or active community members (e.g., Bartowski, TheBloke legacy files, or official organization handles).
⚠️ This repository is no longer actively maintained. Existing files are provided "as is" for archival and legacy hardware purposes.
New releases of Llama.cpp now support K-quantization for previously incompatible models, in particular all Falcon 7B models (While Falcon 40b is and always has been fully compatible with K-Quantisation). This is achieved by employing a fallback solution for model layers that cannot be quantized with real K-quants.
For Falcon 7B models, although only a quarter of the layers can be quantized with true K-quants, this approach still benefits from utilizing different legacy quantization types Q4_0, Q4_1, Q5_0, and Q5_1. As a result, it offers better quality at the same file size or smaller file sizes with comparable performance.
So this solution ensures improved performance and efficiency over legacy Q4_0, Q4_1, Q5_0 and Q5_1 Quantizations.
Brief
Finally got the OpenAssistant falcon sft models working again
gguf is the current file format used by the ggml library.
A growing list of Software is using it and can therefore use this model.
The core project making use of the ggml library is the llama.cpp project by Georgi Gerganov
Quantization variants
There is a bunch of quantized files available to cater to your specific needs. Here's how to choose the best option for you:
Legacy quants
Q4_0, Q4_1, Q5_0, Q5_1 and Q8 are legacy quantization types.
Nevertheless, they are fully supported, as there are several circumstances that cause certain model not to be compatible with the modern K-quants.
Note:
Now there's a new option to use K-quants even for previously 'incompatible' models, although this involves some fallback solution that makes them not real K-quants. More details can be found in affected model descriptions.
(This mainly refers to Falcon 7b and Starcoder models)
K-quants
K-quants are designed with the idea that different levels of quantization in specific parts of the model can optimize performance, file size, and memory load.
So, if possible, use K-quants.
With a Q6_K, you'll likely find it challenging to discern a quality difference from the original model - ask your model two times the same question and you may encounter bigger quality differences.
Original Model Card:
Open-Assistant Falcon 40B SFT MIX Model
This model is a fine-tuning of TII's Falcon 40B LLM.
It was trained on a mixture of OASST top-2 threads (exported on June 2, 2023), Dolly-15k and synthetic instruction datasets (see dataset configuration below).
Two special tokens are used to mark the beginning of user and assistant turns:
<|prompter|> and <|assistant|>. Each turn ends with a <|endoftext|> token.
Input prompt example:
<|prompter|>What is a meme, and what's the history behind this word?<|endoftext|><|assistant|>
The input ends with the <|assistant|> token to signal that the model should
start generating the assistant reply.
Coming Soon: I'm in the process of launching a sponsorship/crowdfunding campaign for my work. I'm evaluating Kickstarter, Patreon, or the new GitHub Sponsors platform, and I am hoping for some support and contribution to the continued availability of these kind of models. Your support will enable me to provide even more valuable resources and maintain the models you rely on. Your patience and ongoing support are greatly appreciated as I work to make this page an even more valuable resource for the community.
