quantization_options = [
"Q4_K_M", "Q4_K_S", "IQ4_NL", "IQ4_XS", "Q5_K_M",
"Q5_K_S", "Q6_K", "Q8_0", "IQ3_M", "IQ3_S", "IQ3_XS", "IQ3_XXS"
]
All credits belong to the author.
If you liked these, check out the work with FantasiaFoundry's GGUF-IQ-Imatrix-Quantization-Script.
It stands for Importance Matrix, a technique used to improve the quality of quantized models.
[1]
The Imatrix is calculated based on calibration data, and it helps determine the importance of different model activations during the quantization process. The idea is to preserve the most important information during quantization, which can help reduce the loss of model performance and lead to better quality preservation, especially when the calibration data is diverse.
[2]
For --imatrix data, included imatrix.dat was used.
Using llama.cpp-b2343:
Base⇢ GGUF(F16)⇢ Imatrix-Data(F16)⇢ GGUF(Imatrix-Quants)
The new IQ3_S quant-option has shown to be better than the old Q3_K_S, so I added that instead of the later. Only supported in koboldcpp-1.59.1 or higher.
If you want any specific quantization to be added, feel free to ask.
The following YAML configuration was used to produce this model:
slices:
- sources:
- model: SpecialEdition
layer_range: [0, 32]
- model: Remix
layer_range: [0, 32]
merge_method: slerp
base_model: SpecialEdition
parameters:
t:
- filter: self_attn
value: [0, 0.5, 0.3, 0.7, 1]
- filter: mlp
value: [1, 0.5, 0.7, 0.3, 0]
- value: 0.5
dtype: bfloat16