Speck1.1-140M-Instruct GGUF
llama.cpp-compatible GGUF builds of specklabs/Speck1.1-140M-Instruct, pinned to
source revision 8370432f0f36a6d0b829a50eabfad5476ca37643.
| File | Quantization | Size |
|---|---|---|
| Speck1.1-140M-Instruct-BF16.gguf | BF16 | 361.2 MB |
| Speck1.1-140M-Instruct-Q4_K_M.gguf | Q4_K_M | 112.9 MB |
| Speck1.1-140M-Instruct-Q5_K_M.gguf | Q5_K_M | 130.3 MB |
| Speck1.1-140M-Instruct-Q8_0.gguf | Q8_0 | 192.4 MB |
Usage
llama-cli -hf specklabs/Speck1.1-140M-Instruct-GGUF:Q4_K_M -cnv
The source Speck architecture and llama.cpp's LFM2 runtime implement the same alternating attention/short-convolution operators. Conversion folds the 640-to-768 input and 768-to-640 output adapters into the embeddings, zero-pads the 384-wide convolution channels to 768, and left-pads 3-tap causal kernels to 5 taps. These transformations preserve the model function apart from normal floating-point and quantization rounding.
The GGUF graph stores 180,165,376 parameters because the source's tied 640-wide embedding and two adapters become separate 768-wide input and output matrices. This compatibility transform does not add layers or model capacity.
The conversion was built with llama.cpp revision 2e88c49c90f0add8796f633fea8c3d65b975f295. Exact checksums and
conversion provenance are in conversion.json.