ioworker0/DeepSeek-V4-Flash-0731-REAP
Full fine-tune (fewer layers) vs unsloth/DeepSeek-V4-Flash-0731-GGUF. unsloth/DeepSeek-V4-Flash-0731-GGUF matches part of ioworker0/DeepSeek-V4-Flash-0731-REAP (73 of its 1328 tensors; the rest of ioworker0/DeepSeek-V4-Flash-0731-REAP has no counterpart). On the shared part: ioworker0/DeepSeek-V4-Flash-0731-REAP is a GGUF (IQ3_XXS/IQ1_S) of different weights. Fine-tune: 100% of the weights changed, across 0 of 0 layers.
Tensors changed: MLP 30/30, Attention 27/27, Norms 13/13, other 3/3