Views
No views yet
freakyskittle/kimi-k2.75-code
(a shard-wise SLERP merge of moonshotai/Kimi-K2.7-Code and moonshotai/Kimi-K2.6,
DeepSeek-V3 style MoE).Parameter count: ~720 B. The bf16 and Q8_0 builds contain 1,096 tensors summing to 719.8 B parameters. Hugging Face's auto-detected badge reads ~61 B because it parses thedeep55/file first, and that build stores weights in a packed-int4 layout (int32weight_packedtensors holding 8 weights each, plusweight_scale/weight_shape) whose logical shapes can't be summed. The true model is ~720 B params (further reduced in the deep-pruneddeep55/variant).
-0000N-of-0000M.gguf); point your loader at the
first shard and the rest are picked up automatically.| Folder | Variant | Approx. size | Notes |
|---|---|---|---|
deep55/ | Deep-pruned, full precision | ~203 GB | Prune ratio 0.55, deepseek_v3 arch override |
pruned-compact-oxidize-q4/ | Q4_K_M (oxidize) | ~406 GB | Quantized from the compacted pruned checkpoint |
llamacpp-q4-partial/ | Q4_K_M (llama.cpp, partial) | ~435 GB | Partial llama.cpp quantization |
unpruned-q8/ | Q8_0 (unpruned) | ~765 GB | Split into 2 shards |
pruned-bf16/ | BF16 (pruned) | ~1.44 TB | Split into 4 shards |
freakyskittle/kimi-k2.75-code.oxidize-convert with --arch deepseek_v3.llama-gguf-split (--split-max-size 450G).Kimi K2.7 Code in the UI.