Bahushruth/Qwen3.6-35B-A3B-abliterated-v4-GGUF

Fix: IQ2_M without MTP (compatible with all runtimes)

Update model card: fix MTP runtime compatibility note

Fix: Q2_K without MTP (compatible with all runtimes)

Fix: IQ3_XXS without MTP (compatible with all runtimes)

Fix: IQ3_M without MTP (compatible with all runtimes)

Fix: Q3_K_M without MTP (compatible with all runtimes)

Fix: IQ4_NL without MTP (compatible with all runtimes)

Fix: IQ4_XS without MTP (compatible with all runtimes)

Fix: Q4_K_M without MTP (compatible with all runtimes)

Fix: Q5_K_M without MTP (compatible with all runtimes)

Fix: Q6_K without MTP (compatible with all runtimes)

Fix: Q8_0 without MTP (compatible with all runtimes)

Add BF16-MTP variant (for llama-server with --spec-type draft-mtp)

Fix: BF16 without MTP (compatible with Ollama/LMStudio/KoboldCPP)

Fix model card: clarify no-MTP default, MTP as separate file for advanced users

Update model card: add all quants, MTP support, imatrix details

Update BF16 GGUF (with MTP)

Update Q2_K quantization (with MTP)

Update IQ3_M quantization (with MTP)

Update Q3_K_M quantization (with MTP)

Update IQ4_NL quantization (with MTP)

Update IQ4_XS quantization (with MTP)

Update Q4_K_M quantization (with MTP)

Update Q5_K_M quantization (with MTP)

Update Q6_K quantization (with MTP)

Update Q8_0 quantization (with MTP)

Add IQ2_M quantization (imatrix)

Add Q2_K quantization (imatrix)

Add IQ3_XXS quantization (imatrix)

Add IQ3_M quantization (imatrix)

Add Q3_K_M quantization (imatrix)

Add IQ4_NL quantization (imatrix)

Add IQ4_XS quantization (imatrix)

Add model card with quantization guide

v4: Q4_K_M GGUF - 1-dir norm-preserving abliteration

v4: Q5_K_M GGUF - 1-dir norm-preserving abliteration

v4: Q6_K GGUF - 1-dir norm-preserving abliteration

v4: Q8_0 GGUF - 1-dir norm-preserving abliteration

v4: BF16 GGUF

initial commit