add perplexity comparison with intel auto-round quant

still testing vulkan backend added note

add perplexity graph

Add small GPU optimized KT quants and quickstart

Upload Qwen3-30B-A3B-Instruct-2507-IQ2_KT.gguf with huggingface_hub

Upload Qwen3-30B-A3B-Instruct-2507-IQ1_KT.gguf with huggingface_hub

Upload Qwen3-30B-A3B-Instruct-2507-IQ2_KL.gguf with huggingface_hub

Add IQ2_KL

Upload Qwen3-30B-A3B-Instruct-2507-IQ3_KS.gguf with huggingface_hub

Upload Qwen3-30B-A3B-Instruct-2507-IQ3_K.gguf with huggingface_hub

Add IQ3_K and IQ3_KS

Upload Qwen3-30B-A3B-Instruct-2507-IQ4_KSS.gguf with huggingface_hub

Upload Qwen3-30B-A3B-Instruct-2507-IQ4_K.gguf with huggingface_hub

Add IQ4_K and IQ4_KSS

Upload imatrix-eaddario-combined-all-medium-Qwen3-30B-A3B-Instruct-2507-BF16.dat with huggingface_hub

Upload Qwen3-30B-A3B-Instruct-2507-IQ5_K.gguf with huggingface_hub

upload IQ5_K

initial commit

initial commit