Create README.md

## Description
This dataset contains structured GPU hardware specifications
and AI performance benchmark metrics.
It includes information such as VRAM size, CUDA cores,
Tensor cores, FP16 TFLOPS, memory bandwidth,
and estimated AI training and inference scores.

README.md

@@ -0,0 +1,40 @@
+---
+license: mit
+task_categories:
+- tabular-classification
+language:
+- en
+tags:
+- artificial-intelligence
+- gpu
+- deep-learning
+size_categories:
+- n<1K
+---
+
+# AI GPU Benchmark Dataset
+
+## Description
+This dataset contains structured GPU hardware specifications
+and AI performance benchmark metrics.
+
+## Columns
+- gpu_name
+- architecture
+- vram_gb
+- cuda_cores
+- tensor_cores
+- fp16_tflops
+- memory_bandwidth_gbps
+- ai_training_score
+- ai_inference_score
+- power_watt
+
+## Purpose
+This dataset is designed for:
+- AI hardware benchmarking
+- Machine learning research simulation
+- Educational deep learning experiments
+
+## Format
+CSV structured dataset