Update README.md

README.md

@@ -43,12 +43,6 @@ For quantized baselines, we use **unsloth/DeepSeek-R1-Distill-Qwen-1.5B-Q8\_0**
 
 <hr style="margin: 4px 0px;">
 
-## Hardware Used
-
-The quantization-aware training (QAT) of the Dheyo models and the benchmarking of all model variants were conducted on a cluster equipped with **AMD Instinct MI300X GPUs**. Each GPU provides 192 GB of HBM3 memory and high memory bandwidth, making them well-suited for running large quantized models efficiently. All evaluations were performed using the **llama.cpp** backend, optimized for low-precision inference.
-
-<hr style="margin: 4px 0px;">
-
 ## Benchmark Results
 
 ### Math500 Pass@1 and GPQA Diamond Pass@1 Benchmarks