Evaluate Hebrew STT models using WER
QuantumLLMInstruct: A 500k LLM Instruction-Tuning Dataset
Analyze images and generate detailed descriptions or captions