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