What is Parameter?Parameter kya hai?

🧸 Think of it likeAise samjho

Like the tiny knobs on a giant music mixer: training turns billions of them until the output sounds right.Ek bade music mixer ke chhote knobs jaisa: training arabon knobs ghumaati hai jab tak awaaz sahi na aaye.

📘 ExplainedVistaar se

Parameters (weights and biases) are what the model 'knows'. More parameters can capture more patterns but need more data, memory and compute. '7B' means about 7 billion parameters. Hyperparameters, by contrast, are settings you choose.Parameters (weights aur biases) hi model ki 'jaankari' hain. Zyaada parameters zyaada patterns pakad sakte hain par zyaada data, memory aur compute chahiye. '7B' matlab lagbhag 7 arab parameters. Hyperparameters, iske ulat, aapki chuni settings hain.

💡 Real exampleAsli example

A 7B parameter model needs about 14 GB of memory in 16 bit precision.7B parameter model ko 16 bit precision mein lagbhag 14 GB memory chahiye.

📖 Learn it step by stepStep by step seekho

• Neural networks: layers that learnNeural networks: layers jo seekhti hain
• Serving LLMs efficiently: GPUs, batching and costLLMs ko efficiently serve karna: GPUs, batching aur cost

🔗 Related termsJude hue shabd

• HyperparameterA setting you choose before training, like learning rate or tree depth, rather than something the model learns.Training se pehle chuni setting, jaise learning rate ya tree depth, jo model khud nahi seekhta.
• QuantisationStoring model weights with fewer bits so models use less memory and run faster.Model weights ko kam bits mein rakhna taaki model kam memory le aur tez chale.
• ModelThe learned function that turns inputs into predictions or generated output.Seekha hua function jo inputs ko predictions ya generated output mein badalta hai.