Multimodal AI Interview Questions Hindi mein

जवाब: Answering "how many people are wearing helmets?" about a photo

Visual question answering needs the model to connect a text question with pixel content. The other three are text-only tasks.

जवाब: They operate on any sequence of vectors, whatever modality produced it

Self-attention works on sequences of embeddings. Once images or audio are cut into vectors, the same layers apply.

जवाब: Late fusion of the two models' scores

Late fusion reuses both models, needs little paired data and degrades gracefully when one modality is absent.

जवाब: The colour channels (R, G, B)

In NCHW layout the second dimension is channels; an RGB image has three.

जवाब: 160,000

Sample rate times duration: 16,000 samples per second x 10 seconds = 160,000 samples.

जवाब: Adjacent frames are highly redundant and every frame costs tokens and compute

Consecutive frames are almost identical, so sampling cuts cost sharply while keeping most of the information.

जवाब: It outputs a sequence of token vectors

A ViT produces one vector per patch, a token sequence that a projector can map into the language model's embedding space.

जवाब: 576

336 / 14 = 24 patches per side, and 24 x 24 = 576.

जवाब: A log-mel spectrogram of 16 kHz audio

Whisper resamples to 16 kHz and computes a log-mel spectrogram, processed in 30-second windows.

जवाब: They point in exactly the same direction

Cosine similarity is 1 when the angle between the vectors is zero, i.e. they point the same way.

जवाब: In the other captions in the same batch

Every other caption in the batch serves as a negative, which is why larger batches help.

जवाब: A text prompt describing the class

Zero-shot classes are defined by text prompts; the text tower embeds any new description.

जवाब: Map vision-encoder outputs into the language model's embedding space

The projector converts patch vectors to the dimensionality and space the LLM's token embeddings use.

जवाब: The text tokens

Text tokens act as queries; image features supply keys and values, so each text token gathers relevant visual information.

जवाब: Only the projector

Stage 1 freezes the encoder and LLM and trains just the projector on image-caption pairs.

जवाब: The noise that was added at step t

The standard objective is to predict the added noise, trained with a mean squared error loss.

जवाब: The latent is much smaller than the pixel image, so each step is cheaper

Compressing to a small latent makes every denoising step far cheaper; the VAE decoder restores full resolution at the end.

जवाब: Latency adds up and the LLM loses tone and emotion from the speech

Each stage adds delay, and converting speech to text discards paralinguistic information before the LLM sees it.

जवाब: Smaller images upload faster, fit size limits and usually cost fewer image tokens

Large images waste bandwidth, can exceed limits, and often cost more tokens without improving answers for typical questions.

जवाब: Prepares images, text and chat templates exactly as that checkpoint expects

The processor bundles the image preprocessing, tokenizer and chat template matched to the checkpoint.

जवाब: The vision encoder and projector that turn images into tokens for the LLM

The multimodal projector file holds the vision side; it must match the language model weights.

जवाब: Details the caption omitted cannot be retrieved

Retrieval only sees the generated caption, so anything the captioner skipped is invisible to search.

जवाब: Verifying that extracted line items sum to the extracted total

Deterministic business-rule checks catch inconsistent values; asking the model to self-assess is unreliable.

जवाब: Too few or no frames are being sent to the model for that segment

Visual questions need frames; a transcript-heavy pipeline with sparse or missing frames cannot see the screen.

जवाब: It quadruples

Patch count is (H/P) x (W/P), so doubling both sides multiplies it by four.

जवाब: 0.3

WER = (S + D + I) / N = (1 + 1 + 1) / 10 = 0.3.

जवाब: Precompute and store image embeddings once; embed only the query at search time

Catalogue embeddings are static; compute them at ingest and only encode the query text online.

जवाब: Send a higher-resolution crop of the total region and allow a 'not readable' answer

Detail errors usually come from downsampling; a crop restores resolution and an escape option prevents guessing.

जवाब: Require validation against business rules or human approval before any payment

Prompt instructions alone do not reliably stop injection; gating privileged actions limits the damage when the model is fooled.

जवाब: EXIF may contain GPS location and device details that reveal personal information

EXIF often includes location and device data; removing it protects users when photos are stored or shared.

जवाब: CLIP - contrastive loss; diffusion - noise prediction

CLIP learns with a contrastive image-text loss; diffusion models learn to predict the noise added at each step.

जवाब: Each pair's negatives are the other items in the batch, so more items give more and harder negatives

In-batch negatives are the core of the loss; bigger batches supply more of them per step.

जवाब: Asking for a clearer photo when image quality is low and allowing an 'unsure' answer

Detecting poor inputs and permitting uncertainty prevents the model from guessing a diagnosis from pixels it cannot see clearly.