← All Models | GPT - OpenAI's Generative Pre-trained Transformers:

openai/whisper-tiny

📊 Model Parameters

Total Parameters 49,468,416
Context Length 2,048
Hidden Size 384
Layers 4
Attention Heads 6
KV Heads 6

💾 Memory Requirements

FP32 (Full) 188.7 MB
FP16 (Half) 94.4 MB
INT8 (Quantized) 47.2 MB
INT4 (Quantized) 23.6 MB

🔑 KV Cache (Inference)

Per Token (FP16) 6.14 KB
Max Context FP32 24.0 MB
Max Context FP16 12.0 MB
Max Context INT8 6.0 MB

⚙️ Model Configuration

Core Architecture

Vocabulary Size51,865
Number of Layers4

Attention Configuration

Attention Dropout0%
Tied EmbeddingsYes

Activation & Normalization

Activation Functiongelu

Special Tokens

BOS Token ID50,257
Pad Token ID50,257
EOS Token ID50257

Data Type

Model Dtypefloat32
Layer Types:
Attention
MLP/FFN
Normalization
Embedding