Guide-yakabatana mibvunzo · Pakati Nhanho

Yakasanganiswa Precision Yekudzidzisa Quiz

Covers FP16 range and loss scaling, bfloat16, master weights, autocast, Tensor Core accumulation, and measured hardware-dependent gains.

Nzira dzekutungamira dzinoenderanaKudzidziswa kweMixed Precision
Mubvunzo 1 ye 8

In an FP16 training recipe that keeps an FP32 master weight, why retain it?

Tarisa mubvunzo wese8 mibvunzo

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

  1. In an FP16 training recipe that keeps an FP32 master weight, why retain it?
    • Kuti uchengetedze huremu muhuremu hudiki hunogadziridza iyo 16-bit inogona kutenderedza kure
    • Kuita kuti modhi faira rive diki pane dhisiki
    • Kubvumidza kudzidziswa pasina GPU
    • Kuti encrypt iyo modhi paramita
  2. Idambudziko ripi rinogadziriswa kurasikirwa kuyera muFP16 kudzidziswa?
    • Models inodarika iyo data
    • Gradients inoyerera kusvika zero nekuti FP16's renji idiki
    • MaGPU ari kupera dhisiki nzvimbo
    • Inononoka kurodha data kubva kudhisiki
  3. Ndeupi mukana une bfloat16 pamusoro peFP16?
    • Inoshandisa zvidimbu zvishoma
    • Inochengeta zvinyorwa panzvimbo yenhamba
    • Iyo ine yakafanana yakakura exponent renji seFP32, ichidzikisa kudiwa kwekurasikirwa kwekuyera
    • Inomhanya chete paCPUs
  4. For many supported Tensor Core matrix operations, how are partial sums commonly accumulated?
    • In integer arithmetic with no rounding
    • In text before converting to FP16
    • In FP16 only, which prevents every numerical error
    • In FP32, which reduces rounding error compared with low-precision accumulation
  5. What is a common benefit of using 16-bit values for supported training operations?
    • They can reduce tensor memory and improve throughput, with gains depending on hardware and workload
    • They guarantee higher model accuracy
    • They remove the need to compute gradients
    • They always double end-to-end training speed
  6. Which current PyTorch APIs are commonly used for automatic mixed-precision training?
    • torch.cuda.amp only, which is the recommended new namespace
    • torch.amp.autocast and, for FP16 gradient scaling, torch.amp.GradScaler
    • torch.load and torch.save
    • torch.nn.DataParallel and torch.manual_seed
  7. Why does bfloat16 often need less loss scaling than FP16?
    • It has more mantissa bits than FP32
    • It stores gradients as integers
    • It has an exponent range like FP32, so very small gradients are less likely to underflow
    • It never rounds values
  8. What does autocast do in automatic mixed precision?
    • Converts every tensor and operation to FP16 unconditionally
    • Replaces the optimizer with a different algorithm
    • Guarantees identical bitwise results to FP32
    • Selects operation dtypes so supported operations can use lower precision while others retain safer precision