augustin64/tipe
1
1
Fork 0
mirror of https://github.com/augustin64/projet-tipe synced 2025-02-02 19:39:39 +01:00
Code Issues Packages Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
augustin64 2023-04-15 21:19:20 +02:00
parent 6e1854502d
commit c4cf2c90a0
1 changed files with 12 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
README.md
View File

@ -58,15 +58,15 @@ src/dense
Résultats sur MNIST: Résultats sur MNIST:
``` ```
$ build/dense-main train -e 5 -i data/mnist/train-images-idx3-ubyte -l data/mnist/train-labels-idx1-ubyte $ build/dense-main train -e 5 -i data/mnist/train-images-idx3-ubyte -l data/mnist/train-labels-idx1-ubyte
Threads [8] Époque [0/5] Image [60000/60000] Accuracy: 86.7% Threads [8] Époque [0/5] Image [60000/60000] Accuracy: 86.7%
Threads [8] Époque [1/5] Image [60000/60000] Accuracy: 90.0% Threads [8] Époque [1/5] Image [60000/60000] Accuracy: 90.0%
Threads [8] Époque [2/5] Image [60000/60000] Accuracy: 90.5% Threads [8] Époque [2/5] Image [60000/60000] Accuracy: 90.5%
Threads [8] Époque [3/5] Image [60000/60000] Accuracy: 90.8% Threads [8] Époque [3/5] Image [60000/60000] Accuracy: 90.8%
Threads [8] Époque [4/5] Image [60000/60000] Accuracy: 90.9% Threads [8] Époque [4/5] Image [60000/60000] Accuracy: 90.9%
Execution time: 50 s. Execution time: 50 s.
$ build/dense-main test -i data/mnist/t10k-images-idx3-ubyte -l data/mnist/t10k-labels-idx1-ubyte -m out.bin $ build/dense-main test -i data/mnist/t10k-images-idx3-ubyte -l data/mnist/t10k-labels-idx1-ubyte -m out.bin
10000 Images Accuracy: 90.4% 10000 Images Accuracy: 90.4%
``` ```
<br/><br/> <br/><br/>
@ -130,15 +130,15 @@ Pas de fichier de sortie spécifié, défaut: out.bin
Taux d'apprentissage initial: 3.00e-04 Taux d'apprentissage initial: 3.00e-04
Initialisation: 185ms Initialisation: 185ms
Threads [8] Époque [0/5] Image [60000/60000] Accuracy: 60.5244% Loss: 0.046990: Temps: 1mn 31s Threads [8] Époque [0/5] Image [60000/60000] Accuracy: 60.5244% Loss: 0.046990: Temps: 1mn 31s
Threads [8] Époque [1/5] Image [60000/60000] Accuracy: 91.3681% Loss: 0.012858: Temps: 1mn 31s Threads [8] Époque [1/5] Image [60000/60000] Accuracy: 91.3681% Loss: 0.012858: Temps: 1mn 31s
Threads [8] Époque [2/5] Image [60000/60000] Accuracy: 93.0354% Loss: 0.010549: Temps: 1mn 30s Threads [8] Époque [2/5] Image [60000/60000] Accuracy: 93.0354% Loss: 0.010549: Temps: 1mn 30s
Threads [8] Époque [3/5] Image [60000/60000] Accuracy: 93.8429% Loss: 0.009243: Temps: 1mn 29s Threads [8] Époque [3/5] Image [60000/60000] Accuracy: 93.8429% Loss: 0.009243: Temps: 1mn 29s
Threads [8] Époque [4/5] Image [60000/60000] Accuracy: 94.4507% Loss: 0.008328: Temps: 1mn 31s Threads [8] Époque [4/5] Image [60000/60000] Accuracy: 94.4507% Loss: 0.008328: Temps: 1mn 31s
Temps total: 7mn 33s Temps total: 7mn 33s
$ build/cnn-main test -m out.bin -d mnist -i data/mnist/t10k-images-idx3-ubyte -l data/mnist/t10k-labels-idx1-ubyte $ build/cnn-main test -m out.bin -d mnist -i data/mnist/t10k-images-idx3-ubyte -l data/mnist/t10k-labels-idx1-ubyte
Accuracy: 94.860001 Loss: 0.007952 Accuracy: 94.860001 Loss: 0.007952
``` ```
<br/><br/> <br/><br/>