DL_Riddle_Quiz2

Adivinanza

Jugadas 3

Crear copia

Sobre esta actividad

DL_Riddle_Quiz2

Creada por

Siva Ranjini

India

Descarga la versión para jugar en papel

Top 10 resultados

Todavía no hay resultados para este juego. ¡Sé el primero en aparecer en el ranking! Inicia sesión para identificarte.

Crea tu propio juego gratis desde nuestro creador de juegos

Compite contra tus amigos para ver quien consigue la mejor puntuación en esta actividad

Crear reto

adivinador de adivinanzas

juan hoyos

Colombia

(22)

consiste en adivinar palabras de una frase incompleta o sin respuestas
Adivina el jugador NBA

guillermo .

España

(18)

¿lo conseguirás?
¡Adivina el brawler 1!

😎🎧 Mia Gamer ⭐💥

Perú

(58)

Adivínalo 😁⭐
Adivina la seleccion de futbol

Andres Castro

Colombia

(71)

Debes adivinar la selección
Criaturas fantásticas

Educaplay Educational Resources

España

(31)

Adivínalas usando el menor número de pistas posible


DL_Riddle_Quiz2Versión en línea DL_Riddle_Quiz2 por Siva Ranjini 1 I am the year-2015 architecture credited with breaking the "degradation problem" of very deep CNNs Hints 1 My block can be written y = F(x) + x. 2 My deepest published variant on ImageNet had 152 layers and I won ILSVRC 2015 classification. 2 U-Net's encoder doubles the channel count at each downsampling stage, starting at 64. Hints 1 After four downsampling stages I am the channel count at the bottleneck. 2 Compute it: 64 → 128 → 256 → 512 → ? 3 I am the operation that transfers feature maps from the contracting path to the expansive path. Hints 1 In the original U-Net paper I am implemented as "copy and crop" because of the unpadded convolutions. 2 I provide the decoder with high-resolution localisation information that pure upsampling cannot recover 4 I was introduced at MICCAI in 2015 by Ronneberger, Fischer and Brox. Hints 1 My architecture has a contracting path and an expansive path that resemble a single English letter. 2 I was trained on roughly 30 annotated images and won the ISBI cell-tracking challenge. 5 I am a class of generative model trained by setting up a two-player game between two neural networks. Hints 1 One of my networks creates fake samples; the other tries to tell real samples from fakes. 2 My acronym is three letters and I was introduced by Ian Goodfellow and colleagues in 2014.

DL_Riddle_Quiz2Versión en línea

DL_Riddle_Quiz2

por Siva Ranjini

I am the year-2015 architecture credited with breaking the "degradation problem" of very deep CNNs

Hints

My block can be written y = F(x) + x.

My deepest published variant on ImageNet had 152 layers and I won ILSVRC 2015 classification.

U-Net's encoder doubles the channel count at each downsampling stage, starting at 64.

Hints

After four downsampling stages I am the channel count at the bottleneck.

Compute it: 64 → 128 → 256 → 512 → ?

I am the operation that transfers feature maps from the contracting path to the expansive path.

Hints

In the original U-Net paper I am implemented as "copy and crop" because of the unpadded convolutions.

I provide the decoder with high-resolution localisation information that pure upsampling cannot recover

I was introduced at MICCAI in 2015 by Ronneberger, Fischer and Brox.

Hints

My architecture has a contracting path and an expansive path that resemble a single English letter.

I was trained on roughly 30 annotated images and won the ISBI cell-tracking challenge.

I am a class of generative model trained by setting up a two-player game between two neural networks.

Hints

One of my networks creates fake samples; the other tries to tell real samples from fakes.

My acronym is three letters and I was introduced by Ian Goodfellow and colleagues in 2014.

Puedes integrar el juego en un LMS compatible con LTI 1.1 o LTI 1.3 como Canvas, Moodle, o Blackboard. De esta manera podrás guardar las puntuaciones automáticamente en el libro de calificaciones de esa plataforma.

DL_Riddle_Quiz2

Adivinanza

Sobre esta actividad

Creada por

Descarga la versión para jugar en papel

Top juegos

Adivinanza

adivinador de adivinanzas

Adivinanza

Adivina el jugador NBA

Adivinanza

¡Adivina el brawler 1!

Adivinanza

Adivina la seleccion de futbol

Adivinanza

Criaturas fantásticas

DL_Riddle_Quiz2Versión en línea

por Siva Ranjini

I am the year-2015 architecture credited with breaking the "degradation problem" of very deep CNNs

Hints

U-Net's encoder doubles the channel count at each downsampling stage, starting at 64.

Hints

I am the operation that transfers feature maps from the contracting path to the expansive path.

Hints

I was introduced at MICCAI in 2015 by Ronneberger, Fischer and Brox.

Hints

I am a class of generative model trained by setting up a two-player game between two neural networks.

Hints

		%
	Anónimo	%
		%
		%