Signed-off-by: divinerapier <sihao.fang@outlook.com>
This commit is contained in:
divinerapier 2025-03-14 17:00:03 +08:00
parent b4555aabcf
commit f929a44f2d
No known key found for this signature in database
GPG Key ID: 18615A772E240111
18 changed files with 18 additions and 18 deletions

View File

@ -24,7 +24,7 @@ Apache Flink is an open-source streaming framework developed by the Apache Softw
### The premise condition ### The premise condition
Make sure the deployed Kubernate, Kubectl, Volcano are installed correctly. Make sure the deployed Kubernetes, Kubectl, Volcano are installed correctly.
### Deployment process ### Deployment process

View File

@ -28,7 +28,7 @@ PS-worker model: Parameter Server performs model-related services, Work Server t
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates has many problems: TensorFlow on Kubernetes has many problems:
- Resource isolation. - Resource isolation.
- Lack of GPU scheduling, Gang Schuler. - Lack of GPU scheduling, Gang Schuler.

View File

@ -24,7 +24,7 @@ Apache Flink is an open-source streaming framework developed by the Apache Softw
### The premise condition ### The premise condition
Make sure the deployed Kubernate, Kubectl, Volcano are installed correctly. Make sure the deployed Kubernetes, Kubectl, Volcano are installed correctly.
### Deployment process ### Deployment process

View File

@ -28,7 +28,7 @@ PS-worker model: Parameter Server performs model-related services, Work Server t
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates has many problems: TensorFlow on Kubernetes has many problems:
- Resource isolation. - Resource isolation.
- Lack of GPU scheduling, Gang Schuler. - Lack of GPU scheduling, Gang Schuler.

View File

@ -24,7 +24,7 @@ Apache Flink is an open-source streaming framework developed by the Apache Softw
### The premise condition ### The premise condition
Make sure the deployed Kubernate, Kubectl, Volcano are installed correctly. Make sure the deployed Kubernetes, Kubectl, Volcano are installed correctly.
### Deployment process ### Deployment process

View File

@ -28,7 +28,7 @@ PS-worker model: Parameter Server performs model-related services, Work Server t
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates has many problems: TensorFlow on Kubernetes has many problems:
- Resource isolation. - Resource isolation.
- Lack of GPU scheduling, Gang Schuler. - Lack of GPU scheduling, Gang Schuler.

View File

@ -24,7 +24,7 @@ Apache Flink is an open-source streaming framework developed by the Apache Softw
### The premise condition ### The premise condition
Make sure the deployed Kubernate, Kubectl, Volcano are installed correctly. Make sure the deployed Kubernetes, Kubectl, Volcano are installed correctly.
### Deployment process ### Deployment process

View File

@ -28,7 +28,7 @@ PS-worker model: Parameter Server performs model-related services, Work Server t
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates has many problems: TensorFlow on Kubernetes has many problems:
- Resource isolation. - Resource isolation.
- Lack of GPU scheduling, Gang Schuler. - Lack of GPU scheduling, Gang Schuler.

View File

@ -24,7 +24,7 @@ Apache Flink is an open-source streaming framework developed by the Apache Softw
### The premise condition ### The premise condition
Make sure the deployed Kubernate, Kubectl, Volcano are installed correctly. Make sure the deployed Kubernetes, Kubectl, Volcano are installed correctly.
### Deployment process ### Deployment process

View File

@ -28,7 +28,7 @@ PS-worker model: Parameter Server performs model-related services, Work Server t
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates has many problems: TensorFlow on Kubernetes has many problems:
- Resource isolation. - Resource isolation.
- Lack of GPU scheduling, Gang Schuler. - Lack of GPU scheduling, Gang Schuler.

View File

@ -24,7 +24,7 @@ Apache Flink is an open-source streaming framework developed by the Apache Softw
### The premise condition ### The premise condition
Make sure the deployed Kubernate, Kubectl, Volcano are installed correctly. Make sure the deployed Kubernetes, Kubectl, Volcano are installed correctly.
### Deployment process ### Deployment process

View File

@ -28,7 +28,7 @@ PS-worker model: Parameter Server performs model-related services, Work Server t
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates has many problems: TensorFlow on Kubernetes has many problems:
- Resource isolation. - Resource isolation.
- Lack of GPU scheduling, Gang Schuler. - Lack of GPU scheduling, Gang Schuler.

View File

@ -28,7 +28,7 @@ PS-worker模型Parameter Server执行模型相关业务Work Server训练
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates存在诸多的问题 TensorFlow on Kubernetes存在诸多的问题
- 资源隔离 - 资源隔离
- 缺乏GPU调度、Gang schduler。 - 缺乏GPU调度、Gang schduler。

View File

@ -28,7 +28,7 @@ PS-worker模型Parameter Server执行模型相关业务Work Server训练
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates存在诸多的问题 TensorFlow on Kubernetes存在诸多的问题
- 资源隔离 - 资源隔离
- 缺乏GPU调度、Gang schduler。 - 缺乏GPU调度、Gang schduler。

View File

@ -28,7 +28,7 @@ PS-worker模型Parameter Server执行模型相关业务Work Server训练
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates存在诸多的问题 TensorFlow on Kubernetes存在诸多的问题
- 资源隔离 - 资源隔离
- 缺乏GPU调度、Gang schduler。 - 缺乏GPU调度、Gang schduler。

View File

@ -28,7 +28,7 @@ PS-worker模型Parameter Server执行模型相关业务Work Server训练
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates存在诸多的问题 TensorFlow on Kubernetes存在诸多的问题
- 资源隔离 - 资源隔离
- 缺乏GPU调度、Gang schduler。 - 缺乏GPU调度、Gang schduler。

View File

@ -28,7 +28,7 @@ PS-worker模型Parameter Server执行模型相关业务Work Server训练
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates存在诸多的问题 TensorFlow on Kubernetes存在诸多的问题
- 资源隔离 - 资源隔离
- 缺乏GPU调度、Gang schduler。 - 缺乏GPU调度、Gang schduler。

View File

@ -28,7 +28,7 @@ PS-worker模型Parameter Server执行模型相关业务Work Server训练
{{<figure library="1" src="ps-worker.png" title="ps-worker">}} {{<figure library="1" src="ps-worker.png" title="ps-worker">}}
TensorFlow on Kubernates存在诸多的问题 TensorFlow on Kubernetes存在诸多的问题
- 资源隔离 - 资源隔离
- 缺乏GPU调度、Gang schduler。 - 缺乏GPU调度、Gang schduler。