284 lines
6.2 KiB
YAML
284 lines
6.2 KiB
YAML
|
|
||
|
|
|
||
|
|
resource_packages:
|
||
|
|
- name: contest_on_demand_a100_x1
|
||
|
|
deploy_engines:
|
||
|
|
- ascend_infer
|
||
|
|
- vllm
|
||
|
|
default_deploy_engine: vllm
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于非常驻服务7b/14b-int4推理
|
||
|
|
- name: contest_on_demand_a100_x2
|
||
|
|
deploy_engines:
|
||
|
|
- ascend_infer
|
||
|
|
- vllm
|
||
|
|
default_deploy_engine: vllm
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 2
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于非常驻服务7b/14b-int4推理
|
||
|
|
|
||
|
|
- name: Ephemeral-RTX3090*1
|
||
|
|
deploy_engines:
|
||
|
|
- ascend_infer
|
||
|
|
- vllm
|
||
|
|
default_deploy_engine: vllm
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于非常驻服务7b/14b-int4推理
|
||
|
|
- name: R200-8F*1
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-kunlunxin
|
||
|
|
default_deploy_engine: slx-infer-kunlunxin
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: R200-8F
|
||
|
|
gpu_key: baidu.com/xpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 20
|
||
|
|
memory: 100Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: R200-8F*2
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-kunlunxin
|
||
|
|
default_deploy_engine: slx-infer-kunlunxin
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: R200-8F
|
||
|
|
gpu_key: baidu.com/xpu
|
||
|
|
gpu_value: 2
|
||
|
|
cpu: 20
|
||
|
|
memory: 100Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: R200-8F*4
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-kunlunxin
|
||
|
|
default_deploy_engine: slx-infer-kunlunxin
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: R200-8F
|
||
|
|
gpu_key: baidu.com/xpu
|
||
|
|
gpu_value: 4
|
||
|
|
cpu: 20
|
||
|
|
memory: 100Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: MLU370-X4*1
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-cambricon
|
||
|
|
default_deploy_engine: slx-infer-cambricon
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: MLU370-X4
|
||
|
|
gpu_key: cambricon.com/mlu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
|
||
|
|
- name: MLU370-X4*2
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-cambricon
|
||
|
|
default_deploy_engine: slx-infer-cambricon
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: MLU370-X4
|
||
|
|
gpu_key: cambricon.com/mlu
|
||
|
|
gpu_value: 2
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: Iluvatar MR-V100*1
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-iluvatar
|
||
|
|
default_deploy_engine: slx-infer-iluvatar
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: Iluvatar MR-V100
|
||
|
|
gpu_key: iluvatar.ai/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: Iluvatar MR-V100*2
|
||
|
|
deploy_engines:
|
||
|
|
- slx-infer-iluvatar
|
||
|
|
default_deploy_engine: slx-infer-iluvatar
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: Iluvatar MR-V100
|
||
|
|
gpu_key: iluvatar.ai/gpu
|
||
|
|
gpu_value: 2
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: Iluvatar MR-V100*4
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: Iluvatar MR-V100
|
||
|
|
gpu_key: iluvatar.ai/gpu
|
||
|
|
gpu_value: 4
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: RTX3090*1
|
||
|
|
deploy_engines:
|
||
|
|
- ascend_infer
|
||
|
|
- vllm
|
||
|
|
default_deploy_engine: vllm
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于7b/14b-int4推理
|
||
|
|
- name: RTX3090*2
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 2
|
||
|
|
cpu: 10
|
||
|
|
memory: 80Gi
|
||
|
|
description: 适用于 sagegpt / 34b-int4
|
||
|
|
- name: RTX3090*4
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 4
|
||
|
|
cpu: 10
|
||
|
|
memory: 40Gi
|
||
|
|
description: 适用于 72b-int4
|
||
|
|
- name: vmemory-24Gi
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: vgpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
memory_key: nvidia.com/gpumem
|
||
|
|
memory_value: 24576
|
||
|
|
cpu: 10
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于14b-int4/7b模型上线推理
|
||
|
|
|
||
|
|
- name: vmemory-2-48Gi
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: vgpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 2
|
||
|
|
memory_key: nvidia.com/gpumem
|
||
|
|
memory_value: 24576
|
||
|
|
cpu: 10
|
||
|
|
memory: 100Gi
|
||
|
|
description: 适用于式说大模型
|
||
|
|
|
||
|
|
- name: vmemory-10Gi
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: vgpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
memory_key: nvidia.com/gpumem
|
||
|
|
memory_value: 10000
|
||
|
|
cpu: 1
|
||
|
|
memory: 10Gi
|
||
|
|
description: 适用于7b-int4/embedding模型上线推理
|
||
|
|
- name: A100-14b模型推理套餐
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: Nvidia A100(80G)
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 4
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于14b模型上线推理
|
||
|
|
- name: A100-14b模型推理套餐1
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: Nvidia A100(80G)
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 1
|
||
|
|
cpu: 4
|
||
|
|
memory: 20Gi
|
||
|
|
description: 适用于14b模型上线推理
|
||
|
|
- name: A100-7b模型训练套餐(此种情况为非vgpu)
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: Nvidia A100(80G)
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 2
|
||
|
|
cpu: 40
|
||
|
|
memory: 128Gi
|
||
|
|
description: 适用于7b模型finetune
|
||
|
|
|
||
|
|
- name: 演示空间入门套餐
|
||
|
|
resources:
|
||
|
|
cpu: 1
|
||
|
|
memory: 2Gi
|
||
|
|
description: 适用于7b模型int4推理
|
||
|
|
- name: 演示空间初级套餐
|
||
|
|
resources:
|
||
|
|
gpu: {}
|
||
|
|
cpu: 2
|
||
|
|
memory: 4Gi
|
||
|
|
description: 适用于演示空间
|
||
|
|
- name: 演示空间中级套餐
|
||
|
|
resources:
|
||
|
|
gpu: {}
|
||
|
|
cpu: 4
|
||
|
|
memory: 16Gi
|
||
|
|
description: 适用于演示空间
|
||
|
|
- name: 演示空间高级套餐
|
||
|
|
resources:
|
||
|
|
gpu: {}
|
||
|
|
cpu: 8
|
||
|
|
memory: 32Gi
|
||
|
|
description: 适用于演示空间
|
||
|
|
- name: RTX3090*8
|
||
|
|
resources:
|
||
|
|
gpu:
|
||
|
|
type: gpu
|
||
|
|
device_label: NVIDIA GeForce RTX 3090
|
||
|
|
gpu_key: nvidia.com/gpu
|
||
|
|
gpu_value: 8
|
||
|
|
cpu: 10
|
||
|
|
memory: 200Gi
|
||
|
|
description: 适用于finetune
|