-
Notifications
You must be signed in to change notification settings - Fork 0
/
.project-metadata.yaml
49 lines (42 loc) · 1.22 KB
/
.project-metadata.yaml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
name: LLAMA-2-7B Standalone Model Deployment
description: This AMP deploys Llama-2-7b model as a CML model endpoint, callable via an API. Model is hosted within CML and requires GPU node with 16GB memory and 4 cores minimum.
author: Cloudera
date: "2023-11-15"
specification_version: 1.0
prototype_version: 1.0
environment_variables:
MODEL_NAME:
default: "LLAMA-2-7B CML"
description: Name to give to your model in CML. NOT IMPLEMENTED.
runtimes:
- editor: Workbench
kernel: Python 3.9
edition: Nvidia GPU
addons: []
tasks:
- type: create_model
name: Llama-2-7b
entity_label: llama-2-7b
description: LLAMA-2-7B model hosted in CML.
short_summary: LLAMA-2-7B
default_resources:
cpu: 4
memory: 16
gpu: 1
default_replication_policy:
type: fixed
num_replicas: 1
- type: build_model
name: Build LLAMA 2 model
entity_label: llama-2-7b
comment: First build by the AMP
examples:
- request:
prompt: What is Cloudera?
temperature: 0
max_new_tokens: 50
repetition_penalty: 0.5
target_file_path: Launch_model.py
target_function_name: api_wrapper
- type: deploy_model
entity_label: llama-2-7b