Spaces:
Running
Running
Upload AWS_and_Pinecone_Accounts_for_LiveRAG.md
Browse files
Operational_Instructions/AWS_and_Pinecone_Accounts_for_LiveRAG.md
ADDED
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# AWS and Pinecone Accounts for LiveRAG
|
2 |
+
|
3 |
+
## Using AWS Accounts
|
4 |
+
|
5 |
+
During the LiveRAG Challenge, participants may use two AWS accounts:
|
6 |
+
- **Personal AWS Account** β This is your private AWS account, where you can apply AWS credits provided for the competition.
|
7 |
+
- **TII-Provided AWS Account** β This account is managed by TII and grants access to the dense (Pinecone) and sparse (OpenSearch) pre-built indices.
|
8 |
+
|
9 |
+
Using these accounts is optional but highly recommended to optimize both cost and effort.
|
10 |
+
|
11 |
+
## Using Your Personal AWS Account with AWS Credits
|
12 |
+
|
13 |
+
Each participating group receives AWS credits for the duration of the competition.
|
14 |
+
The credits are limited, and any usage beyond the allocated amount will be charged to your personal payment method.
|
15 |
+
To ensure your credits are properly applied, follow the official AWS instructions [here](https://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/useconsolidatedbilling-credits.html).
|
16 |
+
|
17 |
+
### Personal AWS Account Setup and Credit Application Steps
|
18 |
+
|
19 |
+
1. **Create an AWS Account** β Sign up for an AWS account and add a payment method (typically a credit card). If you have an existing AWS account you may reuse it.
|
20 |
+
2. **Apply AWS Credits** β Follow the instructions in the provided AWS documentation to redeem and apply your credits.
|
21 |
+
3. **Monitor Costs Proactively** β AWS usage, especially GPU-based resources, can be expensive. We highly recommend to closely monitor your costs.
|
22 |
+
- Use AWS Cost Management tools to monitor spending.
|
23 |
+
- Set up billing alarms using AWS CloudWatch to receive notifications when costs exceed predefined thresholds.
|
24 |
+
|
25 |
+
### Cost Optimization Recommendations
|
26 |
+
|
27 |
+
TII and AWS have estimated the cost of a typical teamβs infrastructure, including GPU usage, and AWS has provided credits accordingly.
|
28 |
+
However, individual teams may have varying requirements, so cost management is essential:
|
29 |
+
- **Shut down unused resources** β Always shut down GPUs and other compute instances when not in use.
|
30 |
+
- **Experiment on smaller datasets first** β This approach speeds up iteration cycles and reduces expenses before scaling up to larger datasets.
|
31 |
+
|
32 |
+
## Using the TII-Provided AWS Account
|
33 |
+
|
34 |
+
The TII-provided AWS account is strictly limited to specific use cases, primarily for accessing pre-built indices. These include:
|
35 |
+
|
36 |
+
- A **Sparse OpenSearch Index** hosted by AWS.
|
37 |
+
- A **Dense Pinecone Index** hosted by Pinecone, with access credentials stored on AWS.
|
38 |
+
|
39 |
+
|
40 |
+
### Access and Setup Instructions
|
41 |
+
|
42 |
+
To gain access and configure the TII-Provided AWS account, follow these steps:
|
43 |
+
|
44 |
+
1. **Receive Sign-In Details** β You will receive an email with login credentials from the organizers.
|
45 |
+
2. **Sign In & Change Password** β Log in to the AWS console and update your password.
|
46 |
+
3. **Create AWS Command Line Interface (CLI) Credentials** β Generate an access key and secret for CLI usage.
|
47 |
+
4. **Configure AWS CLI Profile** β Run the following command to set up the CLI for this account and follow the instructions:
|
48 |
+
|
49 |
+
```sh
|
50 |
+
aws configure --profile sigir-participant
|
51 |
+
```
|
52 |
+
|
53 |
+
5. **Verify Access** β Query the indices to confirm successful configuration.
|
54 |
+
|
55 |
+
For detailed instructions on account setup, credentials, and index usage, refer to this [example notebook](Operational_Instructions/Indices_Usage_Examples_for_LiveRAG.ipynb)
|
56 |
+
|
57 |
+
- **Use AWS CLI Profiles** β We recommend usage of separate AWS profiles. Configure separate AWS CLI profiles for better account management:
|
58 |
+
|
59 |
+
```sh
|
60 |
+
aws configure --profile personal
|
61 |
+
aws configure --profile sigir-participant
|
62 |
+
```
|
63 |
+
- `personal`: Your own AWS account.
|
64 |
+
- `sigir-participant`: The TII-provided AWS account (described below).
|
65 |
+
|
66 |
+
For beginners, it is recommended to complete introductory AWS tutorials available [here](https://aws.amazon.com/getting-started/).
|
67 |
+
|
68 |
+
|
69 |
+
## Using the Pre-Built Indices and Building Your Own Indices
|
70 |
+
We provide you with two basic pre-built dense (Pinecone) and sparse (Opensearch) indices. See usage [here](Indices_Usage_Examples_for_LiveRAG.ipynb). You may use them freely for LiveRAG Challenge tasks only.
|
71 |
+
|
72 |
+
You're welcome to build your own Pinecone/OpenSearch, or other indices. We encourage you to take advantage of the AWS and Pinecone credits we provide for this purpose. OpenSearch can be provisioned and paid by AWS credits.
|
73 |
+
|
74 |
+
For instructions pertaining to building your own Pinecone index see [here](Pinecone_for_LiveRAG.md).
|
75 |
+
|
76 |
+
## Summary of Key Recommendations
|
77 |
+
|
78 |
+
| **Action** | **Recommendation** |
|
79 |
+
|:-----------------------|:-------------------------------------------------------------|
|
80 |
+
| **AWS Credits** | Apply credits early and monitor usage. |
|
81 |
+
| **Billing Alerts** | Use AWS CloudWatch to avoid unexpected charges. |
|
82 |
+
| **Resource Management** | Shut down unused resources to reduce costs. |
|
83 |
+
| **Experimentation** | Start with small datasets before scaling up. |
|
84 |
+
| **Account Management** | Use separate AWS CLI profiles for personal and TII-Provided accounts. |
|
85 |
+
|
86 |
+
By following these best practices, you can effectively manage your resources, optimize costs, and maximize your efficiency during the LiveRAG Challenge.
|