Skip to content

Commit 751e75e

Browse files
LJSpillerLJSpiller
authored
Merge pull request #215651 from ChrisHMSFT/chrhoder/updatestoopenai
[Cog-Svcs] Update Azure OpenAI docs with updated limits and other details
2 parents 6818d66 + dfc3776 commit 751e75e

File tree

7 files changed

+68
-9
lines changed

7 files changed

+68
-9
lines changed

‎articles/cognitive-services/openai/concepts/models.md

Lines changed: 52 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -5,7 +5,7 @@ description: Learn about the different models that are available in Azure OpenAI
55
ms.service: cognitive-services
66
ms.topic: conceptual
77
ms.date: 06/24/2022
8-
ms.custom: event-tier1-build-2022
8+
ms.custom: event-tier1-build-2022, references_regions
99
manager: nitinme
1010
author: ChrisHMSFT
1111
ms.author: chrhoder
@@ -15,7 +15,7 @@ keywords:
1515

1616
# Azure OpenAI models
1717

18-
The service provides access to many different models, grouped by family and capability. A model family typically associates models by their intended task. The following table describes model families currently available in Azure OpenAI.
18+
The service provides access to many different models, grouped by family and capability. A model family typically associates models by their intended task. The following table describes model families currently available in Azure OpenAI. Not all models are available in all regions currently. Please refer to the capability table at the bottom for a full breakdown.
1919

2020
| Model family | Description |
2121
|--|--|
@@ -152,6 +152,56 @@ Similar to text search embedding models, there are two input types supported by
152152

153153
When using our Embeddings models, keep in mind their limitations and risks.
154154

155+
## Model Summary table and region availability
156+
157+
### GPT-3 Models
158+
| Model | Supports Completions | Supports Embeddings | Base model Regions | Fine-Tuning Regions |
159+
| --- | --- | --- | --- | --- |
160+
| Ada | Yes | No | N/A | East US, South Central US, West Europe |
161+
| Text-Ada-001 | Yes | No | East US, South Central US, West Europe | N/A |
162+
| Babbage | Yes | No | N/A | East US, South Central US, West Europe |
163+
| Text-Babbage-001 | Yes | No | East US, South Central US, West Europe | N/A |
164+
| Curie | Yes | No | N/A | East US, South Central US, West Europe |
165+
| Text-curie-001 | Yes | No | East US, South Central US, West Europe | N/A |
166+
| Davinci* | Yes | No | N/A | East US, South Central US, West Europe |
167+
| Text-davinci-001 | Yes | No | South Central US, West Europe | N/A |
168+
| Text-davinci-002 | Yes | No | East US, South Central US, West Europe | N/A |
169+
| Text-davinci-fine-tune-002* | Yes | No | N/A | East US, West Europe |
170+
171+
\*Models available by request only. Please open a support request.
172+
173+
### Codex Models
174+
| Model | Supports Completions | Supports Embeddings | Base model Regions | Fine-Tuning Regions |
175+
| --- | --- | --- | --- | --- |
176+
| Code-Cushman-001* | Yes | No | South Central US, West Europe | East US, South Central US, West Europe |
177+
| Code-Davinci-002 | Yes | No | East US, West Europe | N/A |
178+
| Code-Davinci-Fine-tune-002* | Yes | No | N/A | East US, West Europe |
179+
180+
\*Models available for Fine-tuning by request only. Please open a support request.
181+
182+
183+
184+
### Embeddings Models
185+
| Model | Supports Completions | Supports Embeddings | Base model Regions | Fine-Tuning Regions |
186+
| --- | --- | --- | --- | --- |
187+
| text-similarity-ada-001 | No | Yes | East US, South Central US, West Europe | N/A |
188+
| text-similarity-babbage-001 | No | Yes | South Central US, West Europe | N/A |
189+
| text-similarit-curie-001 | No | Yes | East US, South Central US, West Europe | N/A |
190+
| text-similarity-davinci-001 | No | Yes | South Central US, West Europe | N/A |
191+
| text-search-ada-doc-001 | No | Yes | South Central US, West Europe | N/A |
192+
| text-search-ada-query-001 | No | Yes | South Central US, West Europe | N/A |
193+
| text-search-babbage-doc-001 | No | Yes | South Central US, West Europe | N/A |
194+
| text-search-babbage-query-001 | No | Yes | South Central US, West Europe | N/A |
195+
| text-search-curie-doc-001 | No | Yes | South Central US, West Europe | N/A |
196+
| text-search-curie-query-001 | No | Yes | South Central US, West Europe | N/A |
197+
| text-search-davinci-doc-001 | No | Yes | South Central US, West Europe | N/A |
198+
| text-search-davinci-query-001 | No | Yes | South Central US, West Europe | N/A |
199+
| code-search-ada-code-001 | No | Yes | South Central US, West Europe | N/A |
200+
| code-search-ada-text-001 | No | Yes | South Central US, West Europe | N/A |
201+
| code-search-babbage-code-001 | No | Yes | South Central US, West Europe | N/A |
202+
| code-search-babbage-text-001 | No | Yes | South Central US, West Europe | N/A |
203+
204+
155205
## Next steps
156206

157207
[Learn more about Azure OpenAI](../overview.md).

‎articles/cognitive-services/openai/includes/python.md

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -94,6 +94,8 @@ Run the code a few more times to see what other types of responses you get as th
9494

9595
Since our example of `Write a tagline for an ice cream shop.` provides very little context, it's normal for the model to not always return expected results. You can adjust the maximum number of tokens if the response seems unexpected or truncated.
9696

97+
The Azure OpenAI Service also performs content moderation on the prompt inputs and generated outputs. The prompts or responses may be filtered if harmful content is detected. For more information, see the [content filter](../concepts/content-filter.md) article.
98+
9799
## Clean up resources
98100

99101
If you want to clean up and remove an OpenAI resource, you can delete the resource or resource group. Deleting the resource group also deletes any other resources associated with it.

‎articles/cognitive-services/openai/includes/rest.md

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -100,6 +100,8 @@ The output from the completions API will look as follows.
100100
}
101101
```
102102

103+
The Azure OpenAI Service also performs content moderation on the prompt inputs and generated outputs. The prompts or responses may be filtered if harmful content is detected. For more information, see the [content filter](../concepts/content-filter.md) article.
104+
103105
## Clean up resources
104106

105107
If you want to clean up and remove an OpenAI resource, you can delete the resource or resource group. Deleting the resource group also deletes any other resources associated with it.

‎articles/cognitive-services/openai/includes/studio.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -67,7 +67,9 @@ The best way to start exploring completions is through our Playground. It's simp
6767
Tl;dr A neutron star is the collapsed core of a massive supergiant star, which had a total mass of between 10 and 25 solar masses, possibly more if the star was especially metal-rich.
6868
```
6969
70-
Keep in mind the accuracy of the response will vary depending on what model you've selected under **Deployments**. The response above was generated from a davinci based model, which is well-suited to this type of summarization whereas a Codex based model wouldn't perform as well at this particular task.
70+
Keep in mind the accuracy of the response will vary depending on what model you've selected under **Deployments**. The response above was generated from a Davinci based model, which is well-suited to this type of summarization whereas a Codex based model wouldn't perform as well at this particular task.
71+
72+
The Azure OpenAI Service also performs content moderation on the prompt inputs and generated outputs. The prompts or responses may be filtered if harmful content is detected. For more information, see the [content filter](../concepts/content-filter.md) article.
7173
7274
## Clean up resources
7375

‎articles/cognitive-services/openai/overview.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -23,12 +23,12 @@ The Azure OpenAI service provides REST API access to OpenAI's powerful language
2323
| Feature | Azure OpenAI |
2424
| --- | --- |
2525
| Models available | GPT-3 base series <br> Codex series <br> Embeddings series <br> Learn more in our [Models](./concepts/models.md) page.|
26-
| Fine-tuning | Ada <br> Babbage <br> Curie <br> Cushman* <br> Davinci* <br> \* available by request |
27-
| Billing Model| Coming Soon |
26+
| Fine-tuning | Ada <br> Babbage <br> Curie <br> Cushman* <br> Davinci* <br> \* available by request. Please open a support request|
27+
| Price | [Available here](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) |
2828
| Virtual network support | Yes |
2929
| Managed Identity| Yes, via Azure Active Directory |
3030
| UI experience | **Azure Portal** for account & resource management, <br> **Azure OpenAI Service Studio** for model exploration and fine tuning |
31-
| Regional availability | South Central US <br> West Europe |
31+
| Regional availability | East US <br> South Central US <br> West Europe |
3232
| Content filtering | Prompts and completions are evaluated against our content policy with automated systems. High severity content will be filtered. |
3333

3434
## Responsible AI

‎articles/cognitive-services/openai/quotas-limits.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -23,16 +23,16 @@ The following sections provide you with a quick guide to the quotas and limits t
2323
| Limit Name | Limit Value |
2424
|--|--|
2525
| OpenAI resources per region | 2 |
26-
| Requests per second per deployment | 1 |
26+
| Requests per second per deployment | 5 |
2727
| Max fine-tuned model deployments | 2 |
2828
| Ability to deploy same model to multiple deployments | Not allowed |
2929
| Total number of training jobs per resource | 100 |
3030
| Max simultaneous running training jobs per resource | 1 |
3131
| Max training jobs queued | 20 |
3232
| Max Files per resource | 50 |
33-
| Total size of all files per resource | 1 GB|
33+
| Total size of all files per resource | 1 GB |
3434
| Max training job time (job will fail if exceeded) | 120 hours |
35-
| Max training job size (tokens in training file * # of epochs) | **Ada**: 4-M tokens <br> **Babbage**: 4-M tokens <br> **Curie**: 4-M tokens <br> **Cushman**: 4-M tokens <br> **Davinci**: 500 K |
35+
| Max training job size (tokens in training file * # of epochs) | **Ada**: 40-M tokens <br> **Babbage**: 40-M tokens <br> **Curie**: 40-M tokens <br> **Cushman**: 40-M tokens <br> **Davinci**: 10-M |
3636

3737

3838
### General best practices to mitigate throttling during autoscaling

‎articles/cognitive-services/openai/toc.yml

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -8,6 +8,8 @@ items:
88
href: overview.md
99
- name: Quotas and limits
1010
href: quotas-limits.md
11+
- name: Pricing
12+
href: https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/
1113
- name: Quickstarts
1214
href: quickstart.md
1315
- name: Concepts
@@ -66,6 +68,7 @@ items:
6668
href: https://azure.microsoft.com/global-infrastructure/services/?products=cognitive-services
6769
- name: Terms of use
6870
href: https://azure.microsoft.com/support/legal/preview-supplemental-terms/
71+
6972

7073

7174

0 commit comments

Comments
 (0)