feat: API regeneration: 20260422T070103Z#15582
Conversation
feat: Support VeoLoraTuningSpec in the tuning jobs PiperOrigin-RevId: 903331273 Source-Link: googleapis/googleapis@939ba3b
google-cloud-sdk-librarian-dotnet-robot
requested a review
from a team
as a code owner
|
Pull request diff results Diff level: Minor Comparing with previous NuGet package Diff level: Minor Finished comparisons for Google.Cloud.AIPlatform.V1Beta1 Finding changes in Google.DevicesAndServices.Health.V4... |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces the new Google.DevicesAndServices.Health.V4 client library and updates the Google.Cloud.AIPlatform.V1Beta1 library to support VeoLoraTuningSpec within TuningJob. Additionally, VeoHyperParameters has been expanded with new fields for data mixture ratio, tuning speed, and adapter size. Feedback on the generated code identifies several discrepancies in TuningJob.g.cs where the initialized default values for VeoDataMixtureRatio, TuningSpeed, and AdapterSize do not align with the behavior described in their respective documentation strings.
|
|
||
| /// <summary>Field number for the "veo_data_mixture_ratio" field.</summary> | ||
| public const int VeoDataMixtureRatioFieldNumber = 4; | ||
| private readonly static double VeoDataMixtureRatioDefaultValue = 0D; |
There was a problem hiding this comment.
The documentation for VeoDataMixtureRatio (line 7709) states that the default value is 0.1. However, VeoDataMixtureRatioDefaultValue is initialized to 0D. This creates a discrepancy between the documented behavior and the actual property value when the field is unset.
private readonly static double VeoDataMixtureRatioDefaultValue = 0.1D;|
|
||
| /// <summary>Field number for the "tuning_speed" field.</summary> | ||
| public const int TuningSpeedFieldNumber = 5; | ||
| private readonly static global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.TuningSpeed TuningSpeedDefaultValue = global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.TuningSpeed.Unspecified; |
There was a problem hiding this comment.
The documentation for TUNING_SPEED_UNSPECIFIED (line 8071) states that for Veo 3.0 models, the speed defaults to FAST. However, TuningSpeedDefaultValue is set to Unspecified. If the client is intended to reflect the effective default, this should be updated.
private readonly static global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.TuningSpeed TuningSpeedDefaultValue = global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.TuningSpeed.Fast;|
|
||
| /// <summary>Field number for the "adapter_size" field.</summary> | ||
| public const int AdapterSizeFieldNumber = 6; | ||
| private global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.AdapterSize adapterSize_ = global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.AdapterSize.Unspecified; |
There was a problem hiding this comment.
The documentation for ADAPTER_SIZE_EIGHT (line 8092) states it is the default adapter size for Veo LoRA tuning. However, the adapterSize_ field is initialized to Unspecified.
private global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.AdapterSize adapterSize_ = global::Google.Cloud.AIPlatform.V1Beta1.VeoHyperParameters.Types.AdapterSize.Eight;
2 participants
Changes in this PR