Receiving quota error when trying to use bison chat model in Vertex AI

[dashy]

Hi, I want to try out the new bison chat model. However, when I'm asking anything I'm receiving this error: 

Quota exceeded for aiplatform.googleapis.com/online_prediction_requests_per_base_model with base model: chat-bison. Please submit a quota increase request.

[dashy]

Thanks! Appreciate your help.

[billylo196]

I am still getting quota errors... which type of quota do I need?  thx, Billy.

com.google.cloud.ai.platform.common.errors.AiPlatformException: code=RESOURCE_EXHAUSTED, message=The following quota metrics exceed quota limits: aiplatform.googleapis.com/restricted_image_training_nvidia_a100_80gb_gpus, cause=null; Failed to create custom job.Project number: 801580958718, Job id: 1275918922603823104, Task id: 5042561673047572480, Task name: large-language-model-tuner, Task state: DRIVER_SUCCEEDED, Execution name: projects/801580958718/locations/us-central1/metadataStores/default/executions/1041959984043837747; Failed to create external task or refresh its state. Task:Project number: 801580958718, Job id: 1275918922603823104, Task id: 5042561673047572480, Task name: large-language-model-tuner, Task state: DRIVER_SUCCEEDED, Execution name: projects/801580958718/locations/us-central1/metadataStores/default/executions/1041959984043837747; Failed to handle the pipeline task. Task: Project number: 801580958718, Job id: 1275918922603823104, Task id: 5042561673047572480, Task name: large-language-model-tuner, Task state: DRIVER_SUCCEEDED, Execution name: projects/801580958718/locations/us-central1/metadataStores/default/executions/1041959984043837747

{"display_name": "large-language-model-tuner-1275918922603823104-8789556563019825152", "job_spec": {"worker_pool_specs": [{"replica_count": 1, "machine_spec": {"machine_type": "a2-ultragpu-8g", "accelerator_type": "NVIDIA_A100_80GB", "accelerator_count": 8}, "disk_spec": {"boot_disk_type": "pd-ssd", "boot_disk_size_gb": 500}, "container_spec": {"image_uri": "us-docker.pkg.dev/vertex-ai-restricted/llm/llm-pets-gpu:v2.2.0", "args": ["tune_v2", "--model=text-bison@001", "--task=generation", "--train_data_path=gs://robo_me_billy/801580958718/tune-large-model-20230802084453/convert-jsonl-to-tfrecord_-433815473834950656/output_dir/messages.tfrecord", "--feature_keys=input_text", "--label_key=output_text", "--train_steps=100", "--inputs_length=8192", "--outputs_length=1024", "--ici_mesh_shape=[1, 1, 8]", "--per_core_batch_size=1", "--adapter_length=1", "--train_evaluation_interval=20", "--learning_rate=0.0002",

[Domenic]

Hello, sadly my quota seems to be set to 1.
I was able to request a higher limit, so we'll see what happens there, but shouldn't I have at least 30 to begin with?

Thank you!

[oiwejdsd]

Thanks @mchrestkha1, submitting a prompt in Generative AI Studio does work now.

However, tuning in Generative AI Studio still throws a quota error.

com.google.cloud.ai.platform.common.errors.AiPlatformException: code=RESOURCE_EXHAUSTED, message=The following quota metrics exceed quota limits: aiplatform.googleapis.com/custom_model_training_cpus, cause=null; Failed to create custom job for the task. Task: Project number: xxx, Job id: xxx, Task id: xxx, Task name: validate-pipeline, Task state: DRIVER_SUCCEEDED, Execution name: projects/xxx/locations/europe-west4/metadataStores/default/executions/xxx; Failed to create external task or refresh its state. Task:Project number: xxx, Job id: xxx, Task id: xxx, Task name: validate-pipeline, Task state: DRIVER_SUCCEEDED, Execution name: projects/xxx/locations/europe-west4/metadataStores/default/executions/xxx; Failed to handle the pipeline task. Task: Project number: xxx, Job id: xxx, Task id: xxx, Task name: validate-pipeline, Task state: DRIVER_SUCCEEDED, Execution name: projects/xxx/locations/europe-west4/metadataStores/default/executions/xxx

I'm still on the Free Trial and followed the docs here https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-models#generative-ai-tune-model-dr...

[Hoora]

Customer started using code-bison and gets the same error. Any update? How they can request for Quotas increase? 

[tahar]

i still get the same error how i can requist more quotas ? i need to for project require more then 1000 req per minut 

[bhavanaprasad]

Receiving the same error when trying to create a tuned text bison model. Any inputs are appreciated

[HP20]

Receiving same error in my free account and can't request new quota as well.

[ibnetdevs]

I am also stuck with this error, it shows in one of my projects, but on other project it shows 360 limit. Both projects have billing enabled. The project that shows 0 has monthly billing of average $30. I am not sure whats the issue is. It started appearing a month ago when we were testing with api key different models.