NVIDIA NIM Connector Guide

Back Next

NVIDIA NIM connection actions and objects

When you publish an NVIDIA NIM connection in Application Integration, the

Actions

and

Objects

appear on the

Metadata

tab.

You must escape special characters in payloads as shown in the following table:

Symbol	Character	Replacement
<	left angle bracket or less than	<
>	right anlge bracket or greater than	>
{	left curly bracket or opening brace	{{
}	right curly bracket or closing brace	}}
&	ampersand	&

To use fields with data types, such as array, boolean, double, or integer, you must use annotation as shown in the following sample code:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<field_array m:isArray="true"> 
<property_1> ...</property_1> 
</field_array> 
<field_boolean m:type="xs:boolean">true</field_boolean> 
<field_double m:type="xs:double">10.3</field_double> 
<field_integer m:type="xs:int">5</field_integer> 
</root>

Creating assets and referencing them in payloads is not supported in the current version of the connector.

Some models might return responses larger than 5 MB. It is recommended to publish the connection on an agent to handle large files efficiently.

Consider the following information when you use specific actions and objects:

The Invoke Healthcare Model
action invokes the Healthcare API models, such as deepmind/alphafold2 and nvidia/genmol. It invokes NVIDIA healthcare models to generate predictions or analysis based on the input data.

The connector is tested with the following models:

Deepmind/alphafold2

Nvidia/esmfold

Proteinmpnn/predict

Nvidia/genmol

Nvidia/molmim

You must provide values to the input fields, such as payload, version, and NVCF_POLL_SECONDS.

For example, consider the nvidia/genmol model with the following endpoint URL:

https://health.api.nvidia.com/v1/biology/nvidia/genmol/generate

In this URL,

is the API version and

biology

nvidia

genmol

, and

generate

are the individual URL path segments.

NVCF_POLL_SECONDS - interval in seconds between status checks for a pending request to an NVIDIA Cloud Function (NVCF). The maximum supported value is 59 seconds.

The following snippet is a payload example:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<smiles>C124CN3C1.S3(=O)(=O)CC.C4C#N.[*{20-20}]</smiles> 
<num_molecules m:type="xs:int">30</num_molecules> 
<temperature>0.01</temperature> 
<noise>1</noise> 
<step_size m:type="xs:int">1</step_size> 
<scoring>QED</scoring> 
<unique m:type="xs:boolean">false</unique> 
</root>

The following snippet is a sample response:

200 OK 
<root> 
  <molecules> 
   <score>0.851</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(c2ccc(C(F)(F)F)cc2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.851</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(c2ccc(C(F)(F)F)cc2)C1</smiles> 
  </molecules>  
  ...
  ... 
  <molecules> 
   <score>0.851</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(c2ccc(C(F)(F)F)cc2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.851</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(c2ccc(C(F)(F)F)cc2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.805</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)C2(C)CCC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.802</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(OCc2ccc(Cl)cc2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.798</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)CCC(C)(C)C)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.785</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(OCC2CCCC(O)C2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.78</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)CCC(F)(F)F)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.78</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)CCC(F)(F)F)C1</smiles> 
  </molecules> 
    ...
    ... 
  <molecules> 
   <score>0.78</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)CCC(F)(F)F)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.773</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)C2CCCCCC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.752</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(NC(=O)CCC(C)(F)F)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.75</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(C(=O)CCC2(O)CCC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.668</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(C(=O)CCC2CCCCCC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.655</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(OC(=O)CCC2CC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.627</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(C(=O)CCCN2CCOCC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.562</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(OC(=O)C2CC=CCC2)C1</smiles> 
  </molecules> 
  <molecules> 
   <score>0.481</score> 
   <smiles>CCS(=O)(=O)N1CC(CC#N)(OC(=O)CCCCC(=O)O)C1</smiles> 
  </molecules> 
  <status>success</status> 
</root>

The Get Polling Status
action gets the result of an earlier function invocation request that returned a 202 status. It retrieves the result of a previous function invocation that returned a 202 status code.

For example, when invoking a model, if the response takes longer than the configured response timeout, the server returns a status header as in-progress and a status code 202. It also provides a request ID in the response header, which can be used to poll for the final result.

You must provide values to the input fields, such as host, version, and NVCF_POLL_SECONDS.

By default, the status polling is performed using the following URL:

https://optimize.api.nvidia.com/v1/status/{requestId}

In this URL,

optimize.api.nvidia.com

is the host,

is the API version, and

status

is a URL path segment.

Some models might use different status polling URLs. For more information, see the NVIDIA documentation for model-specific details.

The following snippets are some additional payload examples for other models:

Deepmind/alphafold2

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<sequence>MVPSAGQLALFALGIVLAACQALENSTSPLSADPPVAAAVVSHFNDCPDSHTQFCFHGTCRFLVQEDKPACVCHSGYVGARCEHADLLAVVAASQKKQAITALVVVSIVALAVLIITCVLIHCCQVRKHCEWCRALICRHEKPSALLKGRTACCHSETVV</sequence> 
<algorithm>mmseqs2</algorithm> 
<databases m:isArray="true">uniref90</databases> 
<databases m:isArray="true">small_bfd</databases> 
<e_value m:type="xs:double">0.0001</e_value> 
<iterations m:type="xs:int">1</iterations> 
<relax_prediction m:type="xs:boolean">false</relax_prediction> 
<skip_template_search m:type="xs:boolean">true</skip_template_search> 
</root>

Nvidia/esmfold

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<sequence>MVPSAGQLALFALGIVLAACQALENSTSPLSADPPVAAAVVSHFNDCPDSHTQFCFHGTCRFLVQEDKPACVCHSGYVGARCEHADLLAVVAASQKKQAITALVVVSIVALAVLIITCVLIHCCQVRKHCEWCRALICRHEKPSALLKGRTACCHSETVV</sequence> 
<algorithm>mmseqs2</algorithm> 
<databases m:isArray="true">uniref90</databases> 
<databases m:isArray="true">small_bfd</databases> 
<e_value m:type="xs:double">0.0001</e_value> 
<iterations m:type="xs:int">1</iterations> 
<relax_prediction m:type="xs:boolean">false</relax_prediction> 
<skip_template_search m:type="xs:boolean">true</skip_template_search> 
</root>

Proteinmpnn/predict

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<input_pdb> ATOM      1  N   GLY A   1      11.104  13.207  10.678  1.00 20.00           N   
ATOM      2  CA  GLY A   1      12.560  13.207  10.678  1.00 20.00           C   
ATOM      3  C   GLY A   1      13.104  14.207  11.678  1.00 20.00           C   
ATOM      4  O   GLY A   1      12.560  15.207  11.678  1.00 20.00           O   
TER 
END </input_pdb> 
<input_pdb_chains m:isArray="true">A</input_pdb_chains> 
<ca_only m:type="xs:boolean">false</ca_only> 
<use_soluble_model m:type="xs:boolean">true</use_soluble_model> 
<random_seed m:type="xs:int">42</random_seed> 
<num_seq_per_target m:type="xs:int">1</num_seq_per_target> 
<sampling_temp m:isArray="true" m:type="xs:double">0.1</sampling_temp> 
</root>

Nvidia/molmim

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<algorithm>CMA-ES</algorithm> 
<smi>[H][C@@]12Cc3c[nH]c4cccc(C1=C[C@H](NC(=O)N(CC)CC)CN2C)c34</smi> 
<num_molecules m:type="xs:int">30</num_molecules> 
<iterations m:type="xs:int">10</iterations> 
<property_name>QED</property_name> 
<particles m:type="xs:int">30</particles> 
<minimize m:type="xs:boolean">false</minimize> 
<min_similarity m:type="xs:double">0.3</min_similarity> 
<scaled_radius m:type="xs:double">1</scaled_radius> 
</root>

nvidia/maisi

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
 <num_output_samples m:type="xs:int">1</num_output_samples> 
 <body_region m:isArray="true">abdomen</body_region> 
 <anatomy_list m:isArray="true">liver</anatomy_list> 
 <output_size m:isArray="true" m:type="xs:int">256</output_size> 
 <output_size m:isArray="true" m:type="xs:int">256</output_size> 
 <output_size m:isArray="true" m:type="xs:int">256</output_size> 
 <controllable_anatomy_size m:isArray="true"> 
  <anonymous_array m:isArray="true">hepatic tumor</anonymous_array> 
  <anonymous_array m:isArray="true" m:type="xs:double">0.3</anonymous_array> 
 </controllable_anatomy_size> 
<controllable_anatomy_size m:isArray="true"> 
  <anonymous_array m:isArray="true">liver</anonymous_array> 
  <anonymous_array m:isArray="true" m:type="xs:double">0.5</anonymous_array> 
 </controllable_anatomy_size> 
 <image_output_ext>.nii.gz</image_output_ext> 
 <label_output_ext>.nii.gz</label_output_ext> 
 <pre_signed_url></pre_signed_url> 
 <spacing m:type="xs:double">1</spacing> 
 <spacing m:type="xs:double">1</spacing> 
 <spacing m:type="xs:double">1</spacing> 
</root>

adept/fuyu-8b

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
  <content>Hi! What is in this image?  
    &lt;img src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII==" /&gt; 
  </content> 
  <role>user</role> 
</messages> 
<temperature m:type="xs:double">0.2</temperature> 
<top_p m:type="xs:double">0.7</top_p> 
<max_tokens m:type="xs:int">1024</max_tokens> 
<seed m:type="xs:int">42</seed> 
<stream m:type="xs:boolean">false</stream> 
</root>

Google/deplot

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
<content>Generate underlying data table of the figure below:  
&lt;img src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII=="/&gt; 
</content> 
<role>user</role> 
</messages> 
<temperature m:type="xs:double">0.2</temperature> 
<top_p m:type="xs:double">0.7</top_p> 
<max_tokens m:type="xs:int">1024</max_tokens> 
<stream m:type="xs:boolean">false</stream> 
</root>

Google/paligemma

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
<content>Hi! What is in this image? &lt;img src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII=="/&gt; 
</content> 
<role>user</role> 
</messages> 
<temperature m:type="xs:double">0.2</temperature> 
<top_p m:type="xs:double">0.7</top_p> 
<max_tokens m:type="xs:int">1024</max_tokens> 
<stream m:type="xs:boolean">false</stream> 
</root>

microsoft/kosmos-2

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
<content>Where is the coach in this image? &lt;img src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII=="/&gt; 
</content> 
<role>user</role> 
</messages> 
<grounded_response m:type="xs:boolean">true</grounded_response> 
<temperature m:type="xs:double">0.2</temperature> 
<top_p m:type="xs:double">0.7</top_p> 
<max_tokens m:type="xs:int">1024</max_tokens> 
<response_mode>brief</response_mode> 
<task>VQA</task> 
</root>

microsoft/phi-3-vision-128k-instruct

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
<content>Hi! What is in this image? &lt;img src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII=="/&gt; 
</content> 
<role>user</role> 
</messages> 
<temperature m:type="xs:double">0.2</temperature> 
<top_p m:type="xs:double">0.7</top_p> 
<max_tokens m:type="xs:int">1024</max_tokens> 
<stream m:type="xs:boolean">false</stream> 
</root>

briaai/bria-2.3

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<prompt>Give me image of an orange cat</prompt> 
<aspect_ratio>1:1</aspect_ratio> 
<mode>text-to-image</mode> 
<negative_prompt></negative_prompt> 
<model>bria-2.3</model> 
<seed m:type="xs:int">0</seed> 
<output_format>jpeg</output_format> 
<cfg_scale m:type="xs:double">5</cfg_scale> 
<steps m:type="xs:int">30</steps> 
</root>

nvidia/nv-dinov2

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
<content> 
<type>image_url</type> 
<image_url> 
<url>data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEBLAEsAAD/4gKwSUNDX1BST0ZJTEUAAQEAAAKgbGNtcwQwAABtbnRyUkdC IFhZWiAH6AAIABUADQAcADdhY3NwQVBQTAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA9tYAAQAA AADTLWxjbXMAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA1k ZXNjAAABIAAAAEBjcHJ0AAABYAAAADZ3dHB0AAABmAAAABRjaGFkAAABrAAAACxyWFlaAAAB2AAA ABRiWFlaAAAB7AAAABRnWFlaAAACAAAAABRyVFJDAAACFAAAACBnVFJDAAACFAAAACBiVFJDAAAC FAAAACBjaHJtAAACNAAAACRkbW5kAAACWAAAACRkbWRkAAACfAAAACRtbHVjAAAAAAAAAAEAAAAM ZW5VUwAAACQAAAAcAEcASQBNAFAAIABiAHUAaQBsAHQALQBpAG4AIABzAFIARwBCbWx1YwAAAAAA AAABAAAADGVuVVMAAAAaAAAAHABQAHUAYgBsAGkAYwAgAEQAbwBtAGEAaQBuAABYWVogAAAAAAAA 9tYAAQAAAADTLXNmMzIAAAAAAAEMQgAABd7///MlAAAHkwAA/ZD///uh///9ogAAA9wAAMBuWFla IAAAAAAAAG+gAAA49QAAA5BYWVogAAAAAAAAJJ8AAA+EAAC2xFhZWiAAAAAAAABilwAAt4cAABjZ cGFyYQAAAAAAAwAAAAJmZgAA8qcAAA1ZAAAT0AAACltjaHJtAAAAAAADAAAAAKPXAABUfAAATM0A AJmaAAAmZwAAD1xtbHVjAAAAAAAAAAEAAAAMZW5VUwAAAAgAAAAcAEcASQBNAFBtbHVjAAAAAAAA AAEAAAAMZW5VUwAAAAgAAAAcAHMAUgBHAEL/2wBDACgcHiMeGSgjISMtKygwPGRBPDc3PHtYXUlk kYCZlo+AjIqgtObDoKrarYqMyP/L2u71////m8H////6/+b9//j/2wBDASstLTw1PHZBQXb4pYyl +Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj/wgARCACV AOADAREAAhEBAxEB/8QAGQAAAwEBAQAAAAAAAAAAAAAAAAECAwQF/8QAFwEBAQEBAAAAAAAAAAAA AAAAAAECA//aAAwDAQACEAMQAAAB5AEAAAwQAAAAAAGADGMAKMVAAASjRZTMYABQASAxjLVxoIkZ y0AAIzSWLNYoVONFmMaiywBQqA1NAEIRx0AA0uWBis6M3SgouWLJJJARqWMAEIRxBQMcMS53LXfO puZqo3KlJZ1mwAsYCABCOFWBRUUYWTZcu0sWZ0ijpzdJc9Z2HBQAQUAKFXHNAyi4izn1AuXXOgLm KzsuLlqOmx0DA5xjIADObY5KLObUhFVyuJ1KjfOmnNqZVYz0IYjM5EmkAAdPPox1UVZBz6jJSRVU vTjXNvOdlmkirTOiySQEFgIR3cugKmNCsLMqDSXKzOy5QzsEZqXLZAgEIViEejz2oFLHQcOoHVm8 1CZ0E2CJQE1NpUIQhE2Ajt5dAZVjAzoIObUokkmwBGqBNC5ZFYhAIDpx0ocUWNCkclZ3LNpcLIpD RqhDTQqVCsQgA6ufQplxSOgRFkkmFkUCBGJQQ02lRICoEm+dsqNCykBUACYVmZCqUYhzSsE6ZbJE STYgmmtRpLpGiFSKxCGhWZz11RYznOahOyXQkkRJNmUorLl1l0hE1BlrMoUCEjKLzqpcN4o7JWSQ STYjFWEUuudbZuWpnZNk2AgAAQUHKrmjplZJIrJEZqAMqa259MOnJBSQEAAADARaaqkBCEB//8QA IhAAAgIBBQACAwAAAAAAAAAAAAECESEQEiAwMRNBAyJA/9oACAEBAAEFAv6K6krNqHHltZgfNIoo rolgsTy4Z2HxnxjUh7iiiiuESupey90vCwbxSssfrrStYldlZlwvRCYxxxFfrRRXc/NI+2SQ9Iui OSbpRl+q87qJXrFlaYHE8PxvM/VKKE74zaFNG9G/O8+Rcpe6LhGBKI6jpVqMtrTtFoc0PlRXCaPr bWtiefkJSb08FRKJ+N4yUUYMa50ri/No3Ztztxt0RL3VSKzwyZMlca0ofsVlIm83j7PvhGVFmTPL BZZfBxTFGictIq2/eaY/4JaJtFpj4PgmWWWWX1uJsHAarqxzssvpnFs+NlD0+uEEmtiNiNhsNpt4 rjuN6N6N6JOL0htqolInDWCpdNi1sb6NzFJjbPSOGutSNw5s3MvsQiyy+hH13Lo//8QAHxEAAQME AwEAAAAAAAAAAAAAAQARQBAgITAxUGBw/9oACAEDAQE/AfbHpAhHYUazEjiWFhEywEbRqf5029qP JFDG/8QAHREAAgICAwEAAAAAAAAAAAAAAREAQBAwIVBgYf/aAAgBAgEBPwH2w0u2Z8rsxx4MAnNh u2ZzALZg3rSO3HunuWFF2P8A/8QAIRAAAQMDBAMAAAAAAAAAAAAAMQAhQAERUBAgMGFBUWD/2gAI AQEABj8CwZl21ba8eyfL+N/csyzGbg9bTqOJ68l6DQIISyiiji+5bTWwIQV6Siij8F//xAAnEAAD AAIBBQEAAgEFAAAAAAAAAREQITEgQVFhcTCBkUChsdHh8f/aAAgBAQABPyH/ABoTCEJ+CXkD1tbW YQgmcITR8EdswhBcFGaLBhLqg2idsPOHcWivX9iZ/wDRH/onpLXoo9GU+gSCRIIaEIQnUqhsahM7 i1CE7mcmQexBoKKvts0+MIQkavsT8YJDZpBZhckn00YmN34O6duPPJRSn4QhBISKWo3VHiORV7qJ T4aNBvC1c8PH14VCmjajOBO5bS/CCQhZbWh84g+LSBRd4NN3X+x4Rp8iEt6Yy1Yc80UmnRqk6Agh OY59BqaNkIQSEM7o9kNmlwitVXCGCeOwjqqfZHc5C5XI6iEUWGrlol5GuLjZHkgkJCZZskYacn2H IPbH4Etjjoac4U7Kyk228ktr/QbaTHZfkgTyPh4b8YR+cITPOMFmitGnRqvdjlN+MbsS69izFJae 0adyHHzTXljsfBDPI+iLyQhMIPRjvkUUam23yUC8R5CY3yHvZ3xSj0V2a9H2sEZhgi8kXknl0alw 9og6IsOBIx7HycHPR2nwQaNGvwQsvgZtmzgDkxQMnxrKE6NJXhHwj4R8HwX0X0fwfwIXSlurQl7u 4TuXQidCcZRUaNdQgmJmupwnsa6XRGnwQuC2NTG94WU+T6Ps+ilExMYRSlIPZntvwxuvk+hnqE+E P+FGI5nvhvDKXFKIOyvwxt+GV7GDbKUrym1wz2hnv/sRVv8Asdasqg2sPD6aJjipgewbFL+F1k2s HgpS9LbGw3vDf6cilL0//9oADAMBAAIAAwAAABALbbJttt//ACWztpASTovTStO2tG2ykk/2bPD1 JkyJohIku6vKgIv7ttttZobrmYwgpNpNJJiw7ZS0AkpZtZJ5SQyASlg+7tJoEmAbsdDinSAtI5oC /Kql/h0UZ1WWJtbu88AlI3pJkW1tnXgJ1aJMHIs2VNf5Wt6NUXA5Stptt/TpT6czgzqRNJdqYGK3 gFsu9+UFt5P9S52Ft0emStsy3NJztgg0eAyVeSN904AtqVMA2tp5butEgDF1uSRtApGuAG22wOMp JJ//xAAeEQADAAEFAQEAAAAAAAAAAAAAAREQICEwMUBBUf/aAAgBAwEBPxD0Tka1TD8T2zCEEkRj xCE50h470pj8nTC7J9Ho7Int4IQZcIr1P1zzLyhJD+hF8Y2WE4N3Y2KcFKUuJoeUN4SpEHGroT8D y3naxpN2U09rivBBImKMmxeApSlKMWHj6Jj754QhCaOiiy/O8303yUvF8xNKVGphaYTVCEIQmh6b cLhomI2yxspdSVHK5UxOkDaWDb4rtzNufBt8N3ifF//EAB4RAAMAAgMBAQEAAAAAAAAAAAABERAg ITAxQEFQ/9oACAECAQE/EP4TZWJ7XC+Jc4nBSlG2JoUzS97FjzMw1+i8+O5Muq4K9+Gi9JhkQsvF 8tfEWWNvAm/UJXhKiHKObu8IQhMXFwvBYYlcNj8EhNpzRrWbUpcp4tFnkhLwR1fD+Ci4KXkuGhcr 0g+tMQhBeDZcIfg0LzvjIyaJtFErh5Wz+M1SC6GvhuF8MITpRSi0fAncMpSlLq8UpS9SYsPqaHok JZmkG4KnO2DHIhAkuqc9zEuRYS6l1f/EACgQAQACAgICAQQCAgMAAAAAAAEAESExQVEQYZFxgaHR ILEw8EDB8f/aAAgBAQABPxD/AIdSoEvCRdRZxCT+Woz2BdKzlp9CAq1NNcSpUylvAm1rtwQgOHLL HX3KOJULQgFwDpOEiYZTqJWZP4VKglAlkWN/b3LXN9oYs7lressrid6YzSle1wpEVeoLpw1BNaxU Ear24hjprvvxJIWkyUktaTEDpGNowxXioEC4RGi8e2Iq7lRs+YW12vUekqt6hcrPqLpCqaJsK6cQ mXH9Q0S9EQbzGPhc4KUGe4laLgDRKlRIkSVKlQ8A9PG5cNfbw2kGRGl+CBrykQct/ucp+6aFk5SK 3DeIykX2RiHlWZeA8wdCVKlSpUSJ4WhGSDDfYZ+87EFpleyEBpR3HUwKaYIaqb3LKthpjL6SLiF3 2xglGyoSnLqCoIODuZggydRoWwBjzsEIUlkqVEiSpUBOKvE6AgwS4ib3CcSBVMXvAhKeH90Awv1y ib096+Y5QSZeGG/cUbnGeIuE2w43L1U4YIpgAoleLAW2SkPEdGSPQAxUscwHIr6IeA8EpGZljxC5 VAz0QqwzMUZxjmF3nMNvDEN2h6gMsoziYSbIGn3FVK2sL5nA79wAjXJCWp8aBfeEQWfRBTnEU6lu Jayl3PVDwHmDN7TE5jiEDgbHbGNwu7iqmjZLdIwMICC2XhRcC2PhIc6ItAdDXxKARq00+pLFL4h3 r9484PvEtiV8n7T0SWcRfEkRdEeTwfR4AwINCGJRrB4CX1PQdEQ0BdBCxbrL7iKNKpKiHuDA4P7h MHIqoLTD3BqEP8R/UCbXhxxHktvNETikxx8pfEkfUfaPZUeaFc/nO4R94+5PqIexD3IyQmo1WbvU SiXTcA87a1EatGCKsqtUsUaDAShxd7Ygw03OL94FUTfcbGEJU5WyWAAvKrCfuP8AqInXHn/KL5/K J6/Mev8AM9fwEdQHUA8QPUp1L9SwBHsZqK/cUX32K25uYBo2ykG11/cvelzmg2si2Vl2ZiVBhmek 4emJuqTdGGPt+I+34j7fiY7/ABMdzHcx3CFQSLwGXGqrqcKo0RFmO5Y67gcAZkeY00RwdTiolMVN pfqIZaFzRxBszNNS4jWhTuf+RLyVJ9E+hPpT6EqCKdSnUK6mOpjqY6jJLMM23pG3R9MeA0vEB1zM v0iQW+paq43K6iWX14qGFGmPDF7Su0amO5iEHpA6h9QniAoh4uXNwwPaOK2miB056ZSFEt8TgS+v UrGIbVdPEZ5KhENcjHqfmcFY8KiO3xH/AGIs4Qg8/RixGGUm2I7ZTB2PxCvbNqR5qNciUEU5LIbz 44OsKchAtH6/pGuHiC2LjZV1FOoHTF9Reo+kfSX4CI04Q/EKf+iB/RHpipm5e5ipaYqtlszNsH0Y H+/wApJIdgt7iAAfqRELFn2Rv1LRuMuXCFINASVWUl6rqLco/axbz4XLly5cuCxWNwWJszKQXH2n sj7RlhX8BlAl+hGYuI/4nR41QVRUtFiy5c//2Q==</url> 
</image_url> 
</content> 
</messages> 
</root>

nvidia/nv-grounding-dino

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<model>Grounding Dino</model> 
<messages m:isArray="true"> 
<role>user</role> 
<content m:isArray="true"> 
<type>text</type> 
<text>person, cars, bus, backpack</text> 
</content> 
<content m:isArray="true"> 
<type>media_url</type> 
<media_url> 
<url>data:image/jpeg;base64,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</url> 
</media_url> 
</content> 
</messages> 
<threshold m:type="xs:double">0.7</threshold> 
</root>

stabilityai/stable-diffusion-3-medium

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<aspect_ratio>1:1</aspect_ratio> 
<cfg_scale m:type="xs:double">5</cfg_scale> 
<mode>text-to-image</mode> 
<model>sd3</model> 
<negative_prompt></negative_prompt> 
<output_format>jpeg</output_format> 
<prompt>An orange cat</prompt> 
<seed m:type="xs:int">0</seed> 
<steps m:type="xs:int">50</steps> 
</root>

stabilityai/sdxl-turbo

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<height m:type="xs:int">512</height> 
<width m:type="xs:int">512</width> 
<text_prompts m:isArray="true"> 
<text>A photo of a Shiba Inu dog with a backpack riding a bike</text> 
<weight m:type="xs:double">1</weight> 
</text_prompts> 
<cfg_scale m:type="xs:double">0</cfg_scale> 
<clip_guidance_preset>NONE</clip_guidance_preset> 
<sampler>K_EULER_ANCESTRAL</sampler> 
<samples m:type="xs:int">1</samples> 
<seed m:type="xs:int">0</seed> 
<steps m:type="xs:int">4</steps> 
<style_preset>none</style_preset> 
</root>

stabilityai/stable-diffusion-xl

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<height m:type="xs:int">1024</height> 
<width m:type="xs:int">1024</width> 
<text_prompts m:isArray="true"> 
<text>A photo of a Shiba Inu dog with a backpack riding a bike</text> 
<weight m:type="xs:double">1</weight> 
</text_prompts> 
<cfg_scale m:type="xs:double">5</cfg_scale> 
<clip_guidance_preset>NONE</clip_guidance_preset> 
<sampler>K_DPM_2_ANCESTRAL</sampler> 
<samples m:type="xs:int">1</samples> 
<seed m:type="xs:int">0</seed> 
<steps m:type="xs:int">25</steps> 
<style_preset>none</style_preset> 
</root>

stabilityai/stable-video-diffusion

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<image>data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAIAAACQd1PeAAAAEElEQVR4nGK6HcwNCAAA//8DTgE8HuxwEQAAAABJRU5ErkJggg==</image> 
<seed m:type="xs:int">0</seed> 
<cfg_scale m:type="xs:double">1.8</cfg_scale> 
<motion_bucket_id m:type="xs:int">127</motion_bucket_id> 
</root>

The Invoke Climate Simulation Model
action invokes the Climate Simulation API models, such as nvidia/fourcastnet. It runs models for climate prediction and environmental simulation.

The connector is tested with the Nvidia/fourcastnet model.

You must provide values to the input fields, such as payload, version, and NVCF_POLL_SECONDS.

For example, consider the nvidia/fourcastnet model with the following endpoint URL:

https://climate.api.nvidia.com/v1/nvidia/fourcastnet

In this URL,

is the API version and

nvidia

, and

fourcastnet

are individual URL path segments.

The following snippet is a payload example:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<input_id m:type="xs:int">0</input_id> 
<variables>w10m,z500</variables> 
<simulation_length m:type="xs:int">4</simulation_length> 
<ensemble_size m:type="xs:int">1</ensemble_size> 
<noise_amplitude m:type="xs:double">0</noise_amplitude> 
</root>

The following snippet is a sample response:

200 OK
fourcastnet.zip

The Invoke Medical Imaging Model
action invokes the Healthcare Medical Imaging API models, such as nvidia/vista-3d and nvidia/maisi. It invokes models for analyzing medical images, such as scans or X-rays.

The connector is tested with the following models:

nvidia/vista-3d

nvidia/maisi

You must provide values to the input fields, such as payload, version, and NVCF_POLL_SECONDS.

For example, consider the nvidia/vista-3d model with the following endpoint URL:

https://health.api.nvidia.com/v1/medicalimaging/nvidia/vista-3d

In this URL,

is the API version and

medicalimaging

nvidia

, and

vista-3d

are individual URL path segments.

The following snippet is a payload example:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<image>https://assets.ngc.nvidia.com/products/api-catalog/vista3d/example-1.nii.gz</image> 
<prompts> 
<classes m:isArray="true">spleen</classes> 
<points> 
<spleen m:isArray="true"> 
<anonymous_array m:isArray="true" m:type="xs:int">97</anonymous_array> 
<anonymous_array m:isArray="true" m:type="xs:int">249</anonymous_array> 
<anonymous_array m:isArray="true" m:type="xs:int">87</anonymous_array> 
</spleen> 
</points> 
</prompts> 
</root>

This XML structure uses metadata attributes such as

m:isArray

and

m:type

to represent JSON-like arrays and types. It also leverages custom fields named

anonymous_array

to support anonymous arrays within the payload structure.

The following snippet is an equivalent JSON representation:

{ 
 "image": "https://assets.ngc.nvidia.com/products/api-catalog/vista3d/example-1.nii.gz", 
 "prompts": { 
  "classes": [ 
   "spleen" 
  ], 
  "points": { 
   "spleen": [ 
    [ 
     97, 
     249, 
     87 
    ] 
   ] 
  } 
 } 
}

The following snippet is a sample response:

200 OK 
vista-3d.zip

The Invoke Multimodal Model
action invokes the Multimodal and Visual API models, such as nvidia/neva-22b and nvidia/nv-grounding-dino. It invokes models that process and combine different data types, such as text, images, or video.

The connector is tested with the following models:

adept/fuyu-8b

Google/deplot

google/paligemma

microsoft/kosmos-2

microsoft/phi-3-vision-128k-instruct

nvidia/neva-22b

briaai/bria-2.3

nvidia/nv-dinov2

nvidia/nv-grounding-dino

stabilityai/stable-diffusion-3-medium

stabilityai/sdxl-turbo

stabilityai/stable-diffusion-xl

stabilityai/stable-video-diffusion

You must provide values to the input fields, such as payload, version, and NVCF_POLL_SECONDS.

For example, consider the nvidia/neva-22b model with the following endpoint URL:

https://ai.api.nvidia.com/v1/vlm/nvidia/neva-22b

In this URL,

is the API version and

vlm

nvidia

, and

neva-22b

are individual URL path segments.

The following snippet is a payload example:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data"> 
<messages m:isArray="true"> 
<content>Hi! What is in this image? &lt;img src="data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEBLAEsAAD/4gKwSUNDX1BST0ZJTEUAAQEAAAKgbGNtcwQwAABtbnRyUkdC IFhZWiAH6AAIABUADQAcADdhY3NwQVBQTAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA9tYAAQAA AADTLWxjbXMAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA1k ZXNjAAABIAAAAEBjcHJ0AAABYAAAADZ3dHB0AAABmAAAABRjaGFkAAABrAAAACxyWFlaAAAB2AAA ABRiWFlaAAAB7AAAABRnWFlaAAACAAAAABRyVFJDAAACFAAAACBnVFJDAAACFAAAACBiVFJDAAAC FAAAACBjaHJtAAACNAAAACRkbW5kAAACWAAAACRkbWRkAAACfAAAACRtbHVjAAAAAAAAAAEAAAAM ZW5VUwAAACQAAAAcAEcASQBNAFAAIABiAHUAaQBsAHQALQBpAG4AIABzAFIARwBCbWx1YwAAAAAA AAABAAAADGVuVVMAAAAaAAAAHABQAHUAYgBsAGkAYwAgAEQAbwBtAGEAaQBuAABYWVogAAAAAAAA 9tYAAQAAAADTLXNmMzIAAAAAAAEMQgAABd7///MlAAAHkwAA/ZD///uh///9ogAAA9wAAMBuWFla IAAAAAAAAG+gAAA49QAAA5BYWVogAAAAAAAAJJ8AAA+EAAC2xFhZWiAAAAAAAABilwAAt4cAABjZ cGFyYQAAAAAAAwAAAAJmZgAA8qcAAA1ZAAAT0AAACltjaHJtAAAAAAADAAAAAKPXAABUfAAATM0A AJmaAAAmZwAAD1xtbHVjAAAAAAAAAAEAAAAMZW5VUwAAAAgAAAAcAEcASQBNAFBtbHVjAAAAAAAA AAEAAAAMZW5VUwAAAAgAAAAcAHMAUgBHAEL/2wBDACgcHiMeGSgjISMtKygwPGRBPDc3PHtYXUlk kYCZlo+AjIqgtObDoKrarYqMyP/L2u71////m8H////6/+b9//j/2wBDASstLTw1PHZBQXb4pYyl +Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj4+Pj/wgARCACV AOADAREAAhEBAxEB/8QAGQAAAwEBAQAAAAAAAAAAAAAAAAECAwQF/8QAFwEBAQEBAAAAAAAAAAAA AAAAAAECA//aAAwDAQACEAMQAAAB5AEAAAwQAAAAAAGADGMAKMVAAASjRZTMYABQASAxjLVxoIkZ y0AAIzSWLNYoVONFmMaiywBQqA1NAEIRx0AA0uWBis6M3SgouWLJJJARqWMAEIRxBQMcMS53LXfO puZqo3KlJZ1mwAsYCABCOFWBRUUYWTZcu0sWZ0ijpzdJc9Z2HBQAQUAKFXHNAyi4izn1AuXXOgLm KzsuLlqOmx0DA5xjIADObY5KLObUhFVyuJ1KjfOmnNqZVYz0IYjM5EmkAAdPPox1UVZBz6jJSRVU vTjXNvOdlmkirTOiySQEFgIR3cugKmNCsLMqDSXKzOy5QzsEZqXLZAgEIViEejz2oFLHQcOoHVm8 1CZ0E2CJQE1NpUIQhE2Ajt5dAZVjAzoIObUokkmwBGqBNC5ZFYhAIDpx0ocUWNCkclZ3LNpcLIpD RqhDTQqVCsQgA6ufQplxSOgRFkkmFkUCBGJQQ02lRICoEm+dsqNCykBUACYVmZCqUYhzSsE6ZbJE STYgmmtRpLpGiFSKxCGhWZz11RYznOahOyXQkkRJNmUorLl1l0hE1BlrMoUCEjKLzqpcN4o7JWSQ STYjFWEUuudbZuWpnZNk2AgAAQUHKrmjplZJIrJEZqAMqa259MOnJBSQEAAADARaaqkBCEB//8QA IhAAAgIBBQACAwAAAAAAAAAAAAECESEQEiAwMRNBAyJA/9oACAEBAAEFAv6K6krNqHHltZgfNIoo rolgsTy4Z2HxnxjUh7iiiiuESupey90vCwbxSssfrrStYldlZlwvRCYxxxFfrRRXc/NI+2SQ9Iui OSbpRl+q87qJXrFlaYHE8PxvM/VKKE74zaFNG9G/O8+Rcpe6LhGBKI6jpVqMtrTtFoc0PlRXCaPr bWtiefkJSb08FRKJ+N4yUUYMa50ri/No3Ztztxt0RL3VSKzwyZMlca0ofsVlIm83j7PvhGVFmTPL BZZfBxTFGictIq2/eaY/4JaJtFpj4PgmWWWWX1uJsHAarqxzssvpnFs+NlD0+uEEmtiNiNhsNpt4 rjuN6N6N6JOL0htqolInDWCpdNi1sb6NzFJjbPSOGutSNw5s3MvsQiyy+hH13Lo//8QAHxEAAQME AwEAAAAAAAAAAAAAAQARQBAgITAxUGBw/9oACAEDAQE/AfbHpAhHYUazEjiWFhEywEbRqf5029qP JFDG/8QAHREAAgICAwEAAAAAAAAAAAAAAREAQBAwIVBgYf/aAAgBAgEBPwH2w0u2Z8rsxx4MAnNh u2ZzALZg3rSO3HunuWFF2P8A/8QAIRAAAQMDBAMAAAAAAAAAAAAAMQAhQAERUBAgMGFBUWD/2gAI AQEABj8CwZl21ba8eyfL+N/csyzGbg9bTqOJ68l6DQIISyiiji+5bTWwIQV6Siij8F//xAAnEAAD AAIBBQEAAgEFAAAAAAAAAREQITEgQVFhcTCBkUChsdHh8f/aAAgBAQABPyH/ABoTCEJ+CXkD1tbW YQgmcITR8EdswhBcFGaLBhLqg2idsPOHcWivX9iZ/wDRH/onpLXoo9GU+gSCRIIaEIQnUqhsahM7 i1CE7mcmQexBoKKvts0+MIQkavsT8YJDZpBZhckn00YmN34O6duPPJRSn4QhBISKWo3VHiORV7qJ T4aNBvC1c8PH14VCmjajOBO5bS/CCQhZbWh84g+LSBRd4NN3X+x4Rp8iEt6Yy1Yc80UmnRqk6Agh OY59BqaNkIQSEM7o9kNmlwitVXCGCeOwjqqfZHc5C5XI6iEUWGrlol5GuLjZHkgkJCZZskYacn2H IPbH4Etjjoac4U7Kyk228ktr/QbaTHZfkgTyPh4b8YR+cITPOMFmitGnRqvdjlN+MbsS69izFJae 0adyHHzTXljsfBDPI+iLyQhMIPRjvkUUam23yUC8R5CY3yHvZ3xSj0V2a9H2sEZhgi8kXknl0alw 9og6IsOBIx7HycHPR2nwQaNGvwQsvgZtmzgDkxQMnxrKE6NJXhHwj4R8HwX0X0fwfwIXSlurQl7u 4TuXQidCcZRUaNdQgmJmupwnsa6XRGnwQuC2NTG94WU+T6Ps+ilExMYRSlIPZntvwxuvk+hnqE+E P+FGI5nvhvDKXFKIOyvwxt+GV7GDbKUrym1wz2hnv/sRVv8Asdasqg2sPD6aJjipgewbFL+F1k2s HgpS9LbGw3vDf6cilL0//9oADAMBAAIAAwAAABALbbJttt//ACWztpASTovTStO2tG2ykk/2bPD1 JkyJohIku6vKgIv7ttttZobrmYwgpNpNJJiw7ZS0AkpZtZJ5SQyASlg+7tJoEmAbsdDinSAtI5oC /Kql/h0UZ1WWJtbu88AlI3pJkW1tnXgJ1aJMHIs2VNf5Wt6NUXA5Stptt/TpT6czgzqRNJdqYGK3 gFsu9+UFt5P9S52Ft0emStsy3NJztgg0eAyVeSN904AtqVMA2tp5butEgDF1uSRtApGuAG22wOMp JJ//xAAeEQADAAEFAQEAAAAAAAAAAAAAAREQICEwMUBBUf/aAAgBAwEBPxD0Tka1TD8T2zCEEkRj xCE50h470pj8nTC7J9Ho7Int4IQZcIr1P1zzLyhJD+hF8Y2WE4N3Y2KcFKUuJoeUN4SpEHGroT8D y3naxpN2U09rivBBImKMmxeApSlKMWHj6Jj754QhCaOiiy/O8303yUvF8xNKVGphaYTVCEIQmh6b cLhomI2yxspdSVHK5UxOkDaWDb4rtzNufBt8N3ifF//EAB4RAAMAAgMBAQEAAAAAAAAAAAABERAg ITAxQEFQ/9oACAECAQE/EP4TZWJ7XC+Jc4nBSlG2JoUzS97FjzMw1+i8+O5Muq4K9+Gi9JhkQsvF 8tfEWWNvAm/UJXhKiHKObu8IQhMXFwvBYYlcNj8EhNpzRrWbUpcp4tFnkhLwR1fD+Ci4KXkuGhcr 0g+tMQhBeDZcIfg0LzvjIyaJtFErh5Wz+M1SC6GvhuF8MITpRSi0fAncMpSlLq8UpS9SYsPqaHok JZmkG4KnO2DHIhAkuqc9zEuRYS6l1f/EACgQAQACAgICAQQCAgMAAAAAAAEAESExQVEQYZFxgaHR ILEw8EDB8f/aAAgBAQABPxD/AIdSoEvCRdRZxCT+Woz2BdKzlp9CAq1NNcSpUylvAm1rtwQgOHLL HX3KOJULQgFwDpOEiYZTqJWZP4VKglAlkWN/b3LXN9oYs7lressrid6YzSle1wpEVeoLpw1BNaxU Ear24hjprvvxJIWkyUktaTEDpGNowxXioEC4RGi8e2Iq7lRs+YW12vUekqt6hcrPqLpCqaJsK6cQ mXH9Q0S9EQbzGPhc4KUGe4laLgDRKlRIkSVKlQ8A9PG5cNfbw2kGRGl+CBrykQct/ucp+6aFk5SK 3DeIykX2RiHlWZeA8wdCVKlSpUSJ4WhGSDDfYZ+87EFpleyEBpR3HUwKaYIaqb3LKthpjL6SLiF3 2xglGyoSnLqCoIODuZggydRoWwBjzsEIUlkqVEiSpUBOKvE6AgwS4ib3CcSBVMXvAhKeH90Awv1y ib096+Y5QSZeGG/cUbnGeIuE2w43L1U4YIpgAoleLAW2SkPEdGSPQAxUscwHIr6IeA8EpGZljxC5 VAz0QqwzMUZxjmF3nMNvDEN2h6gMsoziYSbIGn3FVK2sL5nA79wAjXJCWp8aBfeEQWfRBTnEU6lu Jayl3PVDwHmDN7TE5jiEDgbHbGNwu7iqmjZLdIwMICC2XhRcC2PhIc6ItAdDXxKARq00+pLFL4h3 r9484PvEtiV8n7T0SWcRfEkRdEeTwfR4AwINCGJRrB4CX1PQdEQ0BdBCxbrL7iKNKpKiHuDA4P7h MHIqoLTD3BqEP8R/UCbXhxxHktvNETikxx8pfEkfUfaPZUeaFc/nO4R94+5PqIexD3IyQmo1WbvU SiXTcA87a1EatGCKsqtUsUaDAShxd7Ygw03OL94FUTfcbGEJU5WyWAAvKrCfuP8AqInXHn/KL5/K J6/Mev8AM9fwEdQHUA8QPUp1L9SwBHsZqK/cUX32K25uYBo2ykG11/cvelzmg2si2Vl2ZiVBhmek 4emJuqTdGGPt+I+34j7fiY7/ABMdzHcx3CFQSLwGXGqrqcKo0RFmO5Y67gcAZkeY00RwdTiolMVN pfqIZaFzRxBszNNS4jWhTuf+RLyVJ9E+hPpT6EqCKdSnUK6mOpjqY6jJLMM23pG3R9MeA0vEB1zM v0iQW+paq43K6iWX14qGFGmPDF7Su0amO5iEHpA6h9QniAoh4uXNwwPaOK2miB056ZSFEt8TgS+v UrGIbVdPEZ5KhENcjHqfmcFY8KiO3xH/AGIs4Qg8/RixGGUm2I7ZTB2PxCvbNqR5qNciUEU5LIbz 44OsKchAtH6/pGuHiC2LjZV1FOoHTF9Reo+kfSX4CI04Q/EKf+iB/RHpipm5e5ipaYqtlszNsH0Y H+/wApJIdgt7iAAfqRELFn2Rv1LRuMuXCFINASVWUl6rqLco/axbz4XLly5cuCxWNwWJszKQXH2n sj7RlhX8BlAl+hGYuI/4nR41QVRUtFiy5c//2Q=="/&gt; 
</content> 
<name>John</name> 
<role>user</role> 
</messages> 
<temperature m:type="xs:double">0.2</temperature> 
<top_p m:type="xs:double">0.7</top_p> 
<max_tokens m:type="xs:int">1024</max_tokens> 
<seed m:type="xs:int">42</seed> 
<stream m:type="xs:boolean">false</stream> 
</root>

The following snippet is a sample response:

<root> 
  <usage> 
   <completion_tokens>37</completion_tokens> 
   <prompt_tokens>754</prompt_tokens> 
   <total_tokens>791</total_tokens> 
  </usage> 
  <id>cd05322e-71e5-4b01-a76f-153d96b7a8c0</id> 
  <choices> 
   <finish_reason>stop</finish_reason> 
   <index>0</index> 
   <message> 
     <role>assistant</role> 
     <name/> 
     <content>This image shows a motorcycle on a racetrack. The motorcycle is a Kawasaki Ninja H2R. The motorcycle is a super sport motorcycle. It is a track-only motorcycle.</content> 
   </message> 
  </choices> 
</root>

The Invoke Basic Model
action invokes the basic models for chat completions, function calling, text reranking, and structured output actions. It executes actions, such as chat completions, function calling, embeddings, text reranking, and structured output actions.

You must provide values to the input fields, such as host, version, segments, and payload.

The following snippet is an example of executing Chat completion action payload:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data">
    <model>mistralai/mixtral-8x7b-instruct-v0.1</model>
    <max_tokens m:type="xs:int">1024</max_tokens>
    <stream m:type="xs:boolean">false</stream>
    <temperature m:type="xs:double">1.2</temperature>
    <top_p m:type="xs:double">0.8</top_p>
    <frequency_penalty m:type="xs:double">0</frequency_penalty>
    <presence_penalty m:type="xs:double">0</presence_penalty>
    <seed m:type="xs:int">1024</seed>
    <messages m:isArray="true">
        <role>user</role>
        <content>You are a helpful assistant.</content>
    </messages>
    <messages m:isArray="true">
        <role>assistant</role>
        <content>You are a helpful assistant because you help people with their problems and you are always there for them when they need you.</content>
    </messages>
    <messages m:isArray="true">
        <role>user</role>
        <content>Who is Leo Messi.</content>
    </messages>
</root>

The following snippet is an example to execute Chat Completion Function Call action payload:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data">
    <model>meta/llama-3.1-70b-instruct</model>
    <messages m:isArray="true">
        <role>user</role>
        <content>My name is John and I'm 30 years old</content>
    </messages>
    <tools m:isArray="true">
        <type>function</type>
        <function>
            <name>get_name_age</name>
            <description>Please provide your name and age</description>
            <parameters>
                <type>object</type>
                <properties>
                    <name>
                        <type>string</type>
                        <description>The name of participant</description>
                    </name>
                    <age>
                        <type>integer</type>
                        <description>The age of participant</description>
                    </age>
                </properties>
                <required m:isArray="true">name</required>
                <required m:isArray="true">age</required>
            </parameters>
        </function>
    </tools>
    <tool_choice>
        <type>function</type>
        <function>
            <name>get_name_age</name>
        </function>
    </tool_choice>
</root>

The following snippet is an example to execute Embeddings action payload:

<root xmlns:m="urn:informatica:ae:xquery:json2xml:meta-data">
    <input m:isArray="true">Paris is the capital of France</input>
    <model>nvidia/nv-embedqa-e5-v5</model>
    <input_type>query</input_type>
</root>

Reranking and Structured Output actions can be executed on local environment, Docker image, or on-premise hosting eтvironment.

NVIDIA NIM connections

Download Guide

Watch

Comments

Cloud Application Integration Homepage