feat: Adding OpenAI Compatible Prompts API

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

llama_stack/distributions/nvidia/run.yaml

@@ -134,6 +134,11 @@ models:
   provider_id: nvidia
   provider_model_id: meta/llama-3.3-70b-instruct
   model_type: llm
+- metadata: {}
+  model_id: nvidia/vila
+  provider_id: nvidia
+  provider_model_id: nvidia/vila
+  model_type: llm
 - metadata:
     embedding_dimension: 2048
     context_length: 8192

llama_stack/distributions/open-benchmark/open_benchmark.py

@@ -43,7 +43,7 @@ def get_inference_providers() -> tuple[list[Provider], dict[str, list[ProviderMo
             "openai",
             [
                 ProviderModelEntry(
-                    provider_model_id="openai/gpt-4o",
+                    provider_model_id="gpt-4o",
                     model_type=ModelType.llm,
                 )
             ],
@@ -53,7 +53,7 @@ def get_inference_providers() -> tuple[list[Provider], dict[str, list[ProviderMo
             "anthropic",
             [
                 ProviderModelEntry(
-                    provider_model_id="anthropic/claude-3-5-sonnet-latest",
+                    provider_model_id="claude-3-5-sonnet-latest",
                     model_type=ModelType.llm,
                 )
             ],
@@ -206,13 +206,6 @@ def get_distribution_template() -> DistributionTemplate:
                 uri="huggingface://datasets/llamastack/math_500?split=test",
             ),
         ),
-        DatasetInput(
-            dataset_id="bfcl",
-            purpose=DatasetPurpose.eval_messages_answer,
-            source=URIDataSource(
-                uri="huggingface://datasets/llamastack/bfcl_v3?split=train",
-            ),
-        ),
         DatasetInput(
             dataset_id="ifeval",
             purpose=DatasetPurpose.eval_messages_answer,
@@ -250,11 +243,6 @@ def get_distribution_template() -> DistributionTemplate:
             dataset_id="math_500",
             scoring_functions=["basic::regex_parser_math_response"],
         ),
-        BenchmarkInput(
-            benchmark_id="meta-reference-bfcl",
-            dataset_id="bfcl",
-            scoring_functions=["basic::bfcl"],
-        ),
         BenchmarkInput(
             benchmark_id="meta-reference-ifeval",
             dataset_id="ifeval",

llama_stack/distributions/open-benchmark/run.yaml

@@ -136,14 +136,14 @@ inference_store:
   db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/open-benchmark}/inference_store.db
 models:
 - metadata: {}
-  model_id: openai/gpt-4o
+  model_id: gpt-4o
   provider_id: openai
-  provider_model_id: openai/gpt-4o
+  provider_model_id: gpt-4o
   model_type: llm
 - metadata: {}
-  model_id: anthropic/claude-3-5-sonnet-latest
+  model_id: claude-3-5-sonnet-latest
   provider_id: anthropic
-  provider_model_id: anthropic/claude-3-5-sonnet-latest
+  provider_model_id: claude-3-5-sonnet-latest
   model_type: llm
 - metadata: {}
   model_id: gemini/gemini-1.5-flash
@@ -188,12 +188,6 @@ datasets:
     uri: huggingface://datasets/llamastack/math_500?split=test
   metadata: {}
   dataset_id: math_500
-- purpose: eval/messages-answer
-  source:
-    type: uri
-    uri: huggingface://datasets/llamastack/bfcl_v3?split=train
-  metadata: {}
-  dataset_id: bfcl
 - purpose: eval/messages-answer
   source:
     type: uri
@@ -228,11 +222,6 @@ benchmarks:
   - basic::regex_parser_math_response
   metadata: {}
   benchmark_id: meta-reference-math-500
-- dataset_id: bfcl
-  scoring_functions:
-  - basic::bfcl
-  metadata: {}
-  benchmark_id: meta-reference-bfcl
 - dataset_id: ifeval
   scoring_functions:
   - basic::ifeval