add nvidia new models

app.py

@@ -7,7 +7,6 @@ from app_fal import demo as demo_fal
 from app_fireworks import demo as demo_fireworks
 from app_huggingface import demo as demo_huggingface
 from app_meta import demo as demo_meta
-from app_nvidia import demo as demo_nvidia
 from app_omini import demo as demo_omini
 from app_paligemma import demo as demo_paligemma
 from app_perplexity import demo as demo_perplexity
@@ -36,11 +35,13 @@ from app_langchain import demo as demo_langchain
 from app_mistral import demo as demo_mistral
 from app_minimax import demo as demo_minimax
 from app_minimax_coder import demo as demo_minimax_coder
+from app_nvidia import demo as demo_nvidia
 from utils import get_app
 
 # Create mapping of providers to their demos
 PROVIDERS = {
     "Minimax Coder": demo_minimax_coder,
+    "NVIDIA": demo_nvidia,
     "Minimax": demo_minimax,
     "Gemini Camera": demo_gemini_camera,
     "Mistral": demo_mistral,
@@ -77,7 +78,6 @@ PROVIDERS = {
     "Allen AI": demo_allenai,
     "Perplexity": demo_perplexity,
     "Experimental": demo_experimental,
-    "NVIDIA": demo_nvidia,
 }
 
 demo = get_app(

app_nvidia.py

@@ -1,51 +1,21 @@
-import os
+import ai_gradio
 
-import nvidia_gradio
+from utils_ai_gradio import get_app
 
-from utils import get_app
+# Get the nvidia models but keep their full names for loading
+NVIDIA_MODELS_FULL = [k for k in ai_gradio.registry.keys() if k.startswith("nvidia:")]
 
+# Create display names without the prefix
+NVIDIA_MODELS_DISPLAY = [k.replace("nvidia:", "") for k in NVIDIA_MODELS_FULL]
+
+
+# Create and launch the interface using get_app utility
 demo = get_app(
-    models=[
-        "nvidia/llama3-chatqa-1.5-70b",
-        "nvidia/llama3-chatqa-1.5-8b",
-        "nvidia-nemotron-4-340b-instruct",
-        "meta/llama-3.1-70b-instruct",
-        "meta/codellama-70b",
-        "meta/llama2-70b",
-        "meta/llama3-8b",
-        "meta/llama3-70b",
-        "mistralai/codestral-22b-instruct-v0.1",
-        "mistralai/mathstral-7b-v0.1",
-        "mistralai/mistral-large-2-instruct",
-        "mistralai/mistral-7b-instruct",
-        "mistralai/mistral-7b-instruct-v0.3",
-        "mistralai/mixtral-8x7b-instruct",
-        "mistralai/mixtral-8x22b-instruct",
-        "mistralai/mistral-large",
-        "google/gemma-2b",
-        "google/gemma-7b",
-        "google/gemma-2-2b-it",
-        "google/gemma-2-9b-it",
-        "google/gemma-2-27b-it",
-        "google/codegemma-1.1-7b",
-        "google/codegemma-7b",
-        "google/recurrentgemma-2b",
-        "google/shieldgemma-9b",
-        "microsoft/phi-3-medium-128k-instruct",
-        "microsoft/phi-3-medium-4k-instruct",
-        "microsoft/phi-3-mini-128k-instruct",
-        "microsoft/phi-3-mini-4k-instruct",
-        "microsoft/phi-3-small-128k-instruct",
-        "microsoft/phi-3-small-8k-instruct",
-        "qwen/qwen2-7b-instruct",
-        "databricks/dbrx-instruct",
-        "deepseek-ai/deepseek-coder-6.7b-instruct",
-        "upstage/solar-10.7b-instruct",
-        "snowflake/arctic",
-    ],
-    default_model="meta/llama-3.1-70b-instruct",
-    src=nvidia_gradio.registry,
-    accept_token=not os.getenv("NVIDIA_API_KEY"),
+    models=NVIDIA_MODELS_FULL,  # Use the full names with prefix
+    default_model=NVIDIA_MODELS_FULL[0],
+    dropdown_label="Select Nvidia Model",
+    choices=NVIDIA_MODELS_DISPLAY,  # Display names without prefix
+    fill_height=True,
 )
 
 if __name__ == "__main__":

pyproject.toml

@@ -38,7 +38,7 @@ dependencies = [
     "langchain>=0.3.14",
     "chromadb>=0.5.23",
     "openai>=1.55.0",
-    "ai-gradio[crewai,deepseek,gemini,groq,hyperbolic,openai,smolagents,transformers, langchain, mistral,minimax]>=0.2.18",
+    "ai-gradio[crewai,deepseek,gemini,groq,hyperbolic,openai,smolagents,transformers, langchain, mistral,minimax,nvidia]>=0.2.19",
 ]
 
 [tool.uv.sources]

requirements.txt

@@ -2,7 +2,7 @@
 #    uv pip compile pyproject.toml -o requirements.txt
 accelerate==1.2.1
     # via ai-gradio
-ai-gradio==0.2.18
+ai-gradio==0.2.19
     # via anychat (pyproject.toml)
 aiofiles==23.2.1
     # via gradio