add-llama3-for-nvidia-api-catalog (#3631)

api/core/model_runtime/model_providers/nvidia/llm/_position.yaml

@@ -1,5 +1,7 @@
 - google/gemma-7b
 - google/codegemma-7b
 - meta/llama2-70b
+- meta/llama3-8b
+- meta/llama3-70b
 - mistralai/mixtral-8x7b-instruct-v0.1
 - fuyu-8b

api/core/model_runtime/model_providers/nvidia/llm/codegemma-7b.yaml

@@ -11,13 +11,19 @@ model_properties:
 parameter_rules:
   - name: temperature
     use_template: temperature
+    min: 0
+    max: 1
+    default: 0.5
   - name: top_p
     use_template: top_p
+    min: 0
+    max: 1
+    default: 1
   - name: max_tokens
     use_template: max_tokens
-    default: 1024
     min: 1
     max: 1024
+    default: 1024
   - name: frequency_penalty
     use_template: frequency_penalty
     min: -2

api/core/model_runtime/model_providers/nvidia/llm/fuyu-8b.yaml

@@ -22,6 +22,6 @@ parameter_rules:
     max: 1
   - name: max_tokens
     use_template: max_tokens
-    default: 512
+    default: 1024
     min: 1
     max: 1024

api/core/model_runtime/model_providers/nvidia/llm/gemma-7b.yaml

@@ -11,13 +11,19 @@ model_properties:
 parameter_rules:
   - name: temperature
     use_template: temperature
+    min: 0
+    max: 1
+    default: 0.5
   - name: top_p
     use_template: top_p
+    min: 0
+    max: 1
+    default: 1
   - name: max_tokens
     use_template: max_tokens
-    default: 512
     min: 1
     max: 1024
+    default: 1024
   - name: frequency_penalty
     use_template: frequency_penalty
     min: -2

api/core/model_runtime/model_providers/nvidia/llm/llama2-70b.yaml

@@ -7,17 +7,23 @@ features:
   - agent-thought
 model_properties:
   mode: chat
-  context_size: 32768
+  context_size: 4096
 parameter_rules:
   - name: temperature
     use_template: temperature
+    min: 0
+    max: 1
+    default: 0.5
   - name: top_p
     use_template: top_p
+    min: 0
+    max: 1
+    default: 1
   - name: max_tokens
     use_template: max_tokens
-    default: 512
     min: 1
     max: 1024
+    default: 1024
   - name: frequency_penalty
     use_template: frequency_penalty
     min: -2

api/core/model_runtime/model_providers/nvidia/llm/llama3-70b.yaml

@@ -0,0 +1,36 @@
+model: meta/llama3-70b
+label:
+  zh_Hans: meta/llama3-70b
+  en_US: meta/llama3-70b
+model_type: llm
+features:
+  - agent-thought
+model_properties:
+  mode: chat
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    min: 0
+    max: 1
+    default: 0.5
+  - name: top_p
+    use_template: top_p
+    min: 0
+    max: 1
+    default: 1
+  - name: max_tokens
+    use_template: max_tokens
+    min: 1
+    max: 1024
+    default: 1024
+  - name: frequency_penalty
+    use_template: frequency_penalty
+    min: -2
+    max: 2
+    default: 0
+  - name: presence_penalty
+    use_template: presence_penalty
+    min: -2
+    max: 2
+    default: 0

api/core/model_runtime/model_providers/nvidia/llm/llama3-8b.yaml

@@ -0,0 +1,36 @@
+model: meta/llama3-8b
+label:
+  zh_Hans: meta/llama3-8b
+  en_US: meta/llama3-8b
+model_type: llm
+features:
+  - agent-thought
+model_properties:
+  mode: chat
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    min: 0
+    max: 1
+    default: 0.5
+  - name: top_p
+    use_template: top_p
+    min: 0
+    max: 1
+    default: 1
+  - name: max_tokens
+    use_template: max_tokens
+    min: 1
+    max: 1024
+    default: 1024
+  - name: frequency_penalty
+    use_template: frequency_penalty
+    min: -2
+    max: 2
+    default: 0
+  - name: presence_penalty
+    use_template: presence_penalty
+    min: -2
+    max: 2
+    default: 0

api/core/model_runtime/model_providers/nvidia/llm/llm.py

@@ -25,7 +25,10 @@ class NVIDIALargeLanguageModel(OAIAPICompatLargeLanguageModel):
         'mistralai/mixtral-8x7b-instruct-v0.1': '',
         'google/gemma-7b': '',
         'google/codegemma-7b': '',
-        'meta/llama2-70b': ''
+        'meta/llama2-70b': '',
+        'meta/llama3-8b': '',
+        'meta/llama3-70b': ''
+        
     }
 
     def _invoke(self, model: str, credentials: dict,

api/core/model_runtime/model_providers/nvidia/llm/mistralai_mixtral-8x7b-instruct-v0.1.yaml

@@ -11,13 +11,19 @@ model_properties:
 parameter_rules:
   - name: temperature
     use_template: temperature
+    min: 0
+    max: 1
+    default: 0.5
   - name: top_p
     use_template: top_p
+    min: 0
+    max: 1
+    default: 1
   - name: max_tokens
     use_template: max_tokens
-    default: 512
     min: 1
     max: 1024
+    default: 1024
   - name: frequency_penalty
     use_template: frequency_penalty
     min: -2

api/core/model_runtime/model_providers/nvidia/nvidia.yaml

@@ -1,6 +1,9 @@
 provider: nvidia
 label:
   en_US: API Catalog
+description:
+  en_US: API Catalog
+  zh_Hans: API Catalog
 icon_small:
   en_US: icon_s_en.svg
 icon_large: