Update app.py

app.py

@@ -14,8 +14,8 @@ clients = {
     "meta-llama/Llama-3.2-3B-Instruct": InferenceClient(
         model="meta-llama/Llama-3.2-3B-Instruct", token=hf_token
     ),
-    "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B": InferenceClient(
-        model="deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B", token=hf_token
+    "mistralai/Mistral-7B-Instruct-v0.3": InferenceClient(
+        model="mistralai/Mistral-7B-Instruct-v0.3", token=hf_token
     ),
     "meta-llama/Llama-2-7b-chat-hf": InferenceClient(
         model="meta-llama/Llama-2-7b-chat-hf", token=hf_token
@@ -27,35 +27,41 @@ clients = {
 
 # Model presets
 presets = {
-    "mistralai/Mistral-7B-Instruct-v0.3": {
-        "Fast": {"max_new_tokens": 256, "temperature": 1.0, "top_p": 0.8},
-        "Normal": {"max_new_tokens": 512, "temperature": 0.6, "top_p": 0.75},
-        "Quality": {"max_new_tokens": 1024, "temperature": 0.45, "top_p": 0.60},
-        "Unreal Performance": {"max_new_tokens": 2048, "temperature": 0.5, "top_p": 0.7},
-    },
     "meta-llama/Llama-3.2-3B-Instruct": {
-        "Fast": {"max_new_tokens": 256, "temperature": 1.0, "top_p": 0.9},
-        "Normal": {"max_new_tokens": 512, "temperature": 0.7, "top_p": 0.95},
-        "Quality": {"max_new_tokens": 1024, "temperature": 0.5, "top_p": 0.90},
-        "Unreal Performance": {"max_new_tokens": 2048, "temperature": 0.6, "top_p": 0.75},
+        "Fast": {"max_new_tokens": 1024, "temperature": 1.0, "top_p": 0.9},
+        "Normal": {"max_new_tokens": 2048, "temperature": 0.7, "top_p": 0.95},
+        "Quality": {"max_new_tokens": 4096, "temperature": 0.5, "top_p": 0.90},
+        "Unreal Performance": {"max_new_tokens": 8192, "temperature": 0.6, "top_p": 0.75},
     },
-    "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B": {
-        "Fast": {"max_new_tokens": 256, "temperature": 1.0, "top_p": 0.9},
-        "Normal": {"max_new_tokens": 512, "temperature": 0.7, "top_p": 0.95},
-        "Quality": {"max_new_tokens": 1024, "temperature": 0.5, "top_p": 0.90},
-        "Unreal Performance": {"max_new_tokens": 2048, "temperature": 0.6, "top_p": 0.75},
+    "mistralai/Mistral-7B-Instruct-v0.3": {
+        "Fast": {"max_new_tokens": 4096, "temperature": 1.0, "top_p": 0.9},
+        "Normal": {"max_new_tokens": 8192, "temperature": 0.7, "top_p": 0.95},
+        "Quality": {"max_new_tokens": 16384, "temperature": 0.5, "top_p": 0.90},
+        "Unreal Performance": {"max_new_tokens": 32768, "temperature": 0.6, "top_p": 0.75},
     },
     "meta-llama/Llama-2-7b-chat-hf": {
-        "Fast": {"max_new_tokens": 256, "temperature": 1.0, "top_p": 0.9},
-        "Normal": {"max_new_tokens": 512, "temperature": 0.7, "top_p": 0.95},
-        "Quality": {"max_new_tokens": 1024, "temperature": 0.5, "top_p": 0.90},
-        "Unreal Performance": {"max_new_tokens": 2048, "temperature": 0.6, "top_p": 0.75},
+        "Fast": {"max_new_tokens": 512, "temperature": 1.0, "top_p": 0.9},
+        "Normal": {"max_new_tokens": 1024, "temperature": 0.7, "top_p": 0.95},
+        "Quality": {"max_new_tokens": 2048, "temperature": 0.5, "top_p": 0.90},
+        "Unreal Performance": {"max_new_tokens": 4096, "temperature": 0.6, "top_p": 0.75},
     },
     "meta-llama/Llama-3.3-70B-Instruct": {
-        "Fast": {"max_new_tokens": 256, "temperature": 1.0, "top_p": 0.9},
-        "Normal": {"max_new_tokens": 512, "temperature": 0.7, "top_p": 0.95},
-        "Quality": {"max_new_tokens": 1024, "temperature": 0.5, "top_p": 0.90},
-        "Unreal Performance": {"max_new_tokens": 2048, "temperature": 0.6, "top_p": 0.75},
+        "Fast": {"max_new_tokens": 2048, "temperature": 1.0, "top_p": 0.9},
+        "Normal": {"max_new_tokens": 4096, "temperature": 0.7, "top_p": 0.95},
+        "Quality": {"max_new_tokens": 8192, "temperature": 0.5, "top_p": 0.90},
+        "Unreal Performance": {"max_new_tokens": 16384, "temperature": 0.6, "top_p": 0.75},
+    },
+    "meta-llama/Llama-2-13b-chat-hf": {
+        "Fast": {"max_new_tokens": 2048, "temperature": 1.0, "top_p": 0.9},
+        "Normal": {"max_new_tokens": 4096, "temperature": 0.7, "top_p": 0.95},
+        "Quality": {"max_new_tokens": 8192, "temperature": 0.5, "top_p": 0.90},
+        "Unreal Performance": {"max_new_tokens": 16384, "temperature": 0.6, "top_p": 0.75},
+    },
+    "deepseek-ai/deepseek-llm-67b-chat": {
+        "Fast": {"max_new_tokens": 4096, "temperature": 1.0, "top_p": 0.9},
+        "Normal": {"max_new_tokens": 8192, "temperature": 0.7, "top_p": 0.95},
+        "Quality": {"max_new_tokens": 16384, "temperature": 0.5, "top_p": 0.90},
+        "Unreal Performance": {"max_new_tokens": 32768, "temperature": 0.6, "top_p": 0.75},
     }
 }
 
@@ -204,7 +210,6 @@ system_messages = {
 
 # Model registry
 model_registry = [
-    ("deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B", "Lake 1 Flash", "released"),
     ("mistralai/Mistral-7B-Instruct-v0.3", "Lake 1 Base", "released"),
     ("meta-llama/Llama-3.2-3B-Instruct", "Lake 1 Advanced", "released"),
     ("meta-llama/Llama-2-7b-chat-hf", "Lake 2 Chat [Closed Alpha]", "shown"),
@@ -215,15 +220,6 @@ model_registry = [
 
 # Model information
 model_info = {
-    "Lake 1 Flash": {
-        "description": "English only, fast and lightweight model optimized for quick responses",
-        "parameters": "1.5B",
-        "training_data": "Diverse internet text + curated datasets",
-        "developer": "BI Corp",
-        "best_for": "Quick queries, mobile applications",
-        "architecture": "Transformer-based",
-        "context_window": "4096 tokens"
-    },
     "Lake 1 Base": {
         "description": "Balanced model offering good performance across tasks",
         "parameters": "7B",
@@ -259,6 +255,24 @@ model_info = {
         "best_for": "Research & advanced applications",
         "architecture": "Mixture of Experts",
         "context_window": "16384 tokens"
+    },
+    "Lake 2 Advanced": {
+        "description": "Enhanced performance model with advanced reasoning capabilities.",
+        "parameters": "13B",
+        "training_data": "Diverse datasets including tech, literature, and general knowledge.",
+        "developer": "BI Corp",
+        "best_for": "Handling large-scale queries and in-depth topics.",
+        "architecture": "Transformer-based with specialized learning capabilities.",
+        "context_window": "16384 tokens"
+    },
+    "Lake 2 Pro [Planned]": {
+        "description": "Pro-level model with deep learning architecture for high-end applications.",
+        "parameters": "67B",
+        "training_data": "Expert-level data across multiple industries.",
+        "developer": "BI Corp",
+        "best_for": "High-performance computing and enterprise-level tasks.",
+        "architecture": "Hybrid architecture leveraging the latest advances in deep learning.",
+        "context_window": "32768 tokens"
     }
 }