updates 9.0

app.py

@@ -69,7 +69,8 @@ class GPTOSSSizeRecommender:
 			self.tokenizer = AutoTokenizer.from_pretrained(
 				self.model_name,
 				trust_remote_code=True,
-				use_fast=False
+				use_fast=False,
+				cache_dir="/tmp/huggingface_cache"
 			)
 			
 			# Add padding token if not present
@@ -93,7 +94,8 @@ class GPTOSSSizeRecommender:
 					# Load model with 8-bit
 					self.model = AutoModelForCausalLM.from_pretrained(
 						self.model_name,
-						**model_kwargs
+						**model_kwargs,
+						cache_dir="/tmp/huggingface_cache"
 					)
 					print("✅ Model loaded with 8-bit quantization")
 				except Exception as e:
@@ -107,7 +109,8 @@ class GPTOSSSizeRecommender:
 					}
 					self.model = AutoModelForCausalLM.from_pretrained(
 						self.model_name,
-						**model_kwargs
+						**model_kwargs,
+						cache_dir="/tmp/huggingface_cache"
 					)
 					print("✅ Model loaded with bfloat16")
 			else:
@@ -118,7 +121,8 @@ class GPTOSSSizeRecommender:
 				# Load model
 				self.model = AutoModelForCausalLM.from_pretrained(
 					self.model_name,
-					**model_kwargs
+					**model_kwargs,
+					cache_dir="/tmp/huggingface_cache"
 				)
 			
 			if self.device == "cpu":
@@ -242,83 +246,66 @@ class GPTOSSSizeRecommender:
 		gender = user_data.get('gender', 'unisex')
 		estimated_measurements = self.estimate_missing_measurements(user_data, gender)
 		
-		prompt = f"""You are an expert AI clothing size recommendation system with advanced reasoning capabilities. Your task is to recommend the perfect garment size based on available body measurements and garment specifications.
-
-**CUSTOMER PROFILE:**
-Gender/Body Type: {user_data.get('body_type', 'Not specified')} / {gender}
-Height: {estimated_measurements.get('height', 'Not provided')} inches
-Weight: {estimated_measurements.get('weight', 'Not provided')} lbs
-
-**BODY MEASUREMENTS (in inches):**
-- Chest/Bust: {estimated_measurements.get('chest', 0):.1f} {'(estimated)' if user_data.get('chest', 0) == 0 and estimated_measurements.get('chest', 0) > 0 else ''}
-- Shoulder Width: {estimated_measurements.get('shoulder_width', 0):.1f} {'(estimated)' if user_data.get('shoulder_width', 0) == 0 and estimated_measurements.get('shoulder_width', 0) > 0 else ''}
-- Sleeve Length: {estimated_measurements.get('sleeve_length', 0):.1f} {'(estimated)' if user_data.get('sleeve_length', 0) == 0 and estimated_measurements.get('sleeve_length', 0) > 0 else ''}
-- Neck Circumference: {estimated_measurements.get('neck_circumference', 0):.1f} {'(estimated)' if user_data.get('neck_circumference', 0) == 0 and estimated_measurements.get('neck_circumference', 0) > 0 else ''}
-- Shirt Length: {estimated_measurements.get('shirt_length', 0):.1f} {'(estimated)' if user_data.get('shirt_length', 0) == 0 and estimated_measurements.get('shirt_length', 0) > 0 else ''}
-- Armhole Size: {estimated_measurements.get('armhole_size', 0):.1f} {'(estimated)' if user_data.get('armhole_size', 0) == 0 and estimated_measurements.get('armhole_size', 0) > 0 else ''}
-- Waist: {estimated_measurements.get('waist', 0):.1f} {'(estimated)' if user_data.get('waist', 0) == 0 and estimated_measurements.get('waist', 0) > 0 else ''}
-- Inseam: {estimated_measurements.get('inseam', 0):.1f} {'(estimated)' if user_data.get('inseam', 0) == 0 and estimated_measurements.get('inseam', 0) > 0 else ''}
-- Thigh Circumference: {estimated_measurements.get('thigh_circumference', 0):.1f} {'(estimated)' if user_data.get('thigh_circumference', 0) == 0 and estimated_measurements.get('thigh_circumference', 0) > 0 else ''}
-
-**GARMENT DETAILS:**
-- Product: {garment_data.get('product_name', 'Not specified')}
-- Brand: {garment_data.get('brand', 'Not specified')}
-- Category: {garment_data.get('category', 'Not specified')}
-- Available Sizes: {', '.join(garment_data.get('available_sizes', []))}
-
-**GARMENT SIZE SPECIFICATIONS (in inches):**"""
-		
-		# Add size measurements for each available size
+		# Build size comparison table
+		size_comparisons = []
 		for size in garment_data.get('available_sizes', []):
 			measurements = garment_data.get(f'{size.lower()}_measurements', {})
 			if measurements:
-				prompt += f"\n\n{size} Size:"
-				for measurement, value in measurements.items():
-					if value and value != 0:
-						prompt += f"\n  - {measurement.title()}: {value} inches"
+				size_comparisons.append(f"{size}: Chest={measurements.get('chest', 'N/A')}, Waist={measurements.get('waist', 'N/A')}, Length={measurements.get('length', 'N/A')}")
+		
+		prompt = f"""<system>You are a professional clothing size recommendation assistant. You must provide clear, structured recommendations following the exact format specified.</system>
 
-		prompt += """
+<task>Analyze the customer's measurements against the garment specifications and recommend the best size.</task>
 
-**ANALYSIS REQUIREMENTS:**
-1. **Smart Comparison**: Compare customer measurements (including estimated ones) with each available garment size
-2. **Estimation Confidence**: Consider the reliability of estimated vs. provided measurements
-3. **Fit Philosophy**: Account for garment category, brand typical sizing, and comfort requirements
-4. **Multi-factor Analysis**: Consider all relevant measurements, not just chest/waist
-5. **Gender/Body Type**: Factor in the customer's body type and proportions
-6. **Tolerance Analysis**: Account for fabric stretch, fit preferences, and wearing comfort
+<customer_info>
+- Gender: {gender}
+- Body Type: {user_data.get('body_type', 'Regular')}
+- Height: {estimated_measurements.get('height', 0):.0f} inches
+- Weight: {estimated_measurements.get('weight', 0):.0f} lbs
+</customer_info>
 
-**RESPONSE FORMAT:**
-🎯 **RECOMMENDED SIZE: [SIZE]**
-📊 **CONFIDENCE SCORE: [SCORE]/100**
-🔄 **ALTERNATIVE SIZE: [ALT_SIZE]** (if confidence < 85)
+<customer_measurements>
+- Chest: {estimated_measurements.get('chest', 0):.0f}" (User Provided: {'Yes' if user_data.get('chest', 0) > 0 else 'No'})
+- Waist: {estimated_measurements.get('waist', 0):.0f}" (User Provided: {'Yes' if user_data.get('waist', 0) > 0 else 'No'})
+- Shoulder: {estimated_measurements.get('shoulder_width', 0):.0f}" (User Provided: {'Yes' if user_data.get('shoulder_width', 0) > 0 else 'No'})
+- Length Preference: {estimated_measurements.get('shirt_length', 0):.0f}" (User Provided: {'Yes' if user_data.get('shirt_length', 0) > 0 else 'No'})
+</customer_measurements>
 
-📋 **DETAILED FIT ANALYSIS:**
+<garment_info>
+- Product: {garment_data.get('product_name', 'Not specified')}
+- Brand: {garment_data.get('brand', 'Not specified')}
+- Category: {garment_data.get('category', 'Not specified')}
+</garment_info>
 
-**Critical Measurements Match:**
-- Chest/Bust: [Analysis with specific numbers]
-- Waist: [Analysis with specific numbers]
-- Shoulder: [Analysis with specific numbers]
-- Length: [Analysis with specific numbers]
+<available_sizes>
+{chr(10).join(size_comparisons)}
+</available_sizes>
 
-**Estimated Measurements Impact:**
-[Discuss how estimated measurements affect confidence]
+<instructions>
+1. Compare customer measurements to each size
+2. Find the best match considering chest as priority, then waist, then length
+3. Calculate fit tolerance (2-3 inches for comfort)
+4. Consider if estimated measurements affect confidence
+5. Provide your response in EXACTLY this format:
 
-**Fit Assessment:**
-- Overall Comfort: [Excellent/Good/Acceptable/Tight/Loose]
-- Movement Freedom: [Assessment]
-- Style Appropriateness: [Assessment]
+RECOMMENDED SIZE: [Choose one size]
+CONFIDENCE: [Number 1-100]%
+ALTERNATIVE: [Second best size or "None"]
 
-⚠️ **POTENTIAL CONCERNS:**
-[List any fit issues or areas to watch]
+FIT ANALYSIS:
+✓ Chest fit: [Customer chest] vs Size [X] chest [Y] = [Perfect/Good/Tight/Loose]
+✓ Waist fit: [Customer waist] vs Size [X] waist [Y] = [Perfect/Good/Tight/Loose]
+✓ Length fit: [Assessment based on height and garment length]
 
-✅ **FINAL RECOMMENDATION:**
-[Comprehensive summary with actionable advice]
+KEY POINTS:
+• [Most important fit consideration]
+• [Second important point]
+• [Any concerns or adjustments needed]
 
-**Measurement Notes:**
-[Any important notes about sizing for this specific garment/brand]
+RECOMMENDATION: [One sentence summary, e.g. "Size M will provide the best overall fit with comfortable room in the chest and proper length."]
+</instructions>
 
-Reasoning: high
-"""
+Provide your analysis now:"""
 		
 		return prompt
 
@@ -346,13 +333,15 @@ Reasoning: high
 			with torch.no_grad():
 				outputs = self.model.generate(
 					**inputs,
-					max_new_tokens=600,
-					temperature=0.7,
+					max_new_tokens=400,
+					temperature=0.3,
 					do_sample=True,
-					top_p=0.9,
+					top_p=0.95,
+					top_k=50,
 					pad_token_id=self.tokenizer.eos_token_id,
 					eos_token_id=self.tokenizer.eos_token_id,
-					repetition_penalty=1.1
+					repetition_penalty=1.05,
+					num_beams=1
 				)
 			
 			# Decode response
@@ -364,10 +353,64 @@ Reasoning: high
 			else:
 				response = generated_text.strip()
 			
+			# Format response
+			response = self.format_response(response)
+			
 			return response if response else "No response generated"
 			
 		except Exception as e:
 			return f"❌ Model Error: {str(e)}\nPlease try again or restart the application."
+	
+	def format_response(self, response: str) -> str:
+		"""Format and clean up the model response"""
+		# Clean up common issues
+		response = response.strip()
+		
+		# If response doesn't follow format, try to extract key information
+		if "RECOMMENDED SIZE:" not in response:
+			# Try to extract size recommendation from unstructured text
+			import re
+			
+			# Look for size mentions
+			size_pattern = r'\b(XS|S|M|L|XL|XXL|XXXL)\b'
+			sizes_found = re.findall(size_pattern, response.upper())
+			
+			if sizes_found:
+				# Create structured response
+				formatted = f"""🎯 **RECOMMENDED SIZE: {sizes_found[0]}**
+📊 **CONFIDENCE: 75%**
+
+📋 **ANALYSIS:**
+{response[:500]}...
+
+✅ **RECOMMENDATION:**
+Based on the analysis, size {sizes_found[0]} is recommended."""
+				return formatted
+		
+		# Clean up formatting
+		lines = response.split('\n')
+		cleaned_lines = []
+		
+		for line in lines:
+			line = line.strip()
+			if line:
+				# Format headers
+				if line.startswith("RECOMMENDED SIZE:"):
+					line = f"🎯 **{line}**"
+				elif line.startswith("CONFIDENCE:"):
+					line = f"📊 **{line}**"
+				elif line.startswith("ALTERNATIVE:"):
+					line = f"🔄 **{line}**"
+				elif line.startswith("FIT ANALYSIS:"):
+					line = f"\n📋 **{line}**"
+				elif line.startswith("KEY POINTS:"):
+					line = f"\n🔑 **{line}**"
+				elif line.startswith("RECOMMENDATION:"):
+					line = f"\n✅ **{line}**"
+				
+				cleaned_lines.append(line)
+		
+		return '\n'.join(cleaned_lines)
 
 	def recommend_size(self, user_data: Dict, garment_data: Dict, api_key: str = None) -> str:
 		"""Main function to get size recommendation"""
@@ -403,6 +446,14 @@ def predict_size(
 ) -> str:
 	"""Enhanced prediction function with robust input handling"""
 	
+	# Check if model is loaded
+	if not recommender.model_loaded:
+		return """⏳ **AI Model is Loading...**
+		
+Please wait a moment while the AI model initializes. This may take 1-2 minutes on first use.
+
+Once loaded, click 'Get My Perfect Size' again!"""
+	
 	try:
 		# Parse all measurements with units
 		user_data = {
@@ -449,10 +500,29 @@ def predict_size(
 		if total_measurements < 2:
 			return "⚠️ Please provide at least 2 body measurements for accurate recommendations."
 		
+		# Show processing message
+		processing_msg = """🔄 **Analyzing your measurements...**
+		
+Our AI is comparing your measurements with the garment specifications to find your perfect fit.
+		
+This usually takes 5-10 seconds."""
+		
 		# Get recommendation
 		recommendation = recommender.recommend_size(user_data, garment_data)
 		
-		return recommendation
+		# Add visual separator
+		final_output = f"""
+{recommendation}
+
+---
+
+💡 **Shopping Tips:**
+- If between sizes, consider the fabric (stretchy vs rigid)
+- Check the brand's return policy before ordering
+- Save your measurements for future purchases
+"""
+		
+		return final_output
 		
 	except Exception as e:
 		return f"❌ Processing Error: {str(e)}\nPlease check your inputs and try again."