model load from hf model repo

.gitattributes

@@ -1,2 +0,0 @@
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -1,17 +1,4 @@
-# Exclude large training files that aren't needed for inference
-model_chekpoint/optimizer.pt
-model_chekpoint/rng_state.pth
-model_chekpoint/scheduler.pt
-model_chekpoint/trainer_state.json
-model_chekpoint/training_args.bin
-model_chekpoint/model.safetensors
-
-# Exclude large files in root directory
-model.safetensors
-optimizer.pt
-rng_state.pth
-scheduler.pt
-trainer_state.json
-training_args.bin
-tokenizer.json
-vocab.txt
+.venv/
+__pycache__/
+.idea
+model_checkpoint/

app.py

@@ -1,11 +1,10 @@
 import gradio as gr
-import torch
-from utils import load_model_from_checkpoint, get_device
+from utils import load_pipeline_from_huggingface
 
 
 def predict_sentiment(text):
     """
-    Predict sentiment of the input text using the loaded model.
+    Predict sentiment of the input text using the loaded pipeline.
     
     Args:
         text (str): Input text to analyze
@@ -14,25 +13,16 @@ def predict_sentiment(text):
         str: Sentiment prediction
     """
     try:
-        # Load model and tokenizer
-        model, tokenizer = load_model_from_checkpoint()
-        device = get_device()
-        model = model.to(device)
+        # Load pipeline
+        sentiment_pipeline = load_pipeline_from_huggingface()
         
-        # Tokenize input text
-        inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=512)
-        inputs = {k: v.to(device) for k, v in inputs.items()}
+        # Get prediction using pipeline
+        results = sentiment_pipeline(text)
         
-        # Get prediction
-        with torch.no_grad():
-            outputs = model(**inputs)
-            predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)
-            predicted_class = torch.argmax(predictions, dim=-1).item()
-            
-        # Map class to sentiment (adjust based on your model's classes)
-        sentiment_map = {0: "Negative", 1: "Positive", 2: "Neutral"}
-        sentiment = sentiment_map.get(predicted_class, "Unknown")
-        confidence = predictions[0][predicted_class].item()
+        # Extract the highest confidence prediction
+        best_result = max(results[0], key=lambda x: x['score'])
+        sentiment = best_result['label']
+        confidence = best_result['score']
         
         return f"Sentiment: {sentiment} (Confidence: {confidence:.2f})"
         

config.json

@@ -1,37 +0,0 @@
-{
-  "_name_or_path": "ProsusAI/finbert",
-  "architectures": [
-    "BertForSequenceClassification"
-  ],
-  "attention_probs_dropout_prob": 0.1,
-  "classifier_dropout": null,
-  "gradient_checkpointing": false,
-  "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
-  "id2label": {
-    "0": "positive",
-    "1": "negative",
-    "2": "neutral"
-  },
-  "initializer_range": 0.02,
-  "intermediate_size": 3072,
-  "label2id": {
-    "negative": 1,
-    "neutral": 2,
-    "positive": 0
-  },
-  "layer_norm_eps": 1e-12,
-  "max_position_embeddings": 512,
-  "model_type": "bert",
-  "num_attention_heads": 12,
-  "num_hidden_layers": 12,
-  "pad_token_id": 0,
-  "position_embedding_type": "absolute",
-  "problem_type": "single_label_classification",
-  "torch_dtype": "float32",
-  "transformers_version": "4.46.3",
-  "type_vocab_size": 2,
-  "use_cache": true,
-  "vocab_size": 30522
-}

model_chekpoint/config.json

@@ -1,37 +0,0 @@
-{
-  "_name_or_path": "ProsusAI/finbert",
-  "architectures": [
-    "BertForSequenceClassification"
-  ],
-  "attention_probs_dropout_prob": 0.1,
-  "classifier_dropout": null,
-  "gradient_checkpointing": false,
-  "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
-  "id2label": {
-    "0": "positive",
-    "1": "negative",
-    "2": "neutral"
-  },
-  "initializer_range": 0.02,
-  "intermediate_size": 3072,
-  "label2id": {
-    "negative": 1,
-    "neutral": 2,
-    "positive": 0
-  },
-  "layer_norm_eps": 1e-12,
-  "max_position_embeddings": 512,
-  "model_type": "bert",
-  "num_attention_heads": 12,
-  "num_hidden_layers": 12,
-  "pad_token_id": 0,
-  "position_embedding_type": "absolute",
-  "problem_type": "single_label_classification",
-  "torch_dtype": "float32",
-  "transformers_version": "4.46.3",
-  "type_vocab_size": 2,
-  "use_cache": true,
-  "vocab_size": 30522
-}

model_chekpoint/special_tokens_map.json

@@ -1,7 +0,0 @@
-{
-  "cls_token": "[CLS]",
-  "mask_token": "[MASK]",
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "unk_token": "[UNK]"
-}

model_chekpoint/tokenizer_config.json

@@ -1,57 +0,0 @@
-{
-  "added_tokens_decoder": {
-    "0": {
-      "content": "[PAD]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "100": {
-      "content": "[UNK]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "101": {
-      "content": "[CLS]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "102": {
-      "content": "[SEP]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "103": {
-      "content": "[MASK]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    }
-  },
-  "clean_up_tokenization_spaces": true,
-  "cls_token": "[CLS]",
-  "do_basic_tokenize": true,
-  "do_lower_case": true,
-  "mask_token": "[MASK]",
-  "model_max_length": 512,
-  "never_split": null,
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "strip_accents": null,
-  "tokenize_chinese_chars": true,
-  "tokenizer_class": "BertTokenizer",
-  "unk_token": "[UNK]"
-}

requirements.txt

@@ -0,0 +1,5 @@
+gradio 
+torch
+transformers
+huggingface-hub
+pyyaml

special_tokens_map.json

@@ -1,7 +0,0 @@
-{
-  "cls_token": "[CLS]",
-  "mask_token": "[MASK]",
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "unk_token": "[UNK]"
-}

tokenizer_config.json

@@ -1,57 +0,0 @@
-{
-  "added_tokens_decoder": {
-    "0": {
-      "content": "[PAD]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "100": {
-      "content": "[UNK]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "101": {
-      "content": "[CLS]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "102": {
-      "content": "[SEP]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "103": {
-      "content": "[MASK]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    }
-  },
-  "clean_up_tokenization_spaces": true,
-  "cls_token": "[CLS]",
-  "do_basic_tokenize": true,
-  "do_lower_case": true,
-  "mask_token": "[MASK]",
-  "model_max_length": 512,
-  "never_split": null,
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "strip_accents": null,
-  "tokenize_chinese_chars": true,
-  "tokenizer_class": "BertTokenizer",
-  "unk_token": "[UNK]"
-}

upload_model.py

@@ -0,0 +1,7 @@
+from huggingface_hub import upload_folder
+
+upload_folder(
+    repo_id="utyug1/finbert-finetuned-model",
+    folder_path="model_checkpoint",
+    repo_type="model",
+)

utils.py

@@ -1,16 +1,15 @@
 import os
 import yaml
-from transformers import AutoModelForSequenceClassification, AutoTokenizer
-import torch
+from transformers import pipeline
 
 
-def load_model_from_checkpoint():
+def load_pipeline_from_huggingface():
     """
-    Load model from transformers checkpoint.
-    The checkpoint location is specified by model_path in config.yaml.
+    Load sentiment analysis pipeline from Hugging Face repository.
+    The repository name is specified by model_path in config.yaml.
     
     Returns:
-        tuple: (model, tokenizer) loaded from the checkpoint
+        pipeline: Sentiment analysis pipeline loaded from Hugging Face
     """
     # Read config to get model_path
     config_path = "config.yaml"
@@ -24,32 +23,15 @@ def load_model_from_checkpoint():
     if not model_path:
         raise ValueError("model_path not found in config.yaml")
     
-    # Check if checkpoint directory exists
-    if not os.path.exists(model_path):
-        raise FileNotFoundError(f"Model checkpoint not found at: {model_path}")
-    
-    # Load tokenizer and model from the checkpoint
+    # Load pipeline from Hugging Face repository
     try:
-        tokenizer = AutoTokenizer.from_pretrained(model_path)
-        model = AutoModelForSequenceClassification.from_pretrained(model_path)
-        
-        # Set model to evaluation mode
-        model.eval()
+        sentiment_pipeline = pipeline(
+            "sentiment-analysis",
+            model=model_path,
+            return_all_scores=True
+        )
         
-        return model, tokenizer
+        return sentiment_pipeline
         
     except Exception as e:
-        raise RuntimeError(f"Failed to load model from checkpoint {model_path}: {str(e)}")
-
-
-def get_device():
-    """
-    Get the appropriate device for model inference.
-    
-    Returns:
-        torch.device: Device to use for inference
-    """
-    if torch.cuda.is_available():
-        return torch.device('cuda')
-    else:
-        return torch.device('cpu')
+        raise RuntimeError(f"Failed to load pipeline from Hugging Face repository {model_path}: {str(e)}")