Upload 4 files

.gitignore

@@ -0,0 +1,14 @@
+
+data/
+
+.vscode/
+
+.ipynb_checkpoints/
+
+src/__pycache__/
+
+flagged/
+
+__pycache__/
+
+.idea/

README.md

@@ -1,13 +1,20 @@
----
-title: Hindi Character Recognition
-emoji: 💻
-colorFrom: yellow
-colorTo: purple
-sdk: gradio
-sdk_version: 3.19.1
-app_file: app.py
-pinned: false
-license: apache-2.0
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Hindi-Character-Recognition
+
+Install the requirements: `pip install -r requirements.txt`
+
+1. Hindi Character Recognition
+
+    ![App](https://user-images.githubusercontent.com/54038552/210263132-2e95ad65-5049-4a8c-b453-2616cbc4df20.gif)
+
+    Getting the data:
+    - Download the data from [here](https://www.kaggle.com/datasets/suvooo/hindi-character-recognition)
+    - Unzip it. You need to split the data into 4 different directories, since we are training for Hindi digits & letters separately.
+    ![image](https://user-images.githubusercontent.com/54038552/209815855-cd629bdd-5a9a-474e-8ad6-1d4df1954fdc.png)
+    
+    How to run ?
+    - You can create your custom model in the `model.py` file or can go with the `HNet` already present. For custom models created, you need to import them to `train.py`, for them to to use. Remember we are training different models for Hindi Digit & Characters.
+    - Now to train the model with default params do, `python train.py`. You can also specify epochs and lr. Most important, is the `model_type`
+    - To train do, `python train.py --epochs <num-epochs> --lr <learning-rate> --model_type <type-of-model>`
+    
+    Running the app:
+    - Just so, `streamlit run app.py`

gradio_app.py

@@ -0,0 +1,48 @@
+import gradio as gr
+import torch
+import json
+import src.config as CFG
+from src.model import HNet
+
+
+def classify(model, image, mapping):
+    image = torch.from_numpy(image).float()
+    image = image.permute(2, 0, 1).unsqueeze(0)
+
+    outputs = model(image)
+
+    _, preds = torch.max(outputs, 1)
+
+    return f"The predicted character is: {mapping[str(preds[0].item())]}"
+
+
+def upload_and_clasify(image, option):
+
+    mapping, model = None, None
+
+    if option == "Digit":
+        if CFG.BEST_MODEL_DIGIT.exists():
+            model = HNet(num_classes=10)
+            model.load_state_dict(
+                torch.load(CFG.BEST_MODEL_DIGIT, map_location=CFG.DEVICE)
+            )
+            with open(CFG.INDEX_DIGIT, "r") as f:
+                mapping = json.load(f)
+            return classify(model, image, mapping)
+    else:
+        if CFG.BEST_MODEL_VYANJAN.exists():
+            model = HNet(num_classes=36)
+            model.load_state_dict(
+                torch.load(CFG.BEST_MODEL_VYANJAN, map_location=CFG.DEVICE)
+            )
+            with open(CFG.INDEX_VYNAJAN, "r") as f:
+                mapping = json.load(f)
+            return classify(model, image, mapping)
+
+
+demo = gr.Interface(
+    fn=upload_and_clasify,
+    inputs=["image", gr.Dropdown(["Digit", "Vyanjan"])],
+    outputs="text",
+)
+demo.launch(share=True, server_port=8080)

requirements.txt

@@ -0,0 +1,8 @@
+gradio==3.19.1
+numpy==1.21.2
+Pillow==9.4.0
+prettytable==3.6.0
+streamlit==1.4.0
+torch==1.11.0
+torchvision==0.12.0
+tqdm==4.62.3