Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,17 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Muril-Model/Muril-Model/muril_multilingual_dataset.csv filter=lfs diff=lfs merge=lfs -text
+Muril-Model/muril_multilingual_dataset.csv filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_1.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_10.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_11.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_12.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_2.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_3.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_4.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_5.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_6.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_7.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_8.pdf filter=lfs diff=lfs merge=lfs -text
+pdfs/constitution_9.pdf filter=lfs diff=lfs merge=lfs -text

Muril-Model/.gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+muril_multilingual_dataset.csv filter=lfs diff=lfs merge=lfs -text

Muril-Model/.gitignore

@@ -0,0 +1,45 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# Virtual Environment
+venv/
+env/
+.env/
+.venv/
+
+# IDEs and Editors
+.idea/
+.vscode/
+*.swp
+*.swo
+*~
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# Model files and data
+*.h5
+*.pkl
+*.model
+data/
+models/
+logs/

Muril-Model/Dockerfile

@@ -0,0 +1,34 @@
+# Use lightweight Python base image
+FROM python:3.10-slim
+
+# Disable CUDA & set cache locations
+ENV TORCH_DISABLE_CUDA=1
+ENV TRANSFORMERS_CACHE=/app/hf_cache
+ENV HF_HOME=/app/hf_cache
+
+# Working directory
+WORKDIR /app
+
+# Copy dependency list
+COPY requirements.txt .
+
+# Install dependencies efficiently
+RUN apt-get update && apt-get install -y git && \
+    pip install --no-cache-dir -r requirements.txt && \
+    rm -rf /var/lib/apt/lists/*
+
+# Copy app code
+COPY . .
+
+# Create cache folder (writable)
+RUN mkdir -p /app/hf_cache && chmod -R 777 /app/hf_cache
+
+# Expose port for FastAPI
+EXPOSE 8080
+
+# Health check (optional for Cloud Run / Spaces)
+HEALTHCHECK CMD curl --fail http://localhost:8080/ || exit 1
+
+# Run the FastAPI app
+# Using "app:app" ensures fast startup since app.py imports main.py lazily.
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8080"]

Muril-Model/Muril-Model/.gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+muril_multilingual_dataset.csv filter=lfs diff=lfs merge=lfs -text

Muril-Model/Muril-Model/.gitignore

@@ -0,0 +1,45 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# Virtual Environment
+venv/
+env/
+.env/
+.venv/
+
+# IDEs and Editors
+.idea/
+.vscode/
+*.swp
+*.swo
+*~
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# Model files and data
+*.h5
+*.pkl
+*.model
+data/
+models/
+logs/

Muril-Model/Muril-Model/Dockerfile

@@ -0,0 +1,28 @@
+# Lightweight Python image
+FROM python:3.10-slim
+
+# Disable CUDA & set cache
+ENV TORCH_DISABLE_CUDA=1
+ENV HF_HOME=/app/hf_cache
+ENV TRANSFORMERS_CACHE=/app/hf_cache
+
+# Working directory
+WORKDIR /app
+
+# Install dependencies
+COPY requirements.txt .
+RUN apt-get update && apt-get install -y git && \
+    pip install --no-cache-dir -r requirements.txt && \
+    rm -rf /var/lib/apt/lists/*
+
+# Copy app code and model/data
+COPY . .
+
+# Make cache folder writable
+RUN mkdir -p /app/hf_cache && chmod -R 777 /app/hf_cache
+
+# Expose port
+EXPOSE 8080
+
+# Run FastAPI via uvicorn
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8080"]

Muril-Model/Muril-Model/README.md

@@ -0,0 +1,10 @@
+---
+title: Muril Model
+emoji: 🌖
+colorFrom: gray
+colorTo: red
+sdk: docker
+pinned: false
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Muril-Model/Muril-Model/answer_embeddings.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fa2e749567247c8a15144c6a0b1d3423ae8a8a0054aee9f3cc2774f8b9cb555
+size 83854959

Muril-Model/Muril-Model/main.py

@@ -0,0 +1,70 @@
+import os
+import torch
+import pandas as pd
+from fastapi import FastAPI
+from pydantic import BaseModel
+from sentence_transformers import SentenceTransformer, util
+from huggingface_hub import snapshot_download
+
+# --- Cache Configuration ---
+os.environ["HF_HOME"] = "/app/hf_cache"
+os.environ["TRANSFORMERS_CACHE"] = "/app/hf_cache"
+os.environ["TORCH_DISABLE_CUDA"] = "1"
+
+# --- Hugging Face Repo ---
+HF_REPO = "Sp2503/Muril-Model"
+
+# --- Download model & embeddings from Hugging Face Hub ---
+print("📦 Downloading model & embeddings from Hugging Face Hub...")
+model_dir = snapshot_download(repo_id=HF_REPO, repo_type="model")
+print(f"✅ Model snapshot available at: {model_dir}")
+
+MODEL_PATH = model_dir
+CSV_PATH = os.path.join(model_dir, "muril_multilingual_dataset.csv")
+EMBED_PATH = os.path.join(model_dir, "answer_embeddings.pt")
+
+# --- Load resources ---
+print("⚙️ Loading model and embeddings...")
+model = SentenceTransformer(MODEL_PATH)
+df = pd.read_csv(CSV_PATH).dropna(subset=['question', 'answer'])
+answer_embeddings = torch.load(EMBED_PATH, map_location="cpu")
+print("✅ Model and embeddings loaded successfully.")
+
+# --- FastAPI Setup ---
+app = FastAPI(title="MuRIL Multilingual QA API")
+
+class QueryRequest(BaseModel):
+    question: str
+    lang: str = None
+
+class QAResponse(BaseModel):
+    answer: str
+
+@app.get("/")
+def root():
+    return {"status": "✅ API ready", "model_loaded": True}
+
+@app.post("/get-answer", response_model=QAResponse)
+def get_answer_endpoint(request: QueryRequest):
+    question_text = request.question.strip()
+    lang_filter = request.lang
+
+    filtered_df = df
+    filtered_embeddings = answer_embeddings
+    if 'lang' in df.columns and lang_filter:
+        mask = df['lang'] == lang_filter
+        filtered_df = df[mask].reset_index(drop=True)
+        filtered_embeddings = answer_embeddings[mask.values]
+
+    if len(filtered_df) == 0:
+        return {"answer": f"No data found for language '{lang_filter}'."}
+
+    question_emb = model.encode(question_text, convert_to_tensor=True)
+    cosine_scores = util.pytorch_cos_sim(question_emb, filtered_embeddings)
+    best_idx = torch.argmax(cosine_scores).item()
+    answer = filtered_df.iloc[best_idx]['answer']
+    return {"answer": answer}
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run("main:app", host="0.0.0.0", port=8080)

Muril-Model/Muril-Model/muril_combined_multilingual_model/1_Pooling/config.json

@@ -0,0 +1,10 @@
+{
+    "word_embedding_dimension": 768,
+    "pooling_mode_cls_token": false,
+    "pooling_mode_mean_tokens": true,
+    "pooling_mode_max_tokens": false,
+    "pooling_mode_mean_sqrt_len_tokens": false,
+    "pooling_mode_weightedmean_tokens": false,
+    "pooling_mode_lasttoken": false,
+    "include_prompt": true
+}

Muril-Model/Muril-Model/muril_combined_multilingual_model/config.json

@@ -0,0 +1,25 @@
+{
+  "architectures": [
+    "BertModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "dtype": "float32",
+  "embedding_size": 768,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "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",
+  "transformers_version": "4.56.2",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 197285
+}

Muril-Model/Muril-Model/muril_combined_multilingual_model/config_sentence_transformers.json

@@ -0,0 +1,14 @@
+{
+  "model_type": "SentenceTransformer",
+  "__version__": {
+    "sentence_transformers": "5.1.1",
+    "transformers": "4.56.2",
+    "pytorch": "2.8.0+cu126"
+  },
+  "prompts": {
+    "query": "",
+    "document": ""
+  },
+  "default_prompt_name": null,
+  "similarity_fn_name": "cosine"
+}

Muril-Model/Muril-Model/muril_combined_multilingual_model/modules.json

@@ -0,0 +1,14 @@
+[
+  {
+    "idx": 0,
+    "name": "0",
+    "path": "",
+    "type": "sentence_transformers.models.Transformer"
+  },
+  {
+    "idx": 1,
+    "name": "1",
+    "path": "1_Pooling",
+    "type": "sentence_transformers.models.Pooling"
+  }
+]

Muril-Model/Muril-Model/muril_combined_multilingual_model/sentence_bert_config.json

@@ -0,0 +1,4 @@
+{
+    "max_seq_length": 512,
+    "do_lower_case": false
+}

Muril-Model/Muril-Model/muril_combined_multilingual_model/special_tokens_map.json

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

Muril-Model/Muril-Model/muril_combined_multilingual_model/tokenizer_config.json

@@ -0,0 +1,59 @@
+{
+  "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
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "104": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "105": {
+      "content": "[SEP]",
+      "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": false,
+  "extra_special_tokens": {},
+  "lowercase": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": false,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

Muril-Model/Muril-Model/muril_multilingual_dataset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09cca0ed98f57664e558825059722272d15fe99f5238969e95f523629fb50cec
+size 16996056

Muril-Model/Muril-Model/precompute_embeddings.py

@@ -0,0 +1,17 @@
+import torch
+import pandas as pd
+from sentence_transformers import SentenceTransformer
+
+MODEL_PATH = './muril_combined_multilingual_model'
+CSV_PATH = './muril_multilingual_dataset.csv'
+EMB_PATH = './answer_embeddings.pt'
+
+print("🔄 Precomputing embeddings...")
+model = SentenceTransformer(MODEL_PATH)
+df = pd.read_csv(CSV_PATH).dropna(subset=['question', 'answer'])
+
+answers = df['answer'].tolist()
+answer_embeddings = model.encode(answers, convert_to_tensor=True)
+torch.save(answer_embeddings, EMB_PATH)
+
+print(f"✅ Saved {len(answers)} embeddings to {EMB_PATH}")

Muril-Model/Muril-Model/requirements.txt

@@ -0,0 +1,12 @@
+fastapi==0.118.0
+uvicorn==0.37.0
+torch==2.1.0
+sentence-transformers==5.1.1
+transformers==4.43.3
+numpy<2
+pandas==2.1.1
+langdetect==1.0.9
+requests==2.31.0
+tqdm==4.65.0
+PyMuPDF==1.23.0
+huggingface_hub==0.23.4

Muril-Model/README.md

@@ -0,0 +1,10 @@
+---
+title: Muril Model
+emoji: 🌖
+colorFrom: gray
+colorTo: red
+sdk: docker
+pinned: false
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Muril-Model/answer_embeddings.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fa2e749567247c8a15144c6a0b1d3423ae8a8a0054aee9f3cc2774f8b9cb555
+size 83854959

Muril-Model/app.py

@@ -0,0 +1,4 @@
+# app.py
+import importlib
+
+app = importlib.import_module("main").app

Muril-Model/main.py

@@ -0,0 +1,107 @@
+import os
+import torch
+import pandas as pd
+from fastapi import FastAPI
+from pydantic import BaseModel
+from sentence_transformers import SentenceTransformer, util
+from langdetect import detect
+from huggingface_hub import hf_hub_download
+import threading
+import time
+
+# --- Cache Configuration ---
+os.environ["HF_HOME"] = "/app/hf_cache"
+os.environ["TRANSFORMERS_CACHE"] = "/app/hf_cache"
+os.environ["TORCH_DISABLE_CUDA"] = "1"
+
+# --- Paths ---
+MODEL_PATH = './muril_combined_multilingual_model'
+CSV_PATH = './muril_multilingual_dataset.csv'
+HF_REPO = "Sp2503/muril-dataset"
+HF_FILE = "answer_embeddings.pt"
+
+# --- FastAPI Setup ---
+app = FastAPI(title="MuRIL Multilingual QA API")
+
+# Global variables
+model = None
+df = None
+answer_embeddings = None
+
+# --- Helper: Load embeddings from Hugging Face ---
+def load_embeddings():
+    print("📥 Downloading embeddings from Hugging Face...")
+    hf_path = hf_hub_download(
+        repo_id=HF_REPO,
+        filename=HF_FILE,
+        repo_type="dataset",
+        cache_dir="/tmp"
+    )
+    print(f"✅ Embeddings available at {hf_path}")
+    return torch.load(hf_path, map_location="cpu")
+
+# --- Resource Loader ---
+def load_resources():
+    global model, df, answer_embeddings
+    try:
+        print("⚙️ Loading model and dataset...")
+        model = SentenceTransformer(MODEL_PATH)
+        df = pd.read_csv(CSV_PATH).dropna(subset=['question', 'answer'])
+        answer_embeddings = load_embeddings()
+        print("✅ Model and embeddings ready.")
+    except Exception as e:
+        print(f"❌ Error loading resources: {e}")
+
+# --- Background Loader Thread ---
+@app.on_event("startup")
+def startup_event():
+    print("🚀 Starting background model loader thread...")
+    thread = threading.Thread(target=load_resources)
+    thread.daemon = True
+    thread.start()
+
+@app.get("/")
+def root():
+    ready = model is not None and df is not None and answer_embeddings is not None
+    return {"status": "✅ Running MuRIL QA API", "model_loaded": ready}
+
+class QueryRequest(BaseModel):
+    question: str
+    lang: str = None
+
+class QAResponse(BaseModel):
+    answer: str
+
+@app.post("/get-answer", response_model=QAResponse)
+def get_answer_endpoint(request: QueryRequest):
+    if model is None or df is None or answer_embeddings is None:
+        return {"answer": "⏳ Model still loading, please try again shortly."}
+
+    question_text = request.question.strip()
+    lang_filter = request.lang or detect(question_text)
+
+    filtered_df = df
+    filtered_embeddings = answer_embeddings
+    if 'lang' in df.columns and lang_filter:
+        mask = df['lang'] == lang_filter
+        filtered_df = df[mask].reset_index(drop=True)
+        filtered_embeddings = answer_embeddings[mask.values]
+
+    if len(filtered_df) == 0:
+        return {"answer": f"⚠️ No data found for language '{lang_filter}'."}
+
+    question_emb = model.encode(question_text, convert_to_tensor=True)
+    cosine_scores = util.pytorch_cos_sim(question_emb, filtered_embeddings)
+    best_idx = torch.argmax(cosine_scores).item()
+    answer = filtered_df.iloc[best_idx]['answer']
+    return {"answer": answer}
+
+# --- Keep-alive thread for Spaces ---
+def keep_alive():
+    while True:
+        time.sleep(60)
+
+if __name__ == "__main__":
+    import uvicorn
+    threading.Thread(target=keep_alive, daemon=True).start()
+    uvicorn.run("main:app", host="0.0.0.0", port=8080)

Muril-Model/muril_combined_multilingual_model/1_Pooling/config.json

@@ -0,0 +1,10 @@
+{
+    "word_embedding_dimension": 768,
+    "pooling_mode_cls_token": false,
+    "pooling_mode_mean_tokens": true,
+    "pooling_mode_max_tokens": false,
+    "pooling_mode_mean_sqrt_len_tokens": false,
+    "pooling_mode_weightedmean_tokens": false,
+    "pooling_mode_lasttoken": false,
+    "include_prompt": true
+}

Muril-Model/muril_combined_multilingual_model/config.json

@@ -0,0 +1,25 @@
+{
+  "architectures": [
+    "BertModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "dtype": "float32",
+  "embedding_size": 768,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "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",
+  "transformers_version": "4.56.2",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 197285
+}

Muril-Model/muril_combined_multilingual_model/config_sentence_transformers.json

@@ -0,0 +1,14 @@
+{
+  "model_type": "SentenceTransformer",
+  "__version__": {
+    "sentence_transformers": "5.1.1",
+    "transformers": "4.56.2",
+    "pytorch": "2.8.0+cu126"
+  },
+  "prompts": {
+    "query": "",
+    "document": ""
+  },
+  "default_prompt_name": null,
+  "similarity_fn_name": "cosine"
+}

Muril-Model/muril_combined_multilingual_model/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:efc762382be48593e4d42fd7ddb7ba61a013d7447bc027b5d4f345d42bf8427a
+size 950247272

Muril-Model/muril_combined_multilingual_model/modules.json

@@ -0,0 +1,14 @@
+[
+  {
+    "idx": 0,
+    "name": "0",
+    "path": "",
+    "type": "sentence_transformers.models.Transformer"
+  },
+  {
+    "idx": 1,
+    "name": "1",
+    "path": "1_Pooling",
+    "type": "sentence_transformers.models.Pooling"
+  }
+]

Muril-Model/muril_combined_multilingual_model/sentence_bert_config.json

@@ -0,0 +1,4 @@
+{
+    "max_seq_length": 512,
+    "do_lower_case": false
+}

Muril-Model/muril_combined_multilingual_model/special_tokens_map.json

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

Muril-Model/muril_combined_multilingual_model/tokenizer_config.json

@@ -0,0 +1,59 @@
+{
+  "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
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "104": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "105": {
+      "content": "[SEP]",
+      "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": false,
+  "extra_special_tokens": {},
+  "lowercase": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": false,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

Muril-Model/muril_multilingual_dataset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09cca0ed98f57664e558825059722272d15fe99f5238969e95f523629fb50cec
+size 16996056

Muril-Model/precompute_embeddings.py

@@ -0,0 +1,17 @@
+import torch
+import pandas as pd
+from sentence_transformers import SentenceTransformer
+
+MODEL_PATH = './muril_combined_multilingual_model'
+CSV_PATH = './muril_multilingual_dataset.csv'
+EMB_PATH = './answer_embeddings.pt'
+
+print("🔄 Precomputing embeddings...")
+model = SentenceTransformer(MODEL_PATH)
+df = pd.read_csv(CSV_PATH).dropna(subset=['question', 'answer'])
+
+answers = df['answer'].tolist()
+answer_embeddings = model.encode(answers, convert_to_tensor=True)
+torch.save(answer_embeddings, EMB_PATH)
+
+print(f"✅ Saved {len(answers)} embeddings to {EMB_PATH}")

Muril-Model/requirements.txt

@@ -0,0 +1,12 @@
+fastapi==0.118.0
+uvicorn==0.37.0
+torch==2.1.0
+sentence-transformers==5.1.1
+transformers==4.43.3
+numpy<2
+pandas==2.1.1
+langdetect==1.0.9
+requests==2.31.0
+tqdm==4.65.0
+PyMuPDF==1.23.0
+huggingface_hub==0.23.4

pdfs/constitution_1.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:819ba7adc5e5ae4063f26c82472283ba682ddb662e5a7f864cf26984d99b26ed
+size 2413611

pdfs/constitution_10.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:368194dab2e116625e62c9f9f6164d9da142fd4a255ec92fcd93e6f4bfcd71c9
+size 16375014

pdfs/constitution_11.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:690b780cf35413e92949ad48c92f9c55241623adf9fcddc6cc31c8ef2ca68a38
+size 6510505

pdfs/constitution_12.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3e78b81b1551aa11922dd56a128947e483347870518fb40323a20db55b48d0e
+size 30214317

pdfs/constitution_2.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7646d0ad9ef704dd48adddd84f1c1649c297a9233e2e3a4aa21d6e470c0303b7
+size 5499505

pdfs/constitution_3.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdd995f5349c7f32a544a0ae85b389820a0475b9558ea4c1e5f0a2b0a640c39d
+size 8559896

pdfs/constitution_4.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7dae1ef1ad3c201563563b68152b685b2b8ca0a307a75a53e818f1bb81488675
+size 6685693

pdfs/constitution_5.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2409ce877cd34104df1831aa241ab6077ca0c5845451fbf68aa85693850cdca
+size 74393982