Spaces:

ClemSummer
/

ai-lab

Running on CPU Upgrade

App Files Files Community

ClemSummer commited on 9 days ago

Commit

79a071f

1 Parent(s): 3aa3d4c

Added two more games. Container runs locally

Browse files

Files changed (17) hide show

.dockerignore +1 -0
.gitignore +2 -1
Dockerfile +8 -1
cbow_logic.py +52 -0
content_distillery/static/content_distillery.html +120 -0
main.py +84 -18
ppo_summarizer/predict_ppo.py → ppo_logic.py +14 -12
qwen_loRA/README.md +207 -0
qwen_loRA/adapter_config.json +36 -0
qwen_loRA/adapter_model.safetensors +3 -0
qwen_loRA/reward_head.pt +3 -0
requirements.txt +3 -0
save_cbow_model.py +16 -0
templates/cbow.html +248 -0
vit_captioning/generate.py +12 -7
vit_captioning/models/encoder.py +1 -6
vit_captioning/static/landing.html +4 -4

.dockerignore CHANGED Viewed

@@ -27,6 +27,7 @@ clip-checkpoints/
 *.pt
 *.pth
 *.onnx
 # Docker or Space-specific
 docker-compose.yaml

 *.pt
 *.pth
 *.onnx
+models/
 # Docker or Space-specific
 docker-compose.yaml

.gitignore CHANGED Viewed

@@ -1,3 +1,4 @@
 __pycache__/
 *.png
-**/artifacts/

 __pycache__/
 *.png
+**/artifacts/
+models/

Dockerfile CHANGED Viewed

@@ -19,9 +19,16 @@ RUN pip install -r requirements.txt
 RUN mkdir -p /models/clip && \
     python3 -c "from transformers import CLIPModel; CLIPModel.from_pretrained('openai/clip-vit-base-patch32').save_pretrained('/models/clip')"
 RUN python3 -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('bert-base-uncased').save_pretrained('/models/bert-tokenizer')"
 RUN python3 -c "from transformers import CLIPProcessor; CLIPProcessor.from_pretrained('openai/clip-vit-base-patch32').save_pretrained('/models/clip')"
 EXPOSE 7860
 # Install curl if it's not already installed

 RUN mkdir -p /models/clip && \
     python3 -c "from transformers import CLIPModel; CLIPModel.from_pretrained('openai/clip-vit-base-patch32').save_pretrained('/models/clip')"
 RUN python3 -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('bert-base-uncased').save_pretrained('/models/bert-tokenizer')"
 RUN python3 -c "from transformers import CLIPProcessor; CLIPProcessor.from_pretrained('openai/clip-vit-base-patch32').save_pretrained('/models/clip')"
+RUN mkdir -p /models/cbow && \
+    python3 -c "import gensim.downloader as api; model = api.load('glove-twitter-200'); model.save('/models/cbow_model.kv')"
+RUN mkdir -p /models/qwen && \
+    python3 -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('Qwen/Qwen3-0.6B-Base').save_pretrained('/models/qwen')"
+RUN python3 -c "from transformers import AutoModelForCausalLM; AutoModelForCausalLM.from_pretrained('Qwen/Qwen3-0.6B-Base').save_pretrained('/models/qwen')"
 EXPOSE 7860
 # Install curl if it's not already installed

cbow_logic.py ADDED Viewed

	@@ -0,0 +1,52 @@

+# cbow_logic.py
+import gensim
+import os
+import argparse
+from typing import List, Tuple
+import shlex
+class MeaningCalculator:
+    def __init__(self, model_path: str = "/models/cbow_model.kv"):
+        if not os.path.exists(model_path):
+            raise FileNotFoundError(f"Model not found at: {model_path}")
+        self.model = gensim.models.KeyedVectors.load(model_path, mmap='r')
+    def evaluate_expression(self, expression: str, topn: int = 10) -> List[Tuple[str, float]]:
+    # Evaluate expressions like '"new york" - city + capital'.
+        tokens = shlex.split(expression)  # Handles quoted terms properly
+        positive = []
+        negative = []
+        current_op = "+"
+        for token in tokens:
+            print(token)
+            if token in ["+", "-"]:
+                current_op = token
+            else:
+                if current_op == "+":
+                    positive.append(token)
+                else:
+                    negative.append(token)
+        try:
+            return self.model.most_similar(positive=positive, negative=negative, topn=topn)
+        except KeyError as e:
+            return [("InputError", 0.0)]
+from gensim.models import KeyedVectors
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="Evaluate word vector expressions using CBOW.")
+    parser.add_argument("expression", type=str, help="Expression like 'king - man + woman'")
+    parser.add_argument("--model_path", type=str, default="./models/cbow_model.kv", help="Path to CBOW model")
+    args = parser.parse_args()
+    calc = MeaningCalculator(model_path=args.model_path)
+    results = calc.evaluate_expression(args.expression)
+    print(f"\nExpression: {args.expression}\nTop Results:")
+    for word, score in results:
+        print(f"  {word:<15} {score:.4f}")

content_distillery/static/content_distillery.html ADDED Viewed

	@@ -0,0 +1,120 @@

+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8" />
+  <title>📚 Content Distillery</title>
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <script src="https://cdn.tailwindcss.com"></script>
+</head>
+<body class="bg-gray-100 min-h-screen p-6 flex flex-col items-center">
+  <div class="max-w-2xl w-full space-y-4">
+    <h1 class="text-3xl font-bold text-gray-800 text-center">📚 Content Distillery</h1>
+    <!-- Source Dropdown with Refresh Button -->
+    <label class="block font-semibold text-gray-700">Choose a Source:</label>
+    <div class="flex items-center space-x-2">
+      <select id="sourceSelect" class="flex-grow p-2 border rounded">
+        <option value="">-- Select Source --</option>
+        <option value="reddit_romance">Romantic Relationship</option>
+        <option value="reddit_aita">Am I The Asshole</option>
+        <option value="reddit_careers">Career Guidance</option>
+        <option value="reddit_cars">Car discussion</option>
+        <option value="reddit_whatcarshouldibuy">What Car Should I Buy</option>
+        <option value="reddit_nosleep">Horror stories</option>
+        <option value="reddit_maliciouscompliance">People following bad instructions exactly</option>
+        <option value="reddit_talesfromtechsupport">Tech support stories</option>
+        <option value="reddit_decidingtobebetter">Self-improvement and habit change</option>
+        <option value="reddit_askphilosophy">Big-life questions</option>
+      </select>
+      <button id="refreshPost" title="Get another post"
+        class="p-2 bg-gray-200 hover:bg-gray-300 rounded text-lg">🔄</button>
+    </div>
+    <!-- Input Text -->
+    <label class="block font-semibold text-gray-700 mt-4">Original Text: (can modify or create your own)</label>
+    <textarea id="inputText" rows="8" class="w-full p-3 border rounded" placeholder="Paste or fetch post text here..."></textarea>
+    <!-- Generate Button -->
+    <button id="generateBtn" class="w-full bg-red-600 text-white py-2 rounded hover:bg-red-700 font-semibold">
+      🤖 Generate Summary
+    </button>
+    <!-- Output Text -->
+    <label class="block font-semibold text-gray-700 mt-4">Summary:</label>
+    <textarea id="outputText" rows="5" class="w-full p-3 border rounded bg-gray-50" readonly placeholder="Summary will appear here..."></textarea>
+  </div>
+  <!-- Floating Help Button -->
+  <button id="helpButton"
+    class="fixed bottom-4 right-4 bg-blue-600 text-white rounded-full w-12 h-12 text-2xl font-bold shadow-lg hover:bg-blue-700 transition">
+    ?
+  </button>
+  <!-- Help Modal -->
+  <div id="helpModal" class="fixed inset-0 bg-black bg-opacity-50 flex items-center justify-center hidden">
+    <div class="bg-white rounded-lg p-6 max-w-sm w-full shadow-lg text-left">
+      <h2 class="text-xl font-semibold mb-4">About Content Distillery</h2>
+      <p class="text-gray-700 mb-4">
+        This tool fetches content from various online sources and distills it into concise summaries using an a PPO model.
+        Choose a source, optionally edit the text, and press "Generate" to see a summary.
+        The AI reward and PPO models were from Qwen/Qwen3-0.6B-Base, used openai/summarize_from_feedback dataset and trained according to OpenAI paper "Learning to summarize from human feedback".
+      </p>
+      <button id="closeModal"
+        class="mt-2 bg-blue-600 text-white px-4 py-2 rounded hover:bg-blue-700">
+        Close
+      </button>
+    </div>
+  </div>
+  <script>
+    const helpButton = document.getElementById('helpButton');
+    const helpModal = document.getElementById('helpModal');
+    const closeModal = document.getElementById('closeModal');
+    helpButton.addEventListener('click', () => helpModal.classList.remove('hidden'));
+    closeModal.addEventListener('click', () => helpModal.classList.add('hidden'));
+    helpModal.addEventListener('click', (e) => {
+      if (e.target === helpModal) helpModal.classList.add('hidden');
+    });
+    async function fetchPost() {
+      const selected = document.getElementById("sourceSelect").value;
+      if (!selected) return;
+      document.getElementById("inputText").value = "Fetching post...";
+      const res = await fetch(`/get_sample?source=${selected}`);
+      const data = await res.text();
+      document.getElementById("inputText").value = data;
+    }
+    document.getElementById("sourceSelect").addEventListener("change", fetchPost);
+    document.getElementById("refreshPost").addEventListener("click", fetchPost);
+    document.getElementById('generateBtn').addEventListener('click', async function () {
+      const btn = document.getElementById('generateBtn');
+      const post = document.getElementById('inputText').value;
+      btn.disabled = true;
+      btn.textContent = "Generating...";
+      btn.classList.add("opacity-50");
+      try {
+        const res = await fetch('/contentdistillery', {
+          method: 'POST',
+          headers: { 'Content-Type': 'application/x-www-form-urlencoded' },
+          body: new URLSearchParams({ post })
+        });
+        const summary = await res.text();
+        document.getElementById('outputText').value = summary;
+      } catch (e) {
+        document.getElementById('outputText').value = "Error generating summary.";
+      }
+      btn.disabled = false;
+      btn.textContent = "Generate Summary";
+      btn.classList.remove("opacity-50");
+    });
+  </script>
+</body>
+</html>

main.py CHANGED Viewed

@@ -1,16 +1,34 @@
 # app/main.py
-from fastapi import FastAPI, UploadFile, File
-from fastapi.responses import HTMLResponse
 from fastapi.staticfiles import StaticFiles
 import shutil
 from pathlib import Path
 import uvicorn
 import os
 from vit_captioning.generate import CaptionGenerator
 app = FastAPI()
 # Serve static files
 static_dir = Path(__file__).parent / "vit_captioning" / "static"
@@ -21,10 +39,6 @@ app.mount("/static", StaticFiles(directory=static_dir), name="static")
 async def landing():
     return Path("vit_captioning/static/landing.html").read_text()
-# @app.get("/", response_class=HTMLResponse)
-# def root():
-#     return "<h3>✅ Hugging Face Space is alive</h3>"
 @app.get("/health")
 def health_check():
     return {"status": "ok"}
@@ -34,22 +48,17 @@ def health_check():
 async def captioning():
     return Path("vit_captioning/static/captioning/index.html").read_text()
-# ✅ Example: Project 2 placeholder
-@app.get("/project2", response_class=HTMLResponse)
-async def project2():
-    return "<h1>Coming Soon: Project 2</h1>"
-# ✅ Example: Project 2 placeholder
-@app.get("/project3", response_class=HTMLResponse)
-async def project2():
-    return "<h1>Coming Soon: Project 3</h1>"
 # ✅ Caption generation endpoint for captioning app
 # Keep the path consistent with your JS fetch()!
 caption_generator = CaptionGenerator(
     model_type="CLIPEncoder",
     checkpoint_path="./vit_captioning/artifacts/CLIPEncoder_40epochs_unfreeze12.pth",
-    quantized=False
 )
 @app.post("/generate")
@@ -61,5 +70,62 @@ async def generate(file: UploadFile = File(...)):
     captions = caption_generator.generate_caption(temp_file)
     return captions
-# if __name__ == "__main__":
-#     uvicorn.run(app, host="0.0.0.0", port=8000)

 # app/main.py
+from fastapi import FastAPI, UploadFile, File, Request, Form, Query
+from fastapi.responses import HTMLResponse, PlainTextResponse
 from fastapi.staticfiles import StaticFiles
+from fastapi.templating import Jinja2Templates
+from cbow_logic import MeaningCalculator
+from ppo_logic import generate_summary
+import numpy as np
+import json
 import shutil
 from pathlib import Path
 import uvicorn
 import os
+import praw
+import random
 from vit_captioning.generate import CaptionGenerator
+from cbow_logic import MeaningCalculator
+reddit = praw.Reddit(
+    client_id="geuNJZLDwSCdz7sV5vkDNQ",
+    client_secret="IFz7zPVGP3hO6VMy1YU1WX_bX3FpfQ",
+    user_agent="ContentDistilleryBot/0.1 by ClementHa"
+)
 app = FastAPI()
+templates = Jinja2Templates(directory="templates")
+calculator = MeaningCalculator()
 # Serve static files
 static_dir = Path(__file__).parent / "vit_captioning" / "static"
 async def landing():
     return Path("vit_captioning/static/landing.html").read_text()
 @app.get("/health")
 def health_check():
     return {"status": "ok"}
 async def captioning():
     return Path("vit_captioning/static/captioning/index.html").read_text()
+@app.get("/contentdistillery", response_class=HTMLResponse)
+async def contentdistillery():
+    return Path("content_distillery/static/content_distillery.html").read_text()
 # ✅ Caption generation endpoint for captioning app
 # Keep the path consistent with your JS fetch()!
 caption_generator = CaptionGenerator(
     model_type="CLIPEncoder",
     checkpoint_path="./vit_captioning/artifacts/CLIPEncoder_40epochs_unfreeze12.pth",
+    quantized=False,
+    runAsContainer=False
 )
 @app.post("/generate")
     captions = caption_generator.generate_caption(temp_file)
     return captions
+@app.get("/cbow", response_class=HTMLResponse)
+async def cbow_form(request: Request):
+    return templates.TemplateResponse("cbow.html", {"request": request})
+@app.post("/cbow")
+async def cbow(request: Request, expression: str = Form(...)):
+    expression = expression.lower()
+    results = MeaningCalculator().evaluate_expression(expression = expression)
+    # formatted = [
+    #     (word, f"{score:.2f}" if score >= 0.4 else "Irrelevant result")
+    #     for word, score in results[:5]
+    # ]
+    return templates.TemplateResponse("cbow.html", {
+        "request": request,
+        "expression": expression,
+        "results": results
+    })
+@app.get("/contentdistillery", response_class=HTMLResponse)
+async def contentdistillery_page():
+    return Path("contentdistillery.html").read_text(encoding="utf-8")
+@app.post("/contentdistillery", response_class=PlainTextResponse)
+async def generate_summary_from_post(post: str = Form(...)):
+    return generate_summary(post)
+@app.get("/get_sample", response_class=PlainTextResponse)
+def get_sample(source: str = Query(...)):
+    try:
+        if source == "reddit_romance":
+            submissions = reddit.subreddit("relationships").top(limit=10)
+        elif source == "reddit_aita":
+            submissions = reddit.subreddit("AmItheAsshole").hot(limit=10)
+        elif source == "reddit_careers":
+            submissions = reddit.subreddit("careerguidance").hot(limit=10)
+        elif source == "reddit_cars":
+            submissions = reddit.subreddit("cars").hot(limit=10)
+        elif source == "reddit_whatcarshouldibuy":
+            submissions = reddit.subreddit("whatcarshouldibuy").top(limit=10)
+        elif source == "reddit_nosleep":
+            submissions = reddit.subreddit("nosleep").top(limit=10)
+        elif source == "reddit_maliciouscompliance":
+            submissions = reddit.subreddit("maliciouscompliance").hot(limit=10)
+        elif source == "reddit_talesfromtechsupport":
+            submissions = reddit.subreddit("talesfromtechsupport").top(limit=10)
+        elif source == "reddit_decidingtobebetter":
+            submissions = reddit.subreddit("decidingtobebetter").hot(limit=10)
+        elif source == "reddit_askphilosophy":
+            submissions = reddit.subreddit("askphilosophy").top(limit=10)
+        else:
+            return "Unsupported source."
+        posts = [s.selftext.strip() for s in submissions if s.selftext.strip()]
+        if posts:
+            return random.choice(posts)
+        return "No suitable post found."
+    except Exception as e:
+        return f"Error fetching Reddit post: {str(e)}"

ppo_summarizer/predict_ppo.py → ppo_logic.py RENAMED Viewed

@@ -9,10 +9,11 @@ import os
 # -------------------------------
 # Config
 # -------------------------------
-MODEL_NAME = "Qwen/Qwen3-0.6B-Base"
-CHECKPOINT_DIR = "./artifacts/qwen_loRA"
 DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-MAX_NEW_TOKENS = 64
 # -------------------------------
 # Load tokenizer and model
@@ -34,18 +35,21 @@ model = model.to(DEVICE)
 # -------------------------------
 # Generate Summary
 # -------------------------------
-def generate_summary(title: str, post: str) -> str:
-    prompt = f"Title: {title}\n\nPost: {post}\n\nSummary:"
     inputs = tokenizer(prompt, return_tensors="pt", padding=True).to(DEVICE)
     with torch.no_grad():
         outputs = model.generate(
             **inputs,
             max_new_tokens=MAX_NEW_TOKENS,
-            do_sample=True,
-            top_k=50,
-            top_p=0.95,
-            temperature=0.7,
             pad_token_id=tokenizer.pad_token_id,
             use_cache=True
         )
@@ -59,13 +63,11 @@ def generate_summary(title: str, post: str) -> str:
 # -------------------------------
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Generate summary with trained Qwen PPO model")
-    parser.add_argument("--title", type=str, required=True, help="Title of the post")
     parser.add_argument("--post", type=str, required=True, help="Content of the post")
     args = parser.parse_args()
-    print("\n📘 Title:", args.title)
     print("📝 Post:", args.post[:100] + ("..." if len(args.post) > 100 else ""))
     print("\n🤖 Generating summary...\n")
-    summary = generate_summary(args.title, args.post)
     print("✅ Summary:\n", summary)

 # -------------------------------
 # Config
 # -------------------------------
+#MODEL_NAME = "Qwen/Qwen3-0.6B-Base"
+MODEL_NAME = "/models/qwen"
+CHECKPOINT_DIR = "./qwen_loRA"
 DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+MAX_NEW_TOKENS = 256
 # -------------------------------
 # Load tokenizer and model
 # -------------------------------
 # Generate Summary
 # -------------------------------
+def generate_summary(post: str) -> str:
+    #prompt = f"Instruction: Summarize the post in one sentence.\n\nPost:\n{post}\n\nSummary:"
+    # prompt = f"Please summarize the following Reddit post in 1–2 sentences:\n\n{post}\n\nSummary:"
+    prompt = f"Instruction: Summarize the post in 1-2 sentences.\n\nPost:\n{post}\n\nSummary:"
     inputs = tokenizer(prompt, return_tensors="pt", padding=True).to(DEVICE)
     with torch.no_grad():
         outputs = model.generate(
             **inputs,
             max_new_tokens=MAX_NEW_TOKENS,
+            do_sample=False,
+            # top_k=50,
+            # top_p=0.95,
+            temperature=1.0,
             pad_token_id=tokenizer.pad_token_id,
             use_cache=True
         )
 # -------------------------------
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Generate summary with trained Qwen PPO model")
     parser.add_argument("--post", type=str, required=True, help="Content of the post")
     args = parser.parse_args()
     print("📝 Post:", args.post[:100] + ("..." if len(args.post) > 100 else ""))
     print("\n🤖 Generating summary...\n")
+    summary = generate_summary(args.post)
     print("✅ Summary:\n", summary)

qwen_loRA/README.md ADDED Viewed

	@@ -0,0 +1,207 @@

+---
+base_model: Qwen/Qwen3-0.6B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-0.6B-Base
+- lora
+- transformers
+---
+# Model Card for Model ID
+<!-- Provide a quick summary of what the model is/does. -->
+## Model Details
+### Model Description
+<!-- Provide a longer summary of what this model is. -->
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+### Model Sources [optional]
+<!-- Provide the basic links for the model. -->
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+## Uses
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+### Direct Use
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+[More Information Needed]
+### Downstream Use [optional]
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+[More Information Needed]
+### Out-of-Scope Use
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+[More Information Needed]
+## Bias, Risks, and Limitations
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+[More Information Needed]
+### Recommendations
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+## How to Get Started with the Model
+Use the code below to get started with the model.
+[More Information Needed]
+## Training Details
+### Training Data
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+[More Information Needed]
+### Training Procedure
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+#### Preprocessing [optional]
+[More Information Needed]
+#### Training Hyperparameters
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+#### Speeds, Sizes, Times [optional]
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+[More Information Needed]
+## Evaluation
+<!-- This section describes the evaluation protocols and provides the results. -->
+### Testing Data, Factors & Metrics
+#### Testing Data
+<!-- This should link to a Dataset Card if possible. -->
+[More Information Needed]
+#### Factors
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+[More Information Needed]
+#### Metrics
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+[More Information Needed]
+### Results
+[More Information Needed]
+#### Summary
+## Model Examination [optional]
+<!-- Relevant interpretability work for the model goes here -->
+[More Information Needed]
+## Environmental Impact
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+## Technical Specifications [optional]
+### Model Architecture and Objective
+[More Information Needed]
+### Compute Infrastructure
+[More Information Needed]
+#### Hardware
+[More Information Needed]
+#### Software
+[More Information Needed]
+## Citation [optional]
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+**BibTeX:**
+[More Information Needed]
+**APA:**
+[More Information Needed]
+## Glossary [optional]
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+[More Information Needed]
+## More Information [optional]
+[More Information Needed]
+## Model Card Authors [optional]
+[More Information Needed]
+## Model Card Contact
+[More Information Needed]
+### Framework versions
+- PEFT 0.16.0

qwen_loRA/adapter_config.json ADDED Viewed

	@@ -0,0 +1,36 @@

+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-0.6B-Base",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

qwen_loRA/adapter_model.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:c32e6b6f21996e7a38d4d017c3e9addc2a6aa24e0e03f13fe6af64052cfc0701
+size 4602248

qwen_loRA/reward_head.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:12b7d7d367d489625b74f6e6d9560acfb0349e6d4042c8f507b4484dd4a76fa5
+size 6021

requirements.txt CHANGED Viewed

@@ -6,3 +6,6 @@ numpy<2
 transformers
 pillow
 python-multipart

 transformers
 pillow
 python-multipart
+gensim
+peft
+praw

save_cbow_model.py ADDED Viewed

	@@ -0,0 +1,16 @@

+import gensim.downloader as api
+model = api.load("glove-twitter-200")
+print("Model loaded.")
+print("new-york" in model.key_to_index)       # ✅ True if token is present
+print("new" in model.key_to_index)            # ✅ Also true
+print("new york" in model.key_to_index)       # ❌ False — space not valid
+# Optional: print 5 most similar to test
+if "new-york" in model.key_to_index:
+    print(model.most_similar("new-york"))
+compound_terms = [key for key in model.key_to_index if "-" in key]
+print(f"Sample compound tokens: {compound_terms[:10]}")
+model.save("../models/cbow_model.kv")

templates/cbow.html ADDED Viewed

	@@ -0,0 +1,248 @@

+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>CBOW Vector Calculator</title>
+    <style>
+        body {
+            font-family: Arial, sans-serif;
+            max-width: 600px;
+            margin: auto;
+            padding: 1rem;
+            background-color: #f8f9fa;
+            position: relative;
+        }
+        h2 {
+            text-align: center;
+        }
+        form {
+            margin-top: 2rem;
+        }
+        textarea {
+            width: 100%;
+            height: 80px;
+            padding: 0.5rem;
+            font-size: 1rem;
+            resize: vertical;
+        }
+        button {
+            margin-top: 1rem;
+            width: 100%;
+            padding: 0.75rem;
+            font-size: 1rem;
+            background-color: #007bff;
+            color: white;
+            border: none;
+            border-radius: 4px;
+            cursor: pointer;
+        }
+        button:hover {
+            background-color: #0056b3;
+        }
+        .results {
+            margin-top: 2rem;
+            background-color: white;
+            padding: 1rem;
+            border-radius: 6px;
+            box-shadow: 0 0 10px rgba(0,0,0,0.1);
+        }
+        .result-row {
+            margin: 0.5rem 0;
+        }
+        .result-word {
+            font-weight: bold;
+        }
+        .score {
+            color: #666;
+            margin-left: 0.5rem;
+        }
+        .floating-icons {
+            position: fixed;
+            bottom: 1rem;
+            right: 1rem;
+            display: flex;
+            gap: 1rem;
+        }
+        .icon-button {
+            background: white;
+            border: 1px solid #ccc;
+            border-radius: 50%;
+            width: 40px;
+            height: 40px;
+            font-size: 1.2rem;
+            text-align: center;
+            line-height: 40px;
+            cursor: pointer;
+            box-shadow: 0 2px 6px rgba(0, 0, 0, 0.15);
+        }
+        .modal {
+            display: none;
+            position: fixed;
+            z-index: 1000;
+            left: 0; top: 0;
+            width: 100%; height: 100%;
+            background-color: rgba(0,0,0,0.4);
+        }
+        .modal-content {
+            background-color: #fff;
+            margin: 10% auto;
+            padding: 2rem;
+            border-radius: 8px;
+            max-width: 400px;
+            position: relative;
+        }
+        .close {
+            position: absolute;
+            top: 0.5rem;
+            right: 0.75rem;
+            font-size: 1.2rem;
+            cursor: pointer;
+        }
+    </style>
+</head>
+<body>
+    <h2>CBOW Vector Calculator</h2>
+    <form method="post" action="/cbow">
+        <label for="expression">Enter a word vector expression <small>(e.g. <code>king - man + woman</code>)</small>:</label><br>
+        <textarea name="expression" rows="4" style="width: 100%">{{ expression or "" }}</textarea>
+        <button type="submit">Calculate</button>
+    </form>
+    {% if results %}
+<div class="results">
+    {% if results and results|length > 0 %}
+    <p><strong>{{ expression }}</strong> ≈ <strong>{{ results[0][0] }}</strong></p>
+    {% endif %}
+    <h3>Results:</h3>
+    <table style="width: 100%; border-collapse: collapse;">
+        <thead>
+            <tr>
+                <th style="text-align: left; padding: 0.5rem; border-bottom: 1px solid #ccc;">#</th>
+                <th style="text-align: left; padding: 0.5rem; border-bottom: 1px solid #ccc;">Result</th>
+                <th style="text-align: left; padding: 0.5rem; border-bottom: 1px solid #ccc;">Score</th>
+            </tr>
+        </thead>
+        <tbody>
+            {% for word, score in results %}
+            <tr>
+                <td style="padding: 0.5rem;">{{ loop.index }}</td>
+                <td style="padding: 0.5rem;">{{ word }}</td>
+                <td style="padding: 0.5rem;">
+                    {% if score >= 0.4 %}
+                        {{ "%.2f"|format(score) }}
+                    {% else %}
+                        Irrelevant result
+                    {% endif %}
+                </td>
+            </tr>
+            {% endfor %}
+        </tbody>
+    </table>
+</div>
+    {% endif %}
+    <div class="floating-icons">
+        <div class="icon-button" onclick="openModal('suggestionsModal')">💡</div>
+        <div class="icon-button" onclick="openModal('aboutModal')">?</div>
+    </div>
+    <div id="suggestionsModal" class="modal">
+        <div class="modal-content">
+            <span class="close" onclick="closeModal('suggestionsModal')">&times;</span>
+            <h3>Suggestions</h3>
+            <ul>
+                <li>Try: <code>paris - france + italy</code></li>
+                <li>Try: <code>man + smart</code></li>
+                <li>Use <code>-</code> and <code>+</code> operators</li>
+            </ul>
+        </div>
+    </div>
+    <div id="aboutModal" class="modal">
+        <div class="modal-content">
+            <span class="close" onclick="closeModal('aboutModal')">&times;</span>
+            <h3>About</h3>
+            <p>This tool calculates vector arithmetic of words using a pretrained CBOW model "glove-twitter-200". Built by Clement Ha.</p>
+        </div>
+    </div>
+    <script>
+        function openModal(id) {
+            document.getElementById(id).style.display = 'block';
+        }
+        function closeModal(id) {
+            document.getElementById(id).style.display = 'none';
+        }
+        window.onclick = function(event) {
+            const modals = document.querySelectorAll('.modal');
+            modals.forEach(modal => {
+                if (event.target == modal) {
+                    modal.style.display = "none";
+                }
+            });
+        }
+    </script>
+    <script>
+        document.addEventListener("DOMContentLoaded", function () {
+            const textarea = document.querySelector("textarea[name='expression']");
+            const form = document.querySelector("form");
+            textarea.addEventListener("keydown", function (event) {
+                if (event.key === "Enter" && !event.shiftKey) {
+                    event.preventDefault(); // prevent newline
+                    form.submit();          // trigger form submission
+                }
+            });
+        });
+    </script>
+    <script>
+        document.addEventListener("DOMContentLoaded", function () {
+            const form = document.querySelector("form");
+            const textarea = document.querySelector("textarea[name='expression']");
+            let resultShown = {{ 'true' if results else 'false' }};
+            // 1. Validate spacing between tokens
+            form.addEventListener("submit", function (event) {
+                const input = textarea.value.trim();
+                // Simple regex to find missing spaces (e.g., "word+word", "word-word")
+                const spacingIssues = input.match(/\b\w+[\+\-]\w+\b/);
+                if (spacingIssues) {
+                    event.preventDefault();
+                    const problem = spacingIssues[0];
+                    const suggestion = problem.replace(/([\+\-])/, ' $1 ');
+                    alert(`⚠️ It looks like you missed spacing in: "${problem}".\nDid you mean: "${suggestion}"?`);
+                }
+            });
+            // 2. Clear textarea when focused again, only if result was shown
+            if (resultShown) {
+                textarea.addEventListener("focus", () => {
+                    textarea.value = "";
+                    resultShown = false;  // prevent it from clearing again on next focus
+                });
+            }
+        });
+    </script>
+</body>
+</html>

vit_captioning/generate.py CHANGED Viewed

@@ -10,7 +10,7 @@ import argparse
 class CaptionGenerator:
-    def __init__(self, model_type: str, checkpoint_path: str, quantized=False):
         print(f"Loading {model_type} | Quantized: {quantized}")
         # Setup device
         if torch.cuda.is_available():
@@ -25,9 +25,10 @@ class CaptionGenerator:
         # Load tokenizer
         #self.tokenizer = AutoTokenizer.from_pretrained('bert-base-uncased')
-        #HF needs all model downloads to a special read-write cache dir
-        #self.tokenizer = AutoTokenizer.from_pretrained('bert-base-uncased', cache_dir="/tmp")
-        self.tokenizer = AutoTokenizer.from_pretrained('/models/bert-tokenizer')
         # Select encoder, processor, output dim
         if model_type == "ViTEncoder":
@@ -37,8 +38,11 @@ class CaptionGenerator:
         elif model_type == "CLIPEncoder":
             self.encoder = CLIPEncoder().to(self.device)
             self.encoder_dim = 512
-            #self.processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
-            self.processor = CLIPProcessor.from_pretrained("/models/clip")
         else:
             raise ValueError("Unknown model type")
@@ -109,7 +113,8 @@ if __name__ == "__main__":
     generator = CaptionGenerator(
         model_type=args.model,
-        checkpoint_path=args.checkpoint
     )
     captions = generator.generate_caption(args.image)

 class CaptionGenerator:
+    def __init__(self, model_type: str, checkpoint_path: str, quantized=False, runAsContainer=False):
         print(f"Loading {model_type} | Quantized: {quantized}")
         # Setup device
         if torch.cuda.is_available():
         # Load tokenizer
         #self.tokenizer = AutoTokenizer.from_pretrained('bert-base-uncased')
+        if (runAsContainer):
+            self.tokenizer = AutoTokenizer.from_pretrained('/models/bert-tokenizer')
+        else:
+            self.tokenizer = AutoTokenizer.from_pretrained('bert-base-uncased')
         # Select encoder, processor, output dim
         if model_type == "ViTEncoder":
         elif model_type == "CLIPEncoder":
             self.encoder = CLIPEncoder().to(self.device)
             self.encoder_dim = 512
+            if (runAsContainer):
+                self.processor = CLIPProcessor.from_pretrained("/models/clip")
+            else:
+                self.processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
         else:
             raise ValueError("Unknown model type")
     generator = CaptionGenerator(
         model_type=args.model,
+        checkpoint_path=args.checkpoint,
+        runAsContainer=True
     )
     captions = generator.generate_caption(args.image)

vit_captioning/models/encoder.py CHANGED Viewed

@@ -13,9 +13,7 @@ class ViTEncoder(nn.Module):
         #weights = ViT_B_16_Weights.DEFAULT
-        #self.vit = ViTModel.from_pretrained('google/vit-base-patch16-224-in21k')
-        #HF needs all model downloads to a special read-write cache dir
-        self.vit = ViTModel.from_pretrained('google/vit-base-patch16-224-in21k', cache_dir="/tmp")
     def forward(self, pixel_values):
@@ -34,10 +32,7 @@ class CLIPEncoder(nn.Module):
     def __init__(self):
         super(CLIPEncoder, self).__init__()
         #self.clip = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
-        #HF needs all model downloads to a special read-write cache dir
-        #self.clip = CLIPModel.from_pretrained("openai/clip-vit-base-patch32", cache_dir="/tmp")
         self.clip = CLIPModel.from_pretrained("/models/clip")
     def forward(self, pixel_values):
         # ✅ Directly get the pooled image features (already the final representation)
         image_features = self.clip.get_image_features(pixel_values=pixel_values)

         #weights = ViT_B_16_Weights.DEFAULT
+        self.vit = ViTModel.from_pretrained('google/vit-base-patch16-224-in21k')
     def forward(self, pixel_values):
     def __init__(self):
         super(CLIPEncoder, self).__init__()
         #self.clip = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
         self.clip = CLIPModel.from_pretrained("/models/clip")
     def forward(self, pixel_values):
         # ✅ Directly get the pooled image features (already the final representation)
         image_features = self.clip.get_image_features(pixel_values=pixel_values)

vit_captioning/static/landing.html CHANGED Viewed

@@ -13,11 +13,11 @@
       <a href="/captioning" class="block w-full bg-blue-600 hover:bg-blue-700 text-white py-3 rounded-lg shadow text-lg font-semibold">
         🖼️ Image Captioning
       </a>
-      <a href="/project2" class="block w-full bg-green-600 hover:bg-green-700 text-white py-3 rounded-lg shadow text-lg font-semibold">
-        🤖 Coming Soon: Word meaning calculator
       </a>
-      <a href="/project3" class="block w-full bg-green-600 hover:bg-green-700 text-white py-3 rounded-lg shadow text-lg font-semibold">
-        📚 Coming Soon: Text Summerizer
       </a>
       <!-- Add more project links here -->
     </div>

       <a href="/captioning" class="block w-full bg-blue-600 hover:bg-blue-700 text-white py-3 rounded-lg shadow text-lg font-semibold">
         🖼️ Image Captioning
       </a>
+      <a href="/cbow" class="block w-full bg-green-600 hover:bg-green-700 text-white py-3 rounded-lg shadow text-lg font-semibold">
+        🧙‍♂️ Word Alchemy
       </a>
+      <a href="/contentdistillery" class="block w-full bg-red-600 hover:bg-red-700 text-white py-3 rounded-lg shadow text-lg font-semibold">
+        📚 Content Distillery
       </a>
       <!-- Add more project links here -->
     </div>