Update README.md

README.md

@@ -3,7 +3,7 @@ license: apache-2.0
 ---
 # Instinct-1-1B
 
-*Instinct-1-1B is a fully reproducible, from-scratch trained 1B parameter language model trained on 85B tokens using TPU v4 infrastructure.*
+*Instinct-1-1B is a fully reproducible, from-scratch trained 1B parameter language model trained on 20B tokens of PILE using TPU v4 infrastructure.*
 
 **Instinct-1-1B** is a 1 Billion parameter Large Language Model built from scratch under the **AutonomousX** organization.
 
@@ -27,6 +27,8 @@ This model was developed by **Rohit Yadav**, a **B.Tech 3rd year student from NI
 
 Validation was performed using **rolling validation shards of the dataset**.
 
+![image](https://cdn-uploads.huggingface.co/production/uploads/68bf07a31d80a360f1405b72/ICT8s2ycXLVz9MLgc9iBD.png)
+
 ---
 
 ## Architecture Details
@@ -55,7 +57,7 @@ Training pipeline includes:
 * Checkpointing and validation during training
 * Rolling validation shard evaluation
 
-The model was trained on **85B tokens**.
+The model was trained on **20B tokens** and it's is a checkpoint of final version trained on **85B tokens** in total
 
 ---
 
@@ -250,40 +252,122 @@ output_ids = generate(params, input_ids, 200)
 print("\n=== GENERATED TEXT ===\n")
 print(tokenizer.decode(output_ids[0].tolist()))
 ```
-
-</div>
-
----
-
-## Author
-
-**Rohit Yadav**
-
-B.Tech 3rd Year  
+<div id="autonomousx-profile-sections" style="font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Helvetica, Arial, sans-serif; width: 100%; display: flex; flex-direction: column; gap: 24px; margin: 30px 0;">
+
+<style>
+#autonomousx-profile-sections * { box-sizing: border-box; }
+
+.ax-card {
+position: relative;
+overflow: hidden;
+width: 100%;
+padding: 25px;
+border-radius: 12px;
+border: 1px solid #e5e7eb;
+background: #ffffff;
+box-shadow: 0 4px 20px rgba(0, 0, 0, 0.04);
+text-align: left;
+}
+
+@keyframes diagonalShimmer {
+0% { transform: translateX(-150%) skewX(-15deg); }
+50% { transform: translateX(150%) skewX(-15deg); }
+100% { transform: translateX(150%) skewX(-15deg); }
+}
+
+.ax-card::before {
+content: "";
+position: absolute;
+top: 0;
+left: 0;
+width: 100%;
+height: 100%;
+background: linear-gradient(90deg, rgba(255, 255, 255, 0) 0%, rgba(139, 92, 246, 0.05) 40%, rgba(236, 72, 153, 0.1) 50%, rgba(139, 92, 246, 0.05) 60%, rgba(255, 255, 255, 0) 100%);
+animation: diagonalShimmer 5s infinite ease-in-out;
+pointer-events: none;
+z-index: 1;
+}
+
+.ax-card-content { position: relative; z-index: 2; }
+
+.ax-card h1 { margin: 0 0 4px 0; font-size: 14px; color: #6b7280; font-weight: 700; text-transform: uppercase; letter-spacing: 1.5px; }
+.ax-card h2 { margin: 0 0 16px 0; font-size: 32px; color: #8b5cf6; font-weight: 800; letter-spacing: -0.5px; }
+
+/* The new dark interior box */
+.ax-dark-box {
+background: #0f172a;
+color: #e2e8f0;
+padding: 20px;
+border-radius: 8px;
+margin-top: 15px;
+border: 1px solid #1e293b;
+}
+
+.ax-dark-box p { margin: 0 0 12px 0; font-size: 15px; line-height: 1.6; }
+.ax-dark-box a { color: #a78bfa; text-decoration: none; font-weight: 600; }
+.ax-dark-box a:hover { color: #d946ef; text-decoration: underline; }
+
+.ax-icon { margin-right: 8px; font-style: normal; }
+.ax-highlight-text { color: #f472b6; font-weight: 700; }
+
+.ax-badges { display: flex; flex-wrap: wrap; gap: 8px; margin-top: 12px; }
+.ax-badge { background: #1e293b; color: #cbd5e1; padding: 6px 12px; border-radius: 20px; font-size: 13px; font-weight: 600; border: 1px solid #334155; }
+</style>
+
+<div class="ax-card">
+<div class="ax-card-content">
+<h1>Author</h1>
+<h2>Rohit Yadav</h2>
+
+<div class="ax-dark-box">
+<p>
+<strong>B.Tech 3rd Year</strong><br>
 Dr. B.R. Ambedkar National Institute of Technology (NIT) Jalandhar, India
+</p>
+
+<p>
+<span class="ax-icon">📧</span> E-mail: <a href="mailto:yrohit1825@gmail.com">yrohit1825@gmail.com</a><br>
+<span class="ax-icon">🔗</span> LinkedIn: <a href="https://www.linkedin.com/in/rohit-yadav-25535b256/" target="_blank">Rohit Yadav</a><br>
+<span class="ax-icon">💻</span> Github: <a href="https://github.com/YADAV1825" target="_blank">YADAV1825</a>
+</p>
+
+<p class="ax-highlight-text" style="margin-top: 16px;">
+🚀 I am actively seeking Internships and Collaborations!
+</p>
+
+<div style="margin-top: 20px;">
+<h3 style="font-size: 13px; color: #94a3b8; text-transform: uppercase; letter-spacing: 1px; margin: 0 0 10px 0;">Research Interests</h3>
+<div class="ax-badges">
+<span class="ax-badge">Large Language Models</span>
+<span class="ax-badge">MultiModal Pipelines</span>
+<span class="ax-badge">Systems Programming</span>
+<span class="ax-badge">AI Infrastructure</span>
+<span class="ax-badge">Distributed Training</span>
+</div>
+</div>
+</div>
+</div>
+</div>
 
-*E-mail: yrohit1825@gmail.com*
-
-*LinkedIN: https://www.linkedin.com/in/rohit-yadav-25535b256/*
-
-*Github: https://github.com/YADAV1825*
-
-Research interests include:
-
-* Large Language Models
-* MultiModal Pipelines
-* Systems Programming
-* AI Infrastructure
-* Distributed Training
-
----
-
-## Organization
+<div class="ax-card">
+<div class="ax-card-content">
+<h1>Organization</h1>
+<h2 style="color: #3b82f6;">AutonomousX</h2>
 
-**AutonomousX**
+<div class="ax-dark-box">
+<p>
+<strong>AutonomousX</strong> focuses on open-source contributions aimed at building Large Language Models from scratch using custom training pipelines.
+</p>
 
-AutonomousX focuses on open-source contributions aimed at building Large Language Models from scratch using custom training pipelines.
+<p>
+Our work explores different training configurations including optimizers, datasets, and scalable TPU training using <strong>JAX and pmap</strong>. The goal is to provide transparent and reproducible implementations so that researchers, students, and developers can understand how modern LLMs are trained end-to-end.
+</p>
 
-Our work explores different training configurations including optimizers, datasets, and scalable TPU training using **JAX and pmap**. The goal is to provide transparent and reproducible implementations so that researchers, students, and developers can understand how modern LLMs are trained end-to-end.
+<p style="margin-bottom: 0;">
+Due to the current scarcity of complete beginner-friendly guides for training LLMs on TPUs, especially using JAX, AutonomousX aims to bridge this gap by publishing full training pipelines, scripts, and documentation for the open-source community.
+</p>
+</div>
+</div>
+</div>
 
-Due to the current scarcity of complete beginner-friendly guides for training LLMs on TPUs, especially using JAX, AutonomousX aims to bridge this gap by publishing full training pipelines, scripts, and documentation for the open-source community.
+</div>