Initial OMDA model upload

README.md

@@ -0,0 +1,37 @@
+
+# OMDA: Arabic-English Chat LLM
+
+**Model Name:** OMDA  
+**Architecture:** OMDA-Decoder   
+**Tokenizer:** OMDATokenizer  
+**Languages:** Arabic, English  
+**Type:** Chat/Instruction-following  
+**Author:** Binomda  
+**Date:** 2025-06-28
+
+## Model Details
+- Layers: 6
+- Hidden size: 512
+- Attention heads: 8
+- FFN dim: 2048
+- Max sequence length: 512
+- Vocab size: 128004
+- Training data: Aggregated Arabic-English chat pairs from CSV, TXT, and JSON sources.
+
+## Intended Use
+- Chatbots, assistants, translation, and educational tools for Arabic/English.
+
+## Training
+- Trained for 5 epochs on 1000 samples.
+- Loss curve and checkpoints included.
+
+## Limitations
+- This is a small-scale demonstration model and may not generalize well to all real-world chat scenarios.
+- Not suitable for production use without further scaling, extensive evaluation, and safety checks.
+- Limited training data and model size may result in hallucinations or inaccurate translations.
+- No advanced filtering for inappropriate or biased outputs.
+- For research and educational purposes only.
+
+## Export & Deployment
+- See below for HuggingFace, llama.cpp, and ollama export instructions.
+

config.json

@@ -0,0 +1,18 @@
+{
+  "model_name": "OMDA",
+  "architecture": "OMDA-Decoder",
+  "vocab_size": 128004,
+  "d_model": 512,
+  "n_layers": 6,
+  "n_heads": 8,
+  "d_ff": 2048,
+  "max_seq_len": 512,
+  "dropout": 0.1,
+  "batch_size": 8,
+  "learning_rate": 0.0001,
+  "num_epochs": 5,
+  "save_steps": 1000,
+  "eval_steps": 500,
+  "model_save_path": "./omda_chat_model",
+  "tokenizer_name": "OMDATokenizer"
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad51ed28b2c059ce3a0e44cbdfcbda1978b8993d50e90a1cead97ec07b7cbde9
+size 601553329