Add files using upload-large-folder tool

README.md

@@ -1,294 +1,31 @@
 ---
-language:
-- hi
-- en
-language_bcp47:
-- en-hi
+language: en
+library_name: mlx
 pipeline_tag: text-generation
 tags:
-- hinglish
-- hindi
-- indian-languages
-- edge-ai
-- mac-mini
-- lora
-- qwen2.5
-- small-language-model
-- on-device-ai
-- apple-silicon
 - mlx
-- fine-tuned-llm
-- lightweight
-- edge-deployment
-- low-resource
-- nlp
-- conversation
-- india
-- bharat
-- tiny-llm
-- hinglish-chat
-license: apache-2.0
-library_name: mlx
-model-index:
-- name: Bharat-Tiny-LLM
-  results: []
-datasets:
-- ankitdhiman/hinglish-conversations
-- maya-research/IndicVault
-- Sujalvc/hinglish-instruct-dataset
-- Subh24ai/yojana-sahayak-instruct
-- VishalMysore/cookGPT
-inference:
-  parameters:
-    temperature: 0.7
-    max_tokens: 200
-widget:
-- text: "Chai peete hain?"
-  example_title: Chai
-- text: "Kal interview hai, nervous ho raha hoon"
-  example_title: Interview Nerves
-- text: "Biryani kaise banate hain?"
-  example_title: Biryani Recipe
-- text: "Delhi me rehne ke liye kya karna padega?"
-  example_title: Delhi Living
-- text: "Mujhe Hindi seekhni hai"
-  example_title: Learn Hindi
----
-
-# Bharat-Tiny-LLM 🇮🇳
-
-**Hinglish AI — trained on a Mac Mini M4, runs on ₹8,000 phones.**
-
-[![Demo](https://img.shields.io/badge/🤗%20Try%20the%20Demo-Live-FF6B6B?style=flat)](https://huggingface.co/spaces/eulogik/Bharat-Tiny-LLM)
-![Mac Mini](https://img.shields.io/badge/Trained%20on-Mac%20Mini%20M4-0071CD?style=flat&logo=apple&logoColor=white)
-![Size](https://img.shields.io/badge/Model%20Size-828%20MB-00B4D8?style=flat)
-![Speed](https://img.shields.io/badge/Inference~57%20tok/s-00E676?style=flat)
-![License](https://img.shields.io/badge/License-Apache%202.0-green?style=flat)
-![Cloud Cost](https://img.shields.io/badge/Cloud%20Cost-%240-orange?style=flat)
-![Val Loss](https://img.shields.io/badge/Val%20Loss-0.781-FF6B6B?style=flat)
-
----
-
-## The Story
-
-**I trained an AI to speak Hinglish on a Mac Mini. No cloud compute. No GPU cluster. Just $750 of Apple Silicon and a lot of chai.**
-
-This is **Bharat-Tiny-LLM** — a Qwen2.5-1.5B model fine-tuned on 376K Hinglish conversations. Quantized to **828 MB** (Q4), it delivers **~57 tokens/second** on a Mac Mini M4 and is small enough to run on an ₹8,000 smartphone.
-
-Most AI models today are skyscrapers — built by billion-dollar companies, requiring data centers, and serving only English. **We built a bicycle.** A tiny, efficient, open-source model that speaks the way 600 million Indians actually talk: Hinglish.
-
-### What Makes This Different
-
-| | Big AI | Bharat-Tiny-LLM |
-|---|---|---|
-| **Hardware** | $100M GPU clusters | Mac Mini M4 (16GB) |
-| **Cloud Cost** | Millions | **$0** |
-| **Language** | English only | Hinglish (हिंग्लिश) |
-| **Model Size** | 70B–405B params | **1.5B** (828 MB Q4) |
-| **Device** | Data centers | **₹8,000 phones** |
-| **Speed** | API latency | **57 tok/s on-device** |
-| **License** | Proprietary/Research | **Apache 2.0** |
-
----
-
-## Model Details
-
-### Architecture
-
-- **Base Model:** [Qwen2.5-1.5B](https://huggingface.co/Qwen/Qwen2.5-1.5B) (Apache 2.0)
-- **Fine-Tuning Method:** LoRA (16 layers, rank 8, alpha 16)
-- **Training Hardware:** Mac Mini M4, 16GB unified memory
-- **Training Duration:** ~3.3 days for 76K iterations
-- **Quantization:** 4-bit (Q4) via MLX — 4.501 bits/weight
-- **Total Trainable Parameters:** 5.276M (0.342% of base)
-
-<details>
-<summary>Base model details (click to expand)</summary>
-
-This model is fine-tuned from [Qwen2.5-1.5B](https://huggingface.co/Qwen/Qwen2.5-1.5B) (Apache 2.0), one of the few small models with native multilingual support (29 languages including Hindi). Unlike English-only models (like SmolLM2, which we tried first — val loss plateaued at 1.32), Qwen2.5 already understands Hindi morphology. LoRA fine-tuning on Hinglish data was enough to make it conversational.
-
-</details>
-
-### Performance
-
-| Metric | Value |
-|--------|-------|
-| Training Data | 376,040 conversations |
-| Validation Loss | **0.781** (best at 75K iters) |
-| Inference Speed | **~57 tok/s** (Mac Mini M4) |
-| Model Size | **828 MB** (Q4 quantized) |
-| Memory Usage | ~1.2 GB at inference |
-| Context Length | 512 tokens |
-
 ---
 
-## Quick Start
+# eulogik/Bharat-Tiny-LLM
 
-### Using mlx-lm (macOS / Apple Silicon)
+## Use with mlx
 
 ```bash
 pip install mlx-lm
-
-python3 -c "
-from mlx_lm import load, generate
-from mlx_lm.sample_utils import make_sampler
-
-model, tokenizer = load('eulogik/Bharat-Tiny-LLM')
-sampler = make_sampler(temp=0.7)
-
-prompt = '<|im_start|>user\nChai peete hain?<|im_end|>\n<|im_start|>assistant\n'
-response = generate(model, tokenizer, prompt=prompt, max_tokens=200, sampler=sampler)
-print(response)
-"
 ```
 
-### Platform Support
-
-| Platform | Status | Instructions |
-|----------|--------|-------------|
-| macOS (Apple Silicon) | ✅ Works | Use `mlx-lm` (above) |
-| Linux / Windows | 🔄 Coming | Need transformers-compatible bf16 export |
-| Android | 🔄 Coming | MLX Android bindings in development |
-| Raspberry Pi 5 | 🔄 Coming | Testing in progress |
-
----
-
-## Example Outputs
-
-| Prompt | Response | Quality |
-|--------|----------|---------|
-| "Kal interview hai, nervous ho raha hoon" | "Chill maaro, you'll be alright. Bas apne aap ko confident rakhna aur jo aata hai woh bol dena. Sab theek hoga." | ✅ |
-| "Biryani kaise banate hain?" | Gives structured recipe with ingredients list and step-by-step instructions in Hinglish | ✅ |
-| "Delhi me rehne ke liye kya karna padega?" | Specific advice on metro, parking, traffic, and areas to live | ✅ |
-| "Weekend pe kya karein?" | 10 ideas with examples — museums, parks, street food, movies | ✅ |
-| "Mujhe Hindi seekhni hai" | Practical listening tips, YouTube channels, and practice strategies | ✅ |
-| "Chai peetey hain?" | Full chai recipe with proportions and tips | ✅ |
-| "Mera phone charge nahi ho raha" | Troubleshooting steps (though slightly repetitive) | ⚠️ |
-| "Job nahi mil rahi" | Can get confused — still improving on emotional/career topics | ❌ |
-
----
-
-## Training Data
-
-We curated **376,040 conversations** from 5 open-source datasets, deduplicated and formatted as chat messages:
-
-| Dataset | Samples | Domain |
-|---------|---------|--------|
-| [Hinglish Conversations](https://huggingface.co/datasets/ankitdhiman/hinglish-conversations) | 201,633 | Natural Hinglish dialogue |
-| [IndicVault (Hindi)](https://huggingface.co/datasets/maya-research/IndicVault) | 74,053 | 20-topic Hindi QA |
-| [IndicVault (Hinglish)](https://huggingface.co/datasets/maya-research/IndicVault) | 77,210 | 20-topic Hinglish QA |
-| [Hinglish Instruct](https://huggingface.co/datasets/Sujalvc/hinglish-instruct-dataset) | 10,378 | Instruction tuning |
-| [Yojana Sahayak](https://huggingface.co/datasets/Subh24ai/yojana-sahayak-instruct) | 6,828 | Government scheme info |
-| [cookGPT](https://huggingface.co/datasets/VishalMysore/cookGPT) | 5,938 | Indian recipes |
-
-All data was:
-- Trimmed to 512 characters per message (to fit within model context)
-- Converted to `messages` format with `user`/`assistant` roles
-- Split 95% train / 5% validation
-
----
-
-## Training Details
-
-### Hyperparameters
-
-```
-Base model:    Qwen2.5-1.5B (bf16 from mlx-community)
-Method:        LoRA (16 layers, rank 8, alpha 16)
-Batch size:    4
-Seq length:    512
-Learning rate: 5e-5 (with cosine schedule)
-Iterations:    76,420
-Optimizer:     AdamW
-Warmup:        100 steps
-Seed:          42
-```
-
-### Training Curve
-
-```
-Val Loss
-1.25 |  █
-1.20 |  ██
-1.15 |  ███
-1.10 |  ████
-1.05 |  █████
-1.00 |  ██████
-0.95 |  ████████
-0.90 |  ██████████
-0.85 |  ████████████
-0.80 |  ██████████████
-0.78 |  ████████████████  ← Best at 75K iters
-     +------------------------
-     0   20K   40K   60K   80K
-```
-
-### Speed Optimization
-
-| Config | it/s | Memory | Notes |
-|--------|------|--------|-------|
-| batch=2, seq=1024 | 0.30 | 10.7 GB | Initial config |
-| batch=8, seq=256 | 0.24 | 10.0 GB | Too much overhead |
-| **batch=4, seq=512** | **0.30** | **10.3 GB** | **Final — best balance** |
-| batch=4, seq=1024 | OOM | — | Exceeded Metal GPU |
-
----
-
-## Limitations
-
-1. **Not fully converged** — Only 76K iterations; val loss was still decreasing at 0.781. More training (target: 110K) would help.
-2. **Occasional English drift** — Some responses switch entirely to English. Needs more Hinglish-only examples.
-3. **Telugu leakage** — IndicVault Hindi data contains some Telugu examples that leak into responses.
-4. **Emotional/career topics** — Responses to queries about jobs, relationships, or mental health can be confused or repetitive.
-5. **Short context** — 512 token limit prevents multi-turn conversations or long-form generation.
-6. **MLX format only** — The Q4 quantized model is in Apple MLX format. We plan to release a transformers-compatible version.
-7. **Not instruction-tuned** — This is a base model fine-tuned on conversations, not RLHF/DPO-tuned. Responses are helpful but not aligned.
-
----
-
-## Roadmap
-
-- [x] **Phase 0:** Fine-tune Qwen2.5-1.5B with LoRA on Hinglish data
-- [x] **Phase 1:** Quantize to Q4 (828 MB) for edge deployment
-- [ ] **Phase 2:** Extend to 8 Indian languages (Tamil, Telugu, Bengali, Marathi, etc.)
-- [ ] **Phase 3:** DPO/RLHF for quality improvement
-- [ ] **Phase 4:** BharatTiny-Bench evaluation suite
-- [ ] **Phase 5:** Web demo with Gradio
-- [ ] **Phase 6:** Android / Raspberry Pi 5 deployment
-
----
-
-## Why "Bharat-Tiny-LLM"?
-
-600 million Indians speak Hinglish — a natural mix of Hindi and English. But every LLM today is either English-only or pure Hindi. There is no model that understands "Yaar, kal ka match dekh liya? Rohit ne toh maara!"
-
-We built this because:
-- **No one builds for edge devices in India.** Everyone targets data centers.
-- **No one speaks Hinglish.** Everyone speaks either English or "pure" Hindi.
-- **No one does it on a budget.** Everyone burns millions in cloud compute.
+```python
+from mlx_lm import load, generate
 
-**We're building the bicycle while everyone else builds skyscrapers.**
+model, tokenizer = load("eulogik/Bharat-Tiny-LLM")
 
----
+prompt = "hello"
 
-## Citation
+if tokenizer.chat_template is not None:
+    messages = [{"role": "user", "content": prompt}]
+    prompt = tokenizer.apply_chat_template(
+        messages, add_generation_prompt=True, return_dict=False,
+    )
 
-```bibtex
-@misc{bharat-tiny-llm-2026,
-  title = {Bharat-Tiny-LLM: Hinglish Edge AI Trained on a Mac Mini},
-  author = {Gautam Kishore},
-  year = {2026},
-  publisher = {Hugging Face},
-  url = {https://huggingface.co/eulogik/Bharat-Tiny-LLM}
-}
+response = generate(model, tokenizer, prompt=prompt, verbose=True)
 ```
-
----
-
-## License
-
-Apache 2.0 — Free for commercial and research use. The base model (Qwen2.5-1.5B) is also Apache 2.0.
-
----
-
-*Built with chai ☕, patience 🧘, and a Mac Mini M4 💻 — somewhere in India.*

chat_template.jinja

@@ -10,7 +10,7 @@
         {{- "\n" }}
         {{- tool | tojson }}
     {%- endfor %}
-    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{{\"name\": <function-name>, \"arguments\": <args-json-object>}}\n</tool_call><|im_end|>\n" }}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
 {%- else %}
     {%- if messages[0]['role'] == 'system' %}
         {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}

config.json

@@ -4,17 +4,49 @@
     ],
     "attention_dropout": 0.0,
     "bos_token_id": 151643,
+    "dtype": "float16",
     "eos_token_id": 151643,
     "hidden_act": "silu",
     "hidden_size": 1536,
     "initializer_range": 0.02,
     "intermediate_size": 8960,
-    "max_position_embeddings": 32768,
+    "layer_types": [
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention"
+    ],
+    "max_position_embeddings": 131072,
     "max_window_layers": 28,
     "model_type": "qwen2",
     "num_attention_heads": 12,
     "num_hidden_layers": 28,
     "num_key_value_heads": 2,
+    "pad_token_id": null,
     "quantization": {
         "group_size": 64,
         "bits": 4,
@@ -26,11 +58,13 @@
         "mode": "affine"
     },
     "rms_norm_eps": 1e-06,
-    "rope_theta": 1000000.0,
-    "sliding_window": 32768,
+    "rope_parameters": {
+        "rope_theta": 1000000.0,
+        "rope_type": "default"
+    },
+    "sliding_window": null,
     "tie_word_embeddings": true,
-    "torch_dtype": "bfloat16",
-    "transformers_version": "4.40.1",
+    "transformers_version": "5.12.1",
     "use_cache": true,
     "use_mrope": false,
     "use_sliding_window": false,

generation_config.json

@@ -0,0 +1,5 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "pad_token_id": 151643
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0755fa9cc496b94aaebca0a1cca1b38ed311eefc2c60c4a68f7353035565fcb3
-size 868629082
+oid sha256:33a771823bb4d0fd013e0b6c64b53147c9014f4c9aa7a86a4a458acc205225a8
+size 868628547

tokenizer_config.json

@@ -3,7 +3,7 @@
   "backend": "tokenizers",
   "bos_token": null,
   "clean_up_tokenization_spaces": false,
-  "eos_token": "<|im_end|>",
+  "eos_token": "<|endoftext|>",
   "errors": "replace",
   "extra_special_tokens": [
     "<|im_start|>",