Update model card

README.md

@@ -0,0 +1,226 @@
+---
+license: apache-2.0
+language:
+  - en
+base_model: facebook/MobileLLM-R1-950M
+datasets:
+  - Ayansk11/FinSenti-Dataset
+pipeline_tag: text-generation
+library_name: transformers
+tags:
+  - finance
+  - financial-sentiment
+  - sentiment-analysis
+  - chain-of-thought
+  - reasoning
+  - grpo
+  - sft
+  - lora
+  - finsenti
+---
+# FinSenti-MobileLLM-R1-950M
+
+FinSenti-MobileLLM-R1-950M is a 0.9B-parameter model fine-tuned to
+read short financial text (headlines, earnings snippets, market commentary)
+and explain its read of them before settling on positive, negative, or
+neutral. It's Meta's purpose-built mobile model. The architecture is shaped for on-device inference (compact embeddings, untied lm_head, shared attention layers) and FinSenti's recipe lifts the financial-sentiment quality without changing that footprint.
+
+The model is part of the [FinSenti
+collection](https://huggingface.co/collections/Ayansk11/finsenti), a
+scaling study of small models trained on the same data with the same recipe.
+
+## What it's good at
+
+- Classifying short financial text (1-3 sentences) into positive / negative
+  / neutral
+- Producing a short reasoning chain you can read or log
+- Following a strict `<reasoning>...</reasoning><answer>...</answer>` output
+  format that's easy to parse downstream
+
+It was trained on news-style headlines and earnings snippets in English, so
+that's where it shines. Outside that domain you'll see the format hold up
+but the labels get noisier.
+
+## How it was trained
+
+Two-stage recipe, same across the whole FinSenti family:
+
+1. **SFT** on the SFT train slice from the [FinSenti
+   dataset](https://huggingface.co/datasets/Ayansk11/FinSenti-Dataset)
+   (~15.2K balanced training samples, drawn from a
+   50.8K-sample pool with held-out val/test splits, chain-of-thought
+   targets generated by a teacher model and filtered for label agreement).
+   This stage took about 1.6 hours on a single A100 80GB
+   for this model.
+2. **GRPO** with four reward functions (sentiment correctness, format
+   compliance, reasoning quality, output consistency), each weighted equally
+   for a maximum reward of 4.0. The training budget was 3000
+   steps with early stopping; the best checkpoint landed near step
+   ~200 with a mean reward of approximately
+   **3.29 / 4.0** on the validation slice.
+
+Trainer stack: PEFT + bitsandbytes (no Unsloth - llama4_text arch unsupported), using Unsloth's pre-quantized mirror
+[`facebook/MobileLLM-R1-950M`](https://huggingface.co/facebook/MobileLLM-R1-950M) as the
+loading shortcut for the upstream
+[`facebook/MobileLLM-R1-950M`](https://huggingface.co/facebook/MobileLLM-R1-950M)
+weights. LoRA adapters (r=16, alpha=32) were
+trained on the attention and MLP projection layers, then merged into the
+base weights before export, so this repo is a self-contained model and
+doesn't need PEFT to load.
+
+## Quick start
+
+Standard `transformers` usage:
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+
+model_id = "Ayansk11/FinSenti-MobileLLM-R1-950M"
+tok = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id, torch_dtype=torch.bfloat16, device_map="auto"
+)
+
+system = (
+    "You are a financial sentiment analyst. For each headline you receive, "
+    "write a short reasoning chain inside <reasoning>...</reasoning> tags, "
+    "then give a single label inside <answer>...</answer> tags. The label "
+    "must be exactly one of: positive, negative, neutral."
+)
+user = "Apple beats Q4 estimates as iPhone sales jump 12% year over year."
+
+messages = [
+    {"role": "system", "content": system},
+    {"role": "user", "content": user},
+]
+prompt = tok.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+
+inputs = tok(prompt, return_tensors="pt").to(model.device)
+out = model.generate(**inputs, max_new_tokens=256, do_sample=False)
+print(tok.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))
+```
+
+Expected output (your reasoning text will vary; the label should match):
+
+```
+<reasoning>
+Beating estimates is a positive earnings surprise. A 12% YoY iPhone sales jump in the company's biggest product line points to demand strength. Both signals push the read positive.
+</reasoning>
+<answer>positive</answer>
+```
+
+## Prompt format
+
+The model expects the system prompt above, verbatim is best. The user turn
+is the headline or short snippet you want classified. Output is two XML-ish
+blocks in this order: `<reasoning>...</reasoning>` then
+`<answer>...</answer>`. The `<answer>` content is one of `positive`,
+`negative`, or `neutral` (lowercase, no punctuation).
+
+If you want labels only and don't care about the reasoning, you can stop
+generation as soon as you see `</answer>` to save tokens.
+
+## Performance notes
+
+The training reward (max 4.0) hit **3.29** on the
+held-out validation slice. That breaks down across the four reward
+functions roughly as:
+
+- Sentiment correctness: dominant contributor; the model gets the label
+  right on the validation split most of the time
+- Format compliance: near-saturated by the end of GRPO; the model almost
+  always produces well-formed `<reasoning>` and `<answer>` tags
+- Reasoning quality: judged on length and presence of finance-relevant
+  signal words; this one's the noisiest of the four
+- Consistency: rewards stable labels across paraphrases of the same headline
+
+Numbers on standard finance benchmarks (FPB, FiQA, Twitter Financial News)
+are forthcoming and will be added once the eval pipeline lands.
+
+## Hardware
+
+At bf16 the weights are about 1.8 GB on disk and need ~3 GB of GPU memory for batch=1 inference. CPU inference is fine too: on a modern laptop you'll get a few tokens per second with the bf16 weights, and 15-30 tok/s with the GGUF Q4_K_M build.
+
+## Limitations
+
+A few things this model isn't built for:
+
+- **Long documents.** Training context was capped at 2048
+  tokens. Anything much longer than a few paragraphs is out of distribution.
+- **Multi-asset reasoning.** It classifies the sentiment of a single piece
+  of text. It won't aggregate across multiple headlines or weigh sources.
+- **Numerical reasoning.** It can read "beats by 12%" and call that
+  positive, but it isn't doing math. Don't ask it to forecast.
+- **Languages other than English.** Training data was English only.
+- **Background knowledge.** If the headline needs you to know what a
+  company does, the model only has whatever was in its base pretraining.
+  It can't look anything up.
+- **Three labels, hard cutoffs.** The output space is positive / negative /
+  neutral. If you need a 5-class scale or a continuous score, you'll need
+  to retrain or post-process.
+
+## Training details
+
+| | |
+|---|---|
+| Upstream base model | [facebook/MobileLLM-R1-950M](https://huggingface.co/facebook/MobileLLM-R1-950M) |
+| Loading mirror | [facebook/MobileLLM-R1-950M](https://huggingface.co/facebook/MobileLLM-R1-950M) (Unsloth's pre-quantized copy) |
+| Dataset | [Ayansk11/FinSenti-Dataset](https://huggingface.co/datasets/Ayansk11/FinSenti-Dataset) (~15.2K train per stage, 50.8K total across splits) |
+| SFT length | ~1.6 hours on A100 80GB |
+| GRPO budget | 3000 steps with early stopping (best near step ~200) |
+| Best GRPO reward | ~3.29 / 4.0 |
+| Adapter | LoRA (r=16, alpha=32) on q/k/v/o/gate/up/down projections |
+| Sequence length | 2048 |
+| Optimizer | AdamW (8-bit), cosine LR schedule |
+| Hardware | NVIDIA A100 80GB (Indiana University BigRed200 cluster) |
+| Frameworks | PEFT + bitsandbytes (no Unsloth - llama4_text arch unsupported) |
+
+## Related FinSenti models
+
+Other sizes and bases trained with the same recipe:
+
+- **Qwen3**: [Qwen3-0.6B](https://huggingface.co/Ayansk11/FinSenti-Qwen3-0.6B), [Qwen3-1.7B](https://huggingface.co/Ayansk11/FinSenti-Qwen3-1.7B), [Qwen3-4B](https://huggingface.co/Ayansk11/FinSenti-Qwen3-4B), [Qwen3-8B](https://huggingface.co/Ayansk11/FinSenti-Qwen3-8B)
+- **Qwen3.5**: [Qwen3.5-0.8B](https://huggingface.co/Ayansk11/FinSenti-Qwen3.5-0.8B), [Qwen3.5-2B](https://huggingface.co/Ayansk11/FinSenti-Qwen3.5-2B), [Qwen3.5-4B](https://huggingface.co/Ayansk11/FinSenti-Qwen3.5-4B), [Qwen3.5-9B](https://huggingface.co/Ayansk11/FinSenti-Qwen3.5-9B)
+- **DeepSeek**: [DeepSeek-R1-1.5B](https://huggingface.co/Ayansk11/FinSenti-DeepSeek-R1-1.5B)
+- **Tiny-LLM**: [Tiny-LLM-10M](https://huggingface.co/Ayansk11/FinSenti-Tiny-LLM-10M)
+- **Llama-3**: [Llama-3.2-1B](https://huggingface.co/Ayansk11/FinSenti-Llama-3.2-1B)
+- **SmolLM**: [SmolLM-1.7B](https://huggingface.co/Ayansk11/FinSenti-SmolLM-1.7B)
+
+There's a GGUF build of this same model at
+[Ayansk11/FinSenti-MobileLLM-R1-950M-GGUF](https://huggingface.co/Ayansk11/FinSenti-MobileLLM-R1-950M-GGUF) for Ollama and
+llama.cpp, and the dataset itself is at
+[Ayansk11/FinSenti-Dataset](https://huggingface.co/datasets/Ayansk11/FinSenti-Dataset).
+
+If you're picking a size, a rough guide:
+
+- **Need it on a phone or browser?** Look at the smallest model in the
+  group (Qwen3-0.6B) or its GGUF.
+- **Laptop with no GPU?** Any model up to ~2B as Q4_K_M GGUF works.
+- **Single 8-12 GB GPU?** The 1.5B-4B sizes are the sweet spot.
+- **Server or workstation?** The 8B / 9B variants give the best reasoning
+  but need the memory.
+
+## Citation
+
+If you use this model in research, please cite:
+
+```bibtex
+@misc{shaikh2026finsenti,
+  title  = {FinSenti: Small Language Models for Financial Sentiment with Chain-of-Thought Reasoning},
+  author = {Shaikh, Ayan},
+  year   = {2026},
+  url    = {https://huggingface.co/collections/Ayansk11/finsenti},
+  note   = {Indiana University}
+}
+```
+
+## License
+
+Apache 2.0, same as the base model.
+
+## Acknowledgements
+
+Trained on the Indiana University BigRed200 cluster (account `r01510`).
+Thanks to the Unsloth and TRL teams for the trainer stack, and to the
+Qwen / DeepSeek teams for the base models.