Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,144 @@
+---
+language:
+- en
+license: apache-2.0
+library_name: peft
+tags:
+- forecasting
+- prediction
+- reinforcement-learning
+- grpo
+- lora
+- mixture-of-experts
+- golf
+- sports
+- future-as-label
+datasets:
+- LightningRodLabs/GolfForecasting
+base_model: openai/gpt-oss-120b
+pipeline_tag: text-generation
+model-index:
+- name: Golf-Forecaster
+  results:
+  - task:
+      type: text-generation
+      name: Probabilistic Forecasting
+    dataset:
+      name: GolfForecasting
+      type: LightningRodLabs/GolfForecasting
+      split: test
+    metrics:
+    - type: brier_score
+      value: 0.207
+      name: Brier Score
+    - type: ece
+      value: 0.062
+      name: Expected Calibration Error
+---
+
+# Golf-Forecaster
+
+### RL-Tuned gpt-oss-120b for Predicting Professional Golf Outcomes
+
+We fine-tuned [gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b) with reinforcement learning to predict professional golf outcomes across PGA Tour, LIV Golf, LPGA, DP World Tour, majors, and the Ryder Cup. Trained on the [GolfForecasting](https://huggingface.co/datasets/LightningRodLabs/GolfForecasting) dataset of 3,178 binary forecasting questions generated with the [Lightning Rod SDK](https://github.com/lightning-rod-labs/lightningrod-python-sdk), Golf-Forecaster beats GPT-5.1 on held-out forecasting questions.
+
+This repo contains a **LoRA adapter** (5.3 GB) for gpt-oss-120b. A standalone `merge.py` script is included to produce a full merged model if needed.
+
+[Dataset](https://huggingface.co/datasets/LightningRodLabs/GolfForecasting) · [Lightning Rod SDK](https://github.com/lightning-rod-labs/lightningrod-python-sdk) · [Future-as-Label paper](https://arxiv.org/abs/2601.06336) · [Outcome-based RL paper](https://arxiv.org/abs/2505.17989)
+
+---
+
+## Results
+
+Evaluated on 855 held-out test questions (temporal split, Aug 2025+). Golf-Forecaster achieves the best Brier score, highest skill score, and best calibration.
+
+| Model | Brier Score | Brier Skill Score | ECE |
+|-------|:---:|:---:|:---:|
+| **Golf-Forecaster** | **0.207** | **+17.0%** | **0.062** |
+| gpt-oss-120b (base) | 0.218 | +12.8% | 0.083 |
+| GPT-5.1 | 0.218 | +12.8% | 0.106 |
+
+### Metrics
+
+- **Brier Score**: Mean squared error between predicted probability and outcome (0 or 1). Lower is better. **Brier Skill Score (BSS)** expresses this as improvement over always predicting the base rate — positive means the model learned something useful beyond historical frequency.
+- **Expected Calibration Error (ECE)**: Measures whether predicted probabilities match actual frequencies. "70%" predictions should resolve "yes" 70% of the time. Lower is better.
+
+---
+
+## Training
+
+- **Base model**: [openai/gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b) (120B MoE, 5.1B active params, 128 experts Top-4)
+- **Method**: GRPO with Brier score reward via [Tinker](https://tinker.computer)
+- **LoRA rank**: 32
+- **Learning rate**: 4e-5
+- **Batch size**: 32, group size 8
+- **Training steps**: 100
+- **Max tokens**: 16,384
+
+---
+
+## Usage
+
+This repo contains a LoRA adapter trained with [Tinker](https://tinker.computer). The adapter uses Tinker's module naming convention, so it requires a merge step before inference. A standalone `merge.py` script is included.
+
+### Merge into full model
+
+```bash
+pip install torch transformers safetensors tqdm huggingface-hub
+python merge.py --output ./golf-forecaster-merged
+```
+
+This downloads the base model, dequantizes to bf16, applies the LoRA adapter, and saves the merged model.
+
+### Inference with the merged model
+
+With [SGLang](https://github.com/sgl-project/sglang) (recommended for MoE):
+
+```python
+import sglang as sgl
+
+engine = sgl.Engine(
+    model_path="./golf-forecaster-merged",
+    tokenizer_path="openai/gpt-oss-120b",
+    trust_remote_code=True,
+    dtype="bfloat16",
+    tp_size=2,
+)
+
+prompt = """You are a forecasting expert. Given the question and context below, predict the probability that the answer is "Yes".
+
+Question: Will Scottie Scheffler win the 2025 Masters?
+
+Respond with your reasoning, then give your final answer as a probability between 0 and 1 inside <answer></answer> tags."""
+
+output = engine.generate(prompt, sampling_params={"max_new_tokens": 4096, "stop": ["</answer>"]})
+print(output["text"])
+```
+
+Or with transformers:
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model = AutoModelForCausalLM.from_pretrained(
+    "./golf-forecaster-merged",
+    torch_dtype="auto",
+    device_map="auto",
+    trust_remote_code=True,
+)
+tokenizer = AutoTokenizer.from_pretrained("openai/gpt-oss-120b", trust_remote_code=True)
+
+inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=4096, do_sample=True, temperature=0.7)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+```
+
+---
+
+## Links
+
+- **Dataset**: [LightningRodLabs/GolfForecasting](https://huggingface.co/datasets/LightningRodLabs/GolfForecasting)
+- **Training platform**: [Tinker](https://tinker.computer)
+- **Data generation**: [Lightning Rod SDK](https://github.com/lightning-rod-labs/lightningrod-python-sdk)
+- **Future-as-Label paper**: [arxiv:2601.06336](https://arxiv.org/abs/2601.06336)
+- **Outcome-based RL paper**: [arxiv:2505.17989](https://arxiv.org/abs/2505.17989)