Upload README.md with huggingface_hub

README.md

@@ -1,202 +1,103 @@
 ---
-base_model: EleutherAI/pythia-160m
+language:
+- en
 library_name: peft
+pipeline_tag: text-generation
+base_model: EleutherAI/pythia-160m
+base_model_relation: adapter
+datasets:
+- cc0de/Enron_email
+tags:
+- lora
+- privacy
+- pii
+- membership-inference
+- masked-fine-tuning
 ---
 
-# Model Card for Model ID
-
-<!-- Provide a quick summary of what the model is/does. -->
-
-
-
-## Model Details
-
-### Model Description
-
-<!-- Provide a longer summary of what this model is. -->
-
-
-
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-
-### Model Sources [optional]
-
-<!-- Provide the basic links for the model. -->
-
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
-
-## Uses
-
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-
-### Direct Use
-
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-
-[More Information Needed]
-
-### Downstream Use [optional]
-
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-
-[More Information Needed]
-
-### Out-of-Scope Use
-
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-
-[More Information Needed]
-
-## Bias, Risks, and Limitations
-
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-
-[More Information Needed]
-
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
-
-## How to Get Started with the Model
-
-Use the code below to get started with the model.
-
-[More Information Needed]
-
-## Training Details
-
-### Training Data
-
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-
-[More Information Needed]
-
-### Training Procedure
-
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-
-#### Preprocessing [optional]
-
-[More Information Needed]
-
-
-#### Training Hyperparameters
-
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-
-#### Speeds, Sizes, Times [optional]
-
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-
-[More Information Needed]
-
-## Evaluation
-
-<!-- This section describes the evaluation protocols and provides the results. -->
-
-### Testing Data, Factors & Metrics
-
-#### Testing Data
-
-<!-- This should link to a Dataset Card if possible. -->
-
-[More Information Needed]
-
-#### Factors
-
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
-
-[More Information Needed]
-
-#### Metrics
-
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-
-[More Information Needed]
-
-### Results
-
-[More Information Needed]
-
-#### Summary
-
-
-
-## Model Examination [optional]
-
-<!-- Relevant interpretability work for the model goes here -->
-
-[More Information Needed]
-
-## Environmental Impact
-
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
-
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
-
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
+# TMFT: Targeted Masked Fine-Tuning
 
-## Technical Specifications [optional]
+This project tests whether masking privacy-sensitive token losses during LoRA
+fine-tuning reduces PII memorization with less utility degradation than random
+masking. It is an empirical mitigation study, not differential privacy or
+machine unlearning.
 
-### Model Architecture and Objective
+## Vessel Setup
 
-[More Information Needed]
+Upload the entire `tmft_project/` directory and open a terminal in that
+directory.
 
-### Compute Infrastructure
+```bash
+python -m pip install -U pip
+python -m pip uninstall -y transformers peft accelerate tokenizers huggingface_hub datasets
+python -m pip install -r requirements.txt
+python -m spacy download en_core_web_sm
+```
 
-[More Information Needed]
+Restart the Jupyter kernel after installation. The tested compatibility stack
+uses PyTorch 2.3.1, Transformers 4.41.2, PEFT 0.11.1, and Datasets 2.20.0.
 
-#### Hardware
+## End-to-End Experiment
 
-[More Information Needed]
+Prepare real PII-containing Enron splits and a real prefix-target evaluation
+set:
 
-#### Software
+```bash
+python main.py --mode prepare --force_prepare
+```
 
-[More Information Needed]
+Train all conditions:
 
-## Citation [optional]
+```bash
+python main.py --mode train --method all
+```
 
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+Evaluate TER, SER, held-out perplexity, MDP, Loss-MIA AUC, and Min-K MIA AUC:
 
-**BibTeX:**
+```bash
+python main.py --mode eval --method all
+```
 
-[More Information Needed]
+Generate result figures:
 
-**APA:**
+```bash
+python main.py --mode plot
+```
 
-[More Information Needed]
+The full pipeline can be launched with:
 
-## Glossary [optional]
+```bash
+python main.py --mode all --method all --force_prepare
+```
 
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+For an interactive run, execute `tmft_experiment.ipynb` from top to bottom.
 
-[More Information Needed]
+## Experimental Conditions
 
-## More Information [optional]
+- `baseline`: standard LoRA fine-tuning
+- `rmft`: random 15% loss masking
+- `tmft_ner`: loss masking at spaCy plus regex PII spans
+- `tmft_mia`: online token masking where the current model is more confident
+  than the frozen base model
+- `tmft_combined`: union of NER and post-warm-up MIA masks
 
-[More Information Needed]
+## Outputs
 
-## Model Card Authors [optional]
+- `data/processed/`: train, validation, and test DatasetDict
+- `data/pii_eval.json`: automatically generated real PII prefix-target attacks
+- `results/<method>/`: LoRA adapters and training metadata
+- `results/tables/main_results.csv`: submission-ready numeric table
+- `results/figures/`: PNG and PDF privacy/utility figures
 
-[More Information Needed]
+Do not report results if preprocessing prints a synthetic fallback warning.
+The final config disables fallback so an unavailable real dataset fails loudly.
 
-## Model Card Contact
+## Hugging Face Upload
 
-[More Information Needed]
-### Framework versions
+```bash
+huggingface-cli login
+python main.py --mode upload --method tmft_combined \
+  --hf_repo_id YOUR_USERNAME/tmft-pythia-160m-tmft-combined
+```
 
-- PEFT 0.11.1
+Use `--public` only after checking that the saved artifacts contain no raw PII.