Add granite-3.3-8b-instruct-lora-pii-detector (#7)

- Add granite-3.3-8b-instruct-lora-pii-detector (88708448e4ddc6bcf54d6007de9b30eb9e42f697)

granite-3.3-8b-instruct-lora-pii-detector/README.md

@@ -0,0 +1,85 @@
+---
+license: apache-2.0
+language:
+- en
+pipeline_tag: text-generation
+library_name: transformers
+---
+
+# Granite 3.3 8B Instruct - PII Detector LoRA
+
+Welcome to Granite Experiments!
+
+Think of Experiments as a preview of what's to come. These projects are still under development, but we wanted to let the open-source community take them for spin! Use them, break them, and help us build what's next for Granite - we'll keep an eye out for feedback and questions. Happy exploring!
+
+Just a heads-up: Experiments are forever evolving, so we can't commit to ongoing support or guarantee performance.
+
+## Model Summary
+
+This is an LoRA adapter for [ibm-granite/granite-3.3-8b-instruct](https://huggingface.co/ibm-granite/granite-3.3-8b-instruct), 
+adding the capability to detect Personally Identifiable Information (PII) in model outputs.
+
+- **Developer:** IBM Research
+- **Model type:** LoRA adapter for [ibm-granite/granite-3.3-8b-instruct](https://huggingface.co/ibm-granite/granite-3.3-8b-instruct)
+- **License:** [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0)
+
+## Usage
+
+### Intended use
+
+This is an experimental LoRA that is designed for detecting PII in model outputs.
+Models with access to personal information via RAG or similar may present additional data protection risks that can be mitigated by using this LoRA to check model outputs.
+
+**PII Detection**: The model identifies PII when the special role `<|start_of_role|>privacy<|end_of_role|>` is included in prompts. Without this role, the model behaves like the base model.
+
+### Quickstart Example
+
+The following code describes how to use the LoRA adapter model to detect PII in a model's outputs.
+
+```python
+import torch
+from transformers import AutoTokenizer,  AutoModelForCausalLM
+from peft import PeftModel
+
+BASE_NAME = "ibm-granite/granite-3.3-8b-instruct"
+LORA_NAME = "intrinsics/granite-3.3-8b-instruct-lora-pii-detector" # LoRA download location. We assume the directory shown in the top level README.md example for the lib was followed.
+PRIVACY_PROMPT = "<|start_of_role|>privacy<|end_of_role|>"
+
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# Load model
+tokenizer = AutoTokenizer.from_pretrained(BASE_NAME, padding_side='right', trust_remote_code=True)
+model_base = AutoModelForCausalLM.from_pretrained(BASE_NAME, device_map="auto")
+pii_detector_model = PeftModel.from_pretrained(model_base, LORA_NAME)
+
+# Detect PII
+model_output = "Taylor Swift lives in New York City and earned over $1bn from her last tour."
+chat = [{"role": "user", "content": model_output}]
+chat = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=False)
+chat = chat + PRIVACY_PROMPT
+
+inputs = tokenizer(chat, return_tensors="pt")
+output = pii_detector_model.generate(inputs["input_ids"].to(device), attention_mask=inputs["attention_mask"].to(device), max_new_tokens=1)
+output_text = tokenizer.decode(output[0][-1])
+print(f"PII detected: {output_text}")
+
+# Y - yes, PII detected.
+# N - no, PII likely not present.
+```
+
+## Training Details
+
+The model was fine-tuned using a combination of open-source datasets, consisting of both benign samples and those with PII.
+The datasets were processed for PII using [IBM's READI library](https://github.ibm.com/security-foundation-models/READI), and filtered to remove examples with non-ASCII text and code.
+
+## Evaluation
+
+The PII Detector LoRA was evaluated against examples from the same distribution as the training dataset, but not seen previously by the model.
+
+| Model                            | Accuracy | TPR   | FPR   | FNR   |
+|----------------------------------|----------|-------|-------|-------| 
+| Granite 3.3 8B LoRA pii detector | 0.970    | 0.976 | 0.024 | 0.036 |
+
+## Contact 
+
+Naoise Holohan, Giulio Zizzo

granite-3.3-8b-instruct-lora-pii-detector/adapter_config.json

@@ -0,0 +1,35 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "ibm-granite/granite-3.3-8b-instruct",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "q_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

granite-3.3-8b-instruct-lora-pii-detector/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93434fdda62784f44e2f6516ec9371883d56c4a3af8d411ab87d50a3042cc7a5
+size 94404160