Add application file

.gitignore

@@ -0,0 +1,3 @@
+.idea
+
+__pycache__

app.py

@@ -0,0 +1,47 @@
+import pathlib
+
+import gradio as gr
+import torch
+from scipy.special import softmax
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+
+# load model
+MODEL_PATH = pathlib.Path("data") / "roberta-large-wmt"
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
+model = AutoModelForSequenceClassification.from_pretrained(MODEL_PATH)
+
+model.to("cuda" if torch.cuda.is_available() else "cpu")
+
+
+def detect(text: str) -> str:
+    inputs = tokenizer(text, return_tensors="pt")
+    with torch.no_grad():
+        logits = model(**inputs).logits
+    predicted_class_id = logits.argmax(dim=1).item()
+
+    label = model.config.id2label[predicted_class_id]
+
+    result = "machine-generated" if label == "0" else "human-generated"
+
+    return ", ".join(
+        [
+            f"The text is {result}",
+            f"with a certainty of ${100 * softmax(logits, axis=1)[0][int(label)]:.2f}%",
+        ]
+    )
+
+
+with gr.Blocks() as demo:
+    gr.Markdown(
+        """
+        # Hello!
+        This is the demo for <a href="https://arxiv.org/abs/2305.12680" target="_blank">G3Detector</a>.
+        """
+    )
+    inp = gr.Textbox(label="Text", placeholder="Paste text here...")
+    out = gr.Textbox(label="Result")
+    detect_btn = gr.Button("Detect")
+    detect_btn.click(fn=detect, inputs=inp, outputs=out, api_name="G3Detector")
+
+demo.launch()

data/roberta-large-wmt/all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.0,
+    "eval_loss": 9.865092277526855,
+    "eval_runtime": 24.5645,
+    "eval_samples": 3000,
+    "eval_samples_per_second": 122.127,
+    "eval_steps_per_second": 15.266,
+    "train_loss": 0.05986976114908854,
+    "train_runtime": 467.209,
+    "train_samples": 4000,
+    "train_samples_per_second": 25.684,
+    "train_steps_per_second": 0.803
+}

data/roberta-large-wmt/config.json

@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "../pretrained_models/roberta-large",
+  "architectures": [
+    "RobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1024,
+  "id2label": {
+    "0": "0",
+    "1": "1"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "label2id": {
+    "0": 0,
+    "1": 1
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.14.0.dev0",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

data/roberta-large-wmt/eval_results.json

@@ -0,0 +1,8 @@
+{
+    "eval_accuracy": 0.0,
+    "eval_loss": 9.865092277526855,
+    "eval_runtime": 24.5645,
+    "eval_samples": 3000,
+    "eval_samples_per_second": 122.127,
+    "eval_steps_per_second": 15.266
+}

data/roberta-large-wmt/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82cd188fc5001427fe27ab72fce2ff7c7c3dc3fcc07e83d31d666c0ea9f31170
+size 1421616717

data/roberta-large-wmt/runs/May20_22-09-26_m3h006/1684584588.9904363/events.out.tfevents.1684584588.m3h006.9656.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee38d96706910519229b8cc97ac25858f84e5f652ab0189d8b19d60cee502dbb
+size 4729

data/roberta-large-wmt/runs/May20_22-09-26_m3h006/events.out.tfevents.1684584588.m3h006.9656.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b9865db1e84c28d5b9e7415514cd9a482c8d23332e663df26b31e29b079a9c6
+size 3776

data/roberta-large-wmt/runs/May20_22-09-26_m3h006/events.out.tfevents.1684585101.m3h006.9656.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50312bc13839ce4f93644a25129405069667f1dc82fecb60e249e1df16989d8a
+size 411

data/roberta-large-wmt/runs/May20_22-19-20_m3h006/events.out.tfevents.1684585194.m3h006.11791.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1a9f24f63934c83f4163a9d22253781de8c55037e598ff8681ffcd60695840b
+size 346

data/roberta-large-wmt/runs/May20_22-21-52_m3h006/events.out.tfevents.1684585347.m3h006.11945.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:109c484dc0944c11ff5bb7cf64e43106f68e58377ce79e3963d4c632da02db7a
+size 346

data/roberta-large-wmt/runs/May21_12-15-58_m3-login1/events.out.tfevents.1684637428.m3-login1.888308.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4af362826610b64151fb885838595134f2eb817a0cd8d01dbbafa6169a013a41
+size 346

data/roberta-large-wmt/runs/May29_12-47-52_m3-login1/events.out.tfevents.1685332847.m3-login1.2429874.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d359a0bce4306bbf1c3b965148f13ff650b371a7c668e503c8252d562470dc7
+size 346

data/roberta-large-wmt/runs/May29_14-17-10_m3-login1/events.out.tfevents.1685334472.m3-login1.3030729.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6215f803425b5027debfb198fa6979aa7e0cbf6a6593c17dc1d7ab93cb59416
+size 346

data/roberta-large-wmt/runs/May30_22-24-13_m3g021/events.out.tfevents.1685449490.m3g021.8428.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e16a988fe5380d0e5412ac8e6921c53a20f15b2e36e9303c8dea3ed175dafbac
+size 346

data/roberta-large-wmt/runs/May30_22-25-29_m3g021/events.out.tfevents.1685449589.m3g021.8584.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9070109320cffa38989216524da0f5528c6934794e69a3b49f486be234fb286
+size 346

data/roberta-large-wmt/special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}

data/roberta-large-wmt/tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "special_tokens_map_file": null, "name_or_path": "../pretrained_models/roberta-large", "tokenizer_class": "RobertaTokenizer"}

data/roberta-large-wmt/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.0,
+    "train_loss": 0.05986976114908854,
+    "train_runtime": 467.209,
+    "train_samples": 4000,
+    "train_samples_per_second": 25.684,
+    "train_steps_per_second": 0.803
+}

data/roberta-large-wmt/trainer_state.json

@@ -0,0 +1,25 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "global_step": 375,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 3.0,
+      "step": 375,
+      "total_flos": 2795794089984000.0,
+      "train_loss": 0.05986976114908854,
+      "train_runtime": 467.209,
+      "train_samples_per_second": 25.684,
+      "train_steps_per_second": 0.803
+    }
+  ],
+  "max_steps": 375,
+  "num_train_epochs": 3,
+  "total_flos": 2795794089984000.0,
+  "trial_name": null,
+  "trial_params": null
+}

data/roberta-large-wmt/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ae53eff0ce2eacf03ac69d3806648b65f0af4937e5d4ddd3e98a56923bdf1c1
+size 2927

requirements.txt

@@ -0,0 +1,4 @@
+gradio
+scipy
+torch
+transformers