Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+data/test.jsonl filter=lfs diff=lfs merge=lfs -text
+data/train.jsonl filter=lfs diff=lfs merge=lfs -text

data/images/1.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20813acc6638870c524e8b5246aae58d537e1a8f3548b7a53faa7993be014725
+size 713968566

data/images/10.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f34cb4bcca9e00495d224329d3e37c06fa753b1ee0b59c742ab9de81843457dd
+size 678272171

data/images/2.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22255d02908c1e757b8fd5039d868afbc03d189c074aa602117dfab2bb1ee3ea
+size 720714560

data/images/3.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce90ca867bb0cc3f19954ad40ac9422ba590f44be0a5900239e510f089e80ac3
+size 678512965

data/images/4.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5a0a2babe3226114406a93655b76f682b9d7c1ae1cff28bc5c13ec9097170e0
+size 665420259

data/images/5.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9be0ce29f958d2dbe6238f5d5e8e4991297146205e368c00ae3b7239c1e419f5
+size 724024231

data/images/6.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:475cf7ef8a8e24c0dd390052b2361fe0df358fe4f7dc2207af5bad9fa1c5e1de
+size 754277919

data/images/7.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e5e70646488430221d06c665b52e11ff5d8f7a7f6588f819b9b74781ed5675f
+size 682945336

data/images/8.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:921246d90f1bfc905de9adac17114a960e0c395ef56206ca2cc92f6faecbdefa
+size 755881389

data/images/9.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c9656453a518dc8729e041fdd68a3524933a0f5929ff68e8aa359b5d03a5c83
+size 731803010

data/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad04eb2fb3417f6127cefa77634040b1def227e1be68fd3b6760422aa488361a
+size 13480332

data/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:267ae5ee1c91d27a18cebfd9d4ecafc80b0ca89c665bdf4ae339812bff3d3bfd
+size 301314134

skyvern-gpt-4o.py

@@ -0,0 +1,128 @@
+import json
+import datasets
+import os
+
+logger = datasets.logging.get_logger(__name__)
+
+
+class Dataset(datasets.GeneratorBasedBuilder):
+    def _info(self):
+        return datasets.DatasetInfo(
+            features=datasets.Features({
+                "images": datasets.Sequence(datasets.Image()),
+                "length": datasets.Value(dtype="int32"),
+                "conversations": datasets.Sequence(datasets.Features({
+                    "from": datasets.Value("string"),
+                    "value": datasets.Value("string")
+                })),
+                "task_name": datasets.Value("string"),
+                "step_name": datasets.Value("string"),
+                "has_retry": datasets.Value("bool"),
+                "retry_index": datasets.Value("int32"),
+                "total_retries": datasets.Value("int32"),
+                "task_num_steps": datasets.Value("int32"),
+                "task_has_solve_captcha": datasets.Value("bool"),
+            })
+        )
+
+    def _split_generators(self, dl_manager: datasets.DownloadManager):
+        dl_manager.download_config.token = True
+        dl_manager.download_config.num_proc = 10
+
+        base_url = "https://huggingface.co/datasets/empower-dev-staging/skyvern-v0/resolve/main/data"
+        image_files = dl_manager.download_and_extract(
+            [f"{base_url}/images/{i + 1}.tar.gz" for i in range(10)])
+
+        image_file_to_full_path_mapping = dict([
+            ('images/' + '/'.join(image_file.split('/')[-2:]), image_file) for image_file in dl_manager.iter_files(image_files)
+        ])
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "filepath": dl_manager.download_and_extract(
+                        f"{base_url}/train.jsonl"),
+                    "image_file_to_full_path_mapping": image_file_to_full_path_mapping
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={
+                    "filepath": dl_manager.download_and_extract(
+                        f"{base_url}/test.jsonl"),
+                    "image_file_to_full_path_mapping": image_file_to_full_path_mapping
+                },
+            ),
+        ]
+
+    def _get_step_info(self, item):
+        first_image_path = item['images'][0]
+        folder = '/'.join(first_image_path.split('/')[-2:-1])
+
+        task = folder.split('-')[0]
+        step = folder.split('-')[1].split('_')
+
+        step_number = step[0]
+        retry_index = int(step[1])
+
+        return {
+            "task_name": task,
+            "step_name": f"{task}-{step_number}",
+            "retry_index": retry_index
+        }
+
+    def _generate_examples(self, filepath, image_file_to_full_path_mapping):
+        with open(filepath, "r") as f:
+            lines = f.readlines()
+
+            items = []
+            step_name_to_retry_indices = {}
+            task_name_to_num_steps = {}
+            task_name_to_having_solve_captcha = {}
+            for id, line in enumerate(lines):
+                item = json.loads(line)
+                actions = json.loads(item["conversations"][1]["value"])[
+                    "actions"]
+                if len(actions) == 0:
+                    continue
+
+                items.append(item)
+
+                step_info = self._get_step_info(item)
+                step_name = step_info["step_name"]
+                task_name = step_info["task_name"]
+
+                if task_name not in task_name_to_having_solve_captcha:
+                    task_name_to_having_solve_captcha[task_name] = False
+                if any(action["action_type"].lower() == "solve_captcha" for action in actions):
+                    task_name_to_having_solve_captcha[task_name] = True
+
+                if step_name not in step_name_to_retry_indices:
+                    step_name_to_retry_indices[step_name] = []
+                    task_name_to_num_steps[task_name] = task_name_to_num_steps.get(
+                        task_name, 0) + 1
+                step_name_to_retry_indices[step_name].append(
+                    step_info["retry_index"])
+
+            step_name_to_retry_indices = dict([
+                (step_name, sorted(retry_indices)) for (step_name, retry_indices) in step_name_to_retry_indices.items()
+            ])
+
+            for id, item in enumerate(items):
+                step_info = self._get_step_info(item)
+                retry_indices = step_name_to_retry_indices[step_info['step_name']]
+                yield id, {
+                    "images": [
+                        image_file_to_full_path_mapping[image] for image in item["images"]
+                    ],
+                    "conversations": item["conversations"],
+                    "length": item["length"],
+                    "task_name": step_info["task_name"],
+                    "step_name": step_info["step_name"],
+                    "has_retry": len(retry_indices) > 1,
+                    "retry_index": retry_indices.index(step_info["retry_index"]),
+                    "total_retries": len(retry_indices),
+                    "task_num_steps": task_name_to_num_steps[step_info["task_name"]],
+                    "task_has_solve_captcha": task_name_to_having_solve_captcha[step_info["task_name"]],
+                }