Upload ByteDance_Dolphin-v2_1.py with huggingface_hub

ByteDance_Dolphin-v2_1.py

@@ -0,0 +1,98 @@
+# /// script
+# requires-python = ">=3.12"
+# dependencies = [
+#     "numpy",
+#     "einops",
+#     "pandas",
+#     "matplotlib",
+#     "protobuf",
+#     "torch",
+#     "sentencepiece",
+#     "torchvision",
+#     "transformers",
+#     "timm",
+#     "diffusers",
+#     "sentence-transformers",
+#     "accelerate",
+#     "peft",
+#     "slack-sdk",
+# ]
+# ///
+
+try:
+    # Load model directly
+    from transformers import AutoProcessor, AutoModelForVision2Seq
+    
+    processor = AutoProcessor.from_pretrained("ByteDance/Dolphin-v2")
+    model = AutoModelForVision2Seq.from_pretrained("ByteDance/Dolphin-v2")
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
+                {"type": "text", "text": "What animal is on the candy?"}
+            ]
+        },
+    ]
+    inputs = processor.apply_chat_template(
+    	messages,
+    	add_generation_prompt=True,
+    	tokenize=True,
+    	return_dict=True,
+    	return_tensors="pt",
+    ).to(model.device)
+    
+    outputs = model.generate(**inputs, max_new_tokens=40)
+    print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))
+    with open('ByteDance_Dolphin-v2_1.txt', 'w', encoding='utf-8') as f:
+        f.write('Everything was good in ByteDance_Dolphin-v2_1.txt')
+except Exception as e:
+    import os
+    from slack_sdk import WebClient
+    client = WebClient(token=os.environ['SLACK_TOKEN'])
+    client.chat_postMessage(
+        channel='#hub-model-metadata-snippets-sprint',
+        text='Problem in <https://huggingface.co/datasets/model-metadata/code_execution_files/blob/main/ByteDance_Dolphin-v2_1.txt|ByteDance_Dolphin-v2_1.txt>',
+    )
+
+    with open('ByteDance_Dolphin-v2_1.txt', 'a', encoding='utf-8') as f:
+        import traceback
+        f.write('''```CODE: 
+# Load model directly
+from transformers import AutoProcessor, AutoModelForVision2Seq
+
+processor = AutoProcessor.from_pretrained("ByteDance/Dolphin-v2")
+model = AutoModelForVision2Seq.from_pretrained("ByteDance/Dolphin-v2")
+messages = [
+    {
+        "role": "user",
+        "content": [
+            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
+            {"type": "text", "text": "What animal is on the candy?"}
+        ]
+    },
+]
+inputs = processor.apply_chat_template(
+	messages,
+	add_generation_prompt=True,
+	tokenize=True,
+	return_dict=True,
+	return_tensors="pt",
+).to(model.device)
+
+outputs = model.generate(**inputs, max_new_tokens=40)
+print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))
+```
+
+ERROR: 
+''')
+        traceback.print_exc(file=f)
+    
+finally:
+    from huggingface_hub import upload_file
+    upload_file(
+        path_or_fileobj='ByteDance_Dolphin-v2_1.txt',
+        repo_id='model-metadata/code_execution_files',
+        path_in_repo='ByteDance_Dolphin-v2_1.txt',
+        repo_type='dataset',
+    )