gradio==5.1.0

.python-version

@@ -0,0 +1 @@
+3.10

README.md

@@ -4,7 +4,7 @@ emoji: 🐢
 colorFrom: purple
 colorTo: purple
 sdk: gradio
-sdk_version: 4.27.0
+sdk_version: 5.1.0
 app_file: app.py
 pinned: false
 ---

app.py

@@ -7,7 +7,12 @@ from typing import Iterator
 import gradio as gr
 import spaces
 import torch
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
+from transformers import (
+    AutoModelForCausalLM,
+    AutoTokenizer,
+    BitsAndBytesConfig,
+    TextIteratorStreamer,
+)
 
 DESCRIPTION = """# Swallow-13B instruct"""
 
@@ -18,7 +23,10 @@ if torch.cuda.is_available():
     model_name = "tokyotech-llm/Swallow-13b-instruct-hf"
     tokenizer = AutoTokenizer.from_pretrained(model_name)
     model = AutoModelForCausalLM.from_pretrained(
-        model_name, load_in_8bit=True, low_cpu_mem_usage=True, device_map="auto"
+        model_name,
+        quantization_config=BitsAndBytesConfig(load_in_8bit=True),
+        low_cpu_mem_usage=True,
+        device_map="auto",
     )
 
 MAX_INPUT_TOKENS = 2048
@@ -30,7 +38,9 @@ PROMPT_DICT = {
         "### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
     ),
     "prompt_no_input": (
-        "以下に、あるタスクを説明する指示があります。" "リクエストを適切に完了するための回答を記述してください。\n\n" "### 指示:\n{instruction}\n\n### 応答:"
+        "以下に、あるタスクを説明する指示があります。"
+        "リクエストを適切に完了するための回答を記述してください。\n\n"
+        "### 指示:\n{instruction}\n\n### 応答:"
     ),
 }
 
@@ -94,13 +104,8 @@ def process_example(instruction: str, input_text: str) -> Iterator[str]:
     yield from run(instruction, input_text)
 
 
-with gr.Blocks(css="style.css") as demo:
+with gr.Blocks(css_paths="style.css") as demo:
     gr.Markdown(DESCRIPTION)
-    gr.DuplicateButton(
-        value="Duplicate Space for private use",
-        elem_id="duplicate-button",
-        visible=os.getenv("SHOW_DUPLICATE_BUTTON") == "1",
-    )
 
     with gr.Row():
         with gr.Column():
@@ -125,8 +130,14 @@ with gr.Blocks(css="style.css") as demo:
 
     gr.Examples(
         examples=[
-            ["以下のトピックに関する詳細な情報を提供してください。", "東京工業大学の主なキャンパスについて教えてください。"],
-            ["以下のトピックに関する詳細な情報を提供してください。", "夢オチとは何かについて教えてください。"],
+            [
+                "以下のトピックに関する詳細な情報を提供してください。",
+                "東京工業大学の主なキャンパスについて教えてください。",
+            ],
+            [
+                "以下のトピックに関する詳細な情報を提供してください。",
+                "夢オチとは何かについて教えてください。",
+            ],
             ["暴れん坊将軍って誰のことですか？", ""],
         ],
         inputs=[instruction, input_text],
@@ -137,4 +148,4 @@ with gr.Blocks(css="style.css") as demo:
     )
 
 if __name__ == "__main__":
-    demo.queue(max_size=20).launch()
+    demo.launch()

pyproject.toml

@@ -0,0 +1,20 @@
+[project]
+name = "swallow-13b-instruct"
+version = "0.1.0"
+description = ""
+readme = "README.md"
+requires-python = ">=3.10"
+dependencies = [
+    "accelerate>=1.0.1",
+    "bitsandbytes>=0.44.1",
+    "blobfile>=3.0.0",
+    "gradio>=5.1.0",
+    "hf-transfer>=0.1.8",
+    "protobuf>=5.28.2",
+    "sentencepiece>=0.2.0",
+    "setuptools>=75.2.0",
+    "spaces>=0.30.4",
+    "tiktoken>=0.8.0",
+    "torch==2.4.0",
+    "transformers>=4.45.2",
+]

requirements.txt

@@ -1,8 +1,261 @@
-accelerate==0.29.3
-bitsandbytes==0.43.1
-gradio==4.27.0
-scipy==1.13.0
-sentencepiece==0.1.99
-spaces==0.26.1
-torch==2.0.0
-transformers==4.40.0
+# This file was autogenerated by uv via the following command:
+#    uv pip compile pyproject.toml -o requirements.txt
+accelerate==1.0.1
+    # via swallow-13b-instruct (pyproject.toml)
+aiofiles==23.2.1
+    # via gradio
+annotated-types==0.7.0
+    # via pydantic
+anyio==4.6.2.post1
+    # via
+    #   gradio
+    #   httpx
+    #   starlette
+bitsandbytes==0.44.1
+    # via swallow-13b-instruct (pyproject.toml)
+blobfile==3.0.0
+    # via swallow-13b-instruct (pyproject.toml)
+certifi==2024.8.30
+    # via
+    #   httpcore
+    #   httpx
+    #   requests
+charset-normalizer==3.4.0
+    # via requests
+click==8.1.7
+    # via
+    #   typer
+    #   uvicorn
+exceptiongroup==1.2.2
+    # via anyio
+fastapi==0.115.2
+    # via gradio
+ffmpy==0.4.0
+    # via gradio
+filelock==3.16.1
+    # via
+    #   blobfile
+    #   huggingface-hub
+    #   torch
+    #   transformers
+    #   triton
+fsspec==2024.9.0
+    # via
+    #   gradio-client
+    #   huggingface-hub
+    #   torch
+gradio==5.1.0
+    # via
+    #   swallow-13b-instruct (pyproject.toml)
+    #   spaces
+gradio-client==1.4.0
+    # via gradio
+h11==0.14.0
+    # via
+    #   httpcore
+    #   uvicorn
+hf-transfer==0.1.8
+    # via swallow-13b-instruct (pyproject.toml)
+httpcore==1.0.6
+    # via httpx
+httpx==0.27.2
+    # via
+    #   gradio
+    #   gradio-client
+    #   spaces
+huggingface-hub==0.26.0
+    # via
+    #   accelerate
+    #   gradio
+    #   gradio-client
+    #   tokenizers
+    #   transformers
+idna==3.10
+    # via
+    #   anyio
+    #   httpx
+    #   requests
+jinja2==3.1.4
+    # via
+    #   gradio
+    #   torch
+lxml==5.3.0
+    # via blobfile
+markdown-it-py==3.0.0
+    # via rich
+markupsafe==2.1.5
+    # via
+    #   gradio
+    #   jinja2
+mdurl==0.1.2
+    # via markdown-it-py
+mpmath==1.3.0
+    # via sympy
+networkx==3.4.1
+    # via torch
+numpy==2.1.2
+    # via
+    #   accelerate
+    #   bitsandbytes
+    #   gradio
+    #   pandas
+    #   transformers
+nvidia-cublas-cu12==12.1.3.1
+    # via
+    #   nvidia-cudnn-cu12
+    #   nvidia-cusolver-cu12
+    #   torch
+nvidia-cuda-cupti-cu12==12.1.105
+    # via torch
+nvidia-cuda-nvrtc-cu12==12.1.105
+    # via torch
+nvidia-cuda-runtime-cu12==12.1.105
+    # via torch
+nvidia-cudnn-cu12==9.1.0.70
+    # via torch
+nvidia-cufft-cu12==11.0.2.54
+    # via torch
+nvidia-curand-cu12==10.3.2.106
+    # via torch
+nvidia-cusolver-cu12==11.4.5.107
+    # via torch
+nvidia-cusparse-cu12==12.1.0.106
+    # via
+    #   nvidia-cusolver-cu12
+    #   torch
+nvidia-nccl-cu12==2.20.5
+    # via torch
+nvidia-nvjitlink-cu12==12.6.77
+    # via
+    #   nvidia-cusolver-cu12
+    #   nvidia-cusparse-cu12
+nvidia-nvtx-cu12==12.1.105
+    # via torch
+orjson==3.10.9
+    # via gradio
+packaging==24.1
+    # via
+    #   accelerate
+    #   gradio
+    #   gradio-client
+    #   huggingface-hub
+    #   spaces
+    #   transformers
+pandas==2.2.3
+    # via gradio
+pillow==10.4.0
+    # via gradio
+protobuf==5.28.2
+    # via swallow-13b-instruct (pyproject.toml)
+psutil==5.9.8
+    # via
+    #   accelerate
+    #   spaces
+pycryptodomex==3.21.0
+    # via blobfile
+pydantic==2.9.2
+    # via
+    #   fastapi
+    #   gradio
+    #   spaces
+pydantic-core==2.23.4
+    # via pydantic
+pydub==0.25.1
+    # via gradio
+pygments==2.18.0
+    # via rich
+python-dateutil==2.9.0.post0
+    # via pandas
+python-multipart==0.0.12
+    # via gradio
+pytz==2024.2
+    # via pandas
+pyyaml==6.0.2
+    # via
+    #   accelerate
+    #   gradio
+    #   huggingface-hub
+    #   transformers
+regex==2024.9.11
+    # via
+    #   tiktoken
+    #   transformers
+requests==2.32.3
+    # via
+    #   huggingface-hub
+    #   spaces
+    #   tiktoken
+    #   transformers
+rich==13.9.2
+    # via typer
+ruff==0.7.0
+    # via gradio
+safetensors==0.4.5
+    # via
+    #   accelerate
+    #   transformers
+semantic-version==2.10.0
+    # via gradio
+sentencepiece==0.2.0
+    # via swallow-13b-instruct (pyproject.toml)
+setuptools==75.2.0
+    # via swallow-13b-instruct (pyproject.toml)
+shellingham==1.5.4
+    # via typer
+six==1.16.0
+    # via python-dateutil
+sniffio==1.3.1
+    # via
+    #   anyio
+    #   httpx
+spaces==0.30.4
+    # via swallow-13b-instruct (pyproject.toml)
+starlette==0.40.0
+    # via fastapi
+sympy==1.13.3
+    # via torch
+tiktoken==0.8.0
+    # via swallow-13b-instruct (pyproject.toml)
+tokenizers==0.20.1
+    # via transformers
+tomlkit==0.12.0
+    # via gradio
+torch==2.4.0
+    # via
+    #   swallow-13b-instruct (pyproject.toml)
+    #   accelerate
+    #   bitsandbytes
+tqdm==4.66.5
+    # via
+    #   huggingface-hub
+    #   transformers
+transformers==4.45.2
+    # via swallow-13b-instruct (pyproject.toml)
+triton==3.0.0
+    # via torch
+typer==0.12.5
+    # via gradio
+typing-extensions==4.12.2
+    # via
+    #   anyio
+    #   fastapi
+    #   gradio
+    #   gradio-client
+    #   huggingface-hub
+    #   pydantic
+    #   pydantic-core
+    #   rich
+    #   spaces
+    #   torch
+    #   typer
+    #   uvicorn
+tzdata==2024.2
+    # via pandas
+urllib3==2.2.3
+    # via
+    #   blobfile
+    #   requests
+uvicorn==0.32.0
+    # via gradio
+websockets==12.0
+    # via gradio-client