added

Dockerfile

@@ -38,4 +38,5 @@ RUN --mount=type=secret,id=HF_TOKEN,mode=0444,required=true \
     git config --global credential.helper store && \
     huggingface-cli login --token $(cat /run/secrets/HF_TOKEN) --add-to-git-credential
 RUN poetry run python model_pull.py
+EXPOSE 7860
 ENTRYPOINT ["/home/user/.local/bin/poetry", "run", "python", "app.py", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -2,17 +2,13 @@ from typing import Optional
 
 import gradio as gr
 import torch
-import transformers
 from peft import PeftModel
 from transformers import GenerationConfig
+from transformers import LlamaForCausalLM
+from transformers import LlamaTokenizer
 
 print("starting server ...")
 
-assert (
-    "LlamaTokenizer" in transformers._import_structure["models.llama"]
-), "LLaMA is now in HuggingFace's main branch.\nPlease reinstall it: pip uninstall transformers && pip install git+https://github.com/huggingface/transformers.git"
-from transformers import LlamaForCausalLM  # noqa
-from transformers import LlamaTokenizer  # noqa
 
 BASE_MODEL = "decapoda-research/llama-13b-hf"
 LORA_WEIGHTS = "izumi-lab/llama-13b-japanese-lora-v0-1ep"
@@ -139,4 +135,5 @@ g = gr.Interface(
     description="izumi-lab/calm-7b-lora-v0-1ep is a 7B-parameter Calm model finetuned to follow instructions. It is trained on the [izumi-lab/llm-japanese-dataset](https://huggingface.co/datasets/izumi-lab/llm-japanese-dataset) dataset and makes use of the Huggingface Calm-7b implementation. For more information, please visit [the project's website](https://llm.msuzuki.me).",
 )
 g.queue(concurrency_count=1)
+print("loading completed")
 g.launch()

model_pull.py

@@ -1,12 +1,7 @@
 import torch
-import transformers
 from peft import PeftModel
-
-assert (
-    "LlamaTokenizer" in transformers._import_structure["models.llama"]
-), "LLaMA is now in HuggingFace's main branch.\nPlease reinstall it: pip uninstall transformers && pip install git+https://github.com/huggingface/transformers.git"
-from transformers import LlamaForCausalLM  # noqa
-from transformers import LlamaTokenizer  # noqa
+from transformers import LlamaForCausalLM
+from transformers import LlamaTokenizer
 
 BASE_MODEL = "decapoda-research/llama-13b-hf"
 LORA_WEIGHTS = "izumi-lab/llama-13b-japanese-lora-v0-1ep"