We couldn't connect to 'https://huggingface.co' to load this file, couldn't find it in the cached files and it looks like bigcode/starcoderbase-1b is not the path to a directory containing a file named config.json

Hi, I am a beginner in huggingface, I am reproducing Fuzz4All project (Article:Fuzz4All: Universal Fuzzing with Large Language Models), reaching the execution level of “./scripts/demo_run.sh” and I got like error :slight_smile: (fuzz4all) root@eebf26c919ec:/home/Fuzz4All# ./scripts/demo_run.sh
BATCH_SIZE: 30
MODEL_NAME: bigcode/starcoderbase-1b
DEVICE: cpu
{‘fuzzing’: {‘output_folder’: ‘outputs/demo/’, ‘num’: 100, ‘total_time’: 1, ‘log_level’: 3, ‘otf’: True, ‘resume’: True, ‘evaluate’: False, ‘use_hand_written_prompt’: False, ‘no_input_prompt’: True, ‘prompt_strategy’: 2, ‘target_name’: ‘/home/gcc-13/bin/g++’}, ‘target’: {‘language’: ‘cpp’, ‘path_documentation’: ‘config/documentation/cpp/cpp_23.md’, ‘path_example_code’: None, ‘trigger_to_generate_input’: ‘/* Please create a very short program which uses new C++ features in a complex way /', ‘input_hint’: ‘#include ’, ‘path_hand_written_prompt’: None, ‘target_string’: ‘’}, ‘llm’: {‘temperature’: 1, ‘batch_size’: 30, ‘device’: ‘cpu’, ‘model_name’: ‘bigcode/starcoderbase-1b’, ‘max_length’: 1024}}
=== Target Config ===
language: cpp
folder: outputs/demo/
bs: 30
temperature: 1
device: cpu
model_name: bigcode/starcoderbase-1b
max_length: 1024
use_hw: False
no_input_prompt: True
prompt_strategy: 2
level: 3
template: fuzzing_with_config_file
config_dict: {‘fuzzing’: {‘output_folder’: ‘outputs/demo/’, ‘num’: 100, ‘total_time’: 1, ‘log_level’: 3, ‘otf’: True, ‘resume’: True, ‘evaluate’: False, ‘use_hand_written_prompt’: False, ‘no_input_prompt’: True, ‘prompt_strategy’: 2, ‘target_name’: ‘/home/gcc-13/bin/g++’}, ‘target’: {‘language’: ‘cpp’, ‘path_documentation’: ‘config/documentation/cpp/cpp_23.md’, ‘path_example_code’: None, ‘trigger_to_generate_input’: '/
Please create a very short program which uses new C++ features in a complex way */’, ‘input_hint’: ‘#include ’, ‘path_hand_written_prompt’: None, ‘target_string’: ‘’}, ‘llm’: {‘temperature’: 1, ‘batch_size’: 30, ‘device’: ‘cpu’, ‘model_name’: ‘bigcode/starcoderbase-1b’, ‘max_length’: 1024}}
target_name: /home/gcc-13/bin/g++

[INFO] Initializing … this may take a while …
[INFO] Loading model …
=== Model Config ===
model_name: bigcode/starcoderbase-1b
model_name: bigcode/starcoderbase-1b
eos: [‘/* Please create a very short program which uses new C++ features in a complex way /', ‘’, '/ Please create a semantically equivalent program to the previous generation /', '/ Please create a mutated program that modifies the previous generation /', '/ Please combine the two previous programs into a single program */’]
device: cpu
max_length: 1024

OSError: We couldn’t connect to ‘https://huggingface.co’ to load this file, couldn’t find it in the cached files and it looks like bigcode/starcoderbase-1b is not the path to a directory
containing a file named config.json.
Checkout your internet connection or see how to run the library in offline mode at ‘Installation’.

I’m getting the same error :confused:

me too 12345678910 caracteres

me too, trying from colab though

I was recently doing the same. I fixed it by following the link: Installation

1.I installed entire “bigcode/starcoderbase” project in my local storage

2.I changed calls in model.py:
self.tokenizer = AutoTokenizer.from_pretrained(
“./your/path/bigcode/starcoderbase”,
local_files_only=True,
)
self.model = (
AutoModelForCausalLM.from_pretrained(
“./your/path/bigcode/starcoderbase”,
local_files_only=True,
)
.to(torch.bfloat16)
.to(device)
)