unslothai/unsloth

[Bug] pulling models from local repository breaks with new name in lower case.

Open

#2,506 opened on May 9, 2025

View on GitHub
 (14 comments) (0 reactions) (1 assignee)Python (64,271 stars) (5,658 forks)batch import
feature requesthelp wantedinactive

Description

We pull models into a local artifactory repository.

os.environ["HF_HUB_ETAG_TIMEOUT"] = "86400" os.environ["HF_HUB_DOWNLOAD_TIMEOUT"] = "86400" os.environ["HF_ENDPOINT"] = "https://artifactory.nowhere.com/artifactory/api/huggingfaceml/AIMLmodels-huggingfaceML-remote"

from unsloth import FastLanguageModel import torch max_seq_length = 1024 # Can increase for longer reasoning traces lora_rank = 32 # Larger rank = smarter, but slower

model, tokenizer = FastLanguageModel.from_pretrained( model_name = "unsloth/Meta-Llama-3.1-8B-Instruct", max_seq_length = max_seq_length, load_in_4bit = True, # False for LoRA 16bit fast_inference = True, # Enable vLLM fast inference max_lora_rank = lora_rank, gpu_memory_utilization = 0.6, # Reduce if out of memory )

Errors out with name in lower case: HTTPError: 401 Client Error: for url: https://artifactory.nowhere.com/artifactory/api/huggingfaceml/AIMLmodels-huggingfaceML-remote/api/models/unsloth/meta-llama-3.1-8b-instruct-unsloth-bnb-4bit

The actual model name is as in the huggingface.co repository: Meta-Llama-3.1-8B-Instruct-unsloth-bnb-4bit

How can this issue be resolved?

Contributor guide